Chevron Gray Pump Manual

100 General Information
Abstract
The General Information section is an overview of all categories of pumps and a
guide which directs readers to more detailed information on specific pump catego-
ries in later sections.
System Hydraulic Design is included in this section because it is the first step in
work involving all pump categories. This is a critical step in pump selection and
often in troubleshooting as well.
This section does not include detailed information on any particular pump or appli-
cation. Except for the section covering system hydraulics, do not use the informa-
tion in this section as a design basis.
Contents Page
110 Introduction 100-3
111 Background
120 Pump Categories 100-4
121 Classification
122 Centrifugal Pumps
123 Positive Displacement Pumps
124 Oil Field and Special Service Pumps
125 Multiphase Pumps
130 System Hydraulic Design 100-7
131 Introduction
132 Overview of the Hydraulic Calculation Sheet, PMP-EF-467
133 PMP-EF-467 Instructions
134 Calculation I (Initial Estimate)
135 Calculation II (Pre-purchase)
136 Calculation III (Final Review)
November 2006 © 1999–2006 Chevron USA Inc. All rights reserved. 100-1
100 General Information Pump Manual
140 Application and Selection Criteria 100-23

141 Introduction
142 Pump Selection Fundamentals
143 Abrasives
145 Reciprocating Pumps
146 Rotary Pumps
147 Miscellaneous Pumps
150 Considerations for H2S Services 100-34
160 Suction Piping Design Considerations 100-34
161 NPSHA and Fluid Velocity
162 Adequate Venting
163 Fluid Velocity Profile Entering Pump Suction
100-2 © 1999–2006 Chevron USA Inc. All rights reserved. November 2006
Pump Manual 100 General Information
110 Introduction
111 Background
Most work involving pumps falls into one of three categories:
• Purchasing and installing new pumps
• Troubleshooting pump problems during startup or while in service
• Modifying pumps to resolve problems or to accommodate operational or
service changes (different flows, liquid, head, etc.)
The overall goal in all three cases is the same: To maximize Company profits while
providing safe, reliable equipment which satisfies operating requirements and local
environmental constraints. Profitability is a long-term goal involving the following
factors:
• Meeting safety and environmental needs (including noise restrictions)
• Initial pump cost
• Installation and commissioning expense
• Reliability
• Maintaining production (this is often the overwhelming economic factor)
• Operability (troublesome equipment wastes resources that can be spent on prof-
itable work)
• Starting up on time, the first time, in critical services
• Maintenance expense
• Energy consumption (this is less important for pumps than for compressors, but
it can still be a significant factor)
Although each of these factors must be considered when making decisions, there is
some degree of conflict between them. For example, buying an inexpensive pump
may keep initial costs down, but the cheaper pump may also be less reliable and
require frequent maintenance or special, costly installation. On the other hand,
buying the ultimately reliable pump may be prohibitively expensive in up-front
costs and unnecessary to ensure acceptable performance.
Use judgment when deciding which factors are the most important. It is especially
important to communicate with the people responsible for operating and main-
taining the pump. Their input on the relative priority of the above factors is invalu-
able. In the end, however, engineering judgment will always be necessary.
120 Pump Categories

All specific pump types covered in this manual fall into one of the five categories
illustrated in Figure 100-1.
Fig. 100-1 Pump Categories
121 Classification
Pumps are divided into two major classifications: centrifugal (kinetic) and positive
displacement. These two divisions are further divided into numerous subdivisions.
Figure 100-2 and Figure 100-3 illustrate the many pump types available. The pumps
shown in boldface are the most common in the petroleum, chemical, and related
industries.

Centrifugal pumps are the most frequently used pumps. They are widely accepted
because they combine a relatively low initial cost with high reliability, compact size,
non-pulsating flow, and easy maintenance. They are also widely available, cover
broad flow/pressure application ranges, and can operate over a wide flow range.
Centrifugal pumps are usually purchased to meet one of two levels of duty:
• General and Chemical pumps are commonly used in non-critical, non-
hazardous services. Most, but not all, of these pumps are built to ANSI (Amer-
ican National Standards Institute) Standard B73.1 (Horizontal End-Suction) or
ANSI Standard B73.2 (Vertical In-Line) specifications.
• Heavy Duty pumps are commonly used in refinery applications. This pump-
type is suitable for critical, hazardous, or “heavy duty” service including chemi-
cals, refining, and producing services. Virtually all pumps in this category are
built to API (American Petroleum Institute) Standard API 610 (“Centrifugal
Pumps for General Refinery Services”) specification. A Company model speci-
fication (PMP-MS-983) which supplements API 610 is also included in this
manual.
Fig. 100-2 Pump Categories Courtesy of the Hydraulic Institute (1 of 2)
Many General/Chemical pumps are in service in refineries and many Heavy Duty
pumps are used outside refineries. The determining factors are the required reli-
ability of design and ruggedness of construction. The Company specifications noted
above contain additional information on their respective applications.
Centrifugal pumps used as Fire Pumps (firewater delivery) are most often Heavy
Duty pumps.
See Section 200 for more information on centrifugal pumps.
123 Positive Displacement Pumps

Although not as common as centrifugal pumps, positive displacement pumps are
also widely employed. The three major positive displacement pump sub-categories
are:
• Reciprocating
• Rotary
• Controlled Volume (Metering)
These subcategories are discussed in more detail in Sections 300, 400, and 500.
Section 140 lists criteria that favor positive-displacement pumps over centrifugals.
Fig. 100-3 Pump Categories Courtesy of the Hydraulic Institute (2 of 2)
124 Oil Field and Special Service Pumps

Sucker rod (“walking beam”) and electric submersible pumps are both very
common and important to producing applications. Extensive Company documenta-
tion for both pump types already exists and is referenced in Section 700 of this
manual.
Several other pump types have been designed to accommodate specific needs.
These are also covered briefly in the Miscellaneous Pumps section and include:
• Air diaphragm pumps
• Regenerative turbine pumps
• Jet pumps (fluid eductors)

• Slurry pumps
Sealless pumps will be covered in this manual at a later date. Other pumps, or fluid-
moving devices, not included in this manual due to their infrequent use include blow
pots or blow cases, pitot or rotating casing pumps, vacuum pumps (which are
usually compressors), and centrifugal-screw pumps. Progressive cavity pumps are
discussed in Section 400, “Rotary Pumps”.
For information on these pumps see the section Other References in this manual or
contact the Vendors of the selected equipment.

Multiphase pumps (MPPs) move liquid, gas, and solids at the same time, through
the same pipeline. They are usually found in the oil fields or offshore platforms
where the fluid from an oil well is a combination of liquid, gas, and solids. A
multiphase pump can pump this combination many miles through one pipeline to a
central gathering station for separation. The multiphase pump saves capital and
operating costs because, without it each well or group of wells would require a sepa-
rator, a pump, a compressor, and two separate pipelines to get the oil and gas to the
central gathering station.
Multiphase pumps are not cheap. Their internal parts have to be specially designed
to handle 100% gas at times. Most multiphase pumps are two-screw (twin-screw)
pumps, but single-screw progressive cavity pumps (PCPs) and helico-axial pumps
are also used. The two-screw type is the same as a liquid two-screw pump with
different screws and mechanical seals. The helico-axial type is a multistage,
specially designed, mixed-flow centrifugal pump. Specially designed, single-screw,
progressive cavity pumps are also used in multiphase service, but because the rotor
contacts the stator, the pumps MUST always operate with some liquid in the pump
to keep the stator from overheating.
Multiphase pumps are also used in artificial lift service. In this application, the
pump is placed down-hole in the well bore.
Multiphase pumps are discussed in more detail in Section 770, Multiphase Pumps.
130 System Hydraulic Design
131 Introduction
This section outlines a method for determining system hydraulics, which can be
used for pump sizing, selection, retrofitting, or troubleshooting. Establishing system
hydraulics is an essential and integral step for all pumps, centrifugal or positive
displacement.
The purpose of determining system hydraulics is to assure that:
• The pump is sized to deliver an adequate flow rate.
• Sufficient driver horsepower is provided.
• The pump develops sufficient discharge pressure to: a) deliver adequate flow
through all downstream pressure drops and elevations, b) allow for adequate
system control at all operating conditions.
• The suction conditions are sufficient to ensure adequate flow into the pump
(cavitation-free service).
THIS PROCEDURE IS GENERAL AND ADDRESSES NEITHER LOCAL
ADMINISTRATIVE NOR APPROVAL REQUIREMENTS. YOU MUST
DETERMINE THESE WITHIN YOUR OWN ORGANIZATION.
132 Overview of the Hydraulic Calculation Sheet, PMP-EF-467

Completing the Hydraulic Calculation Sheet, PMP-EF-467, will determine the
necessary system hydraulics in almost all applications. The form is flexible,
allowing straightforward calculations when the system is well defined (Case I). It
also allows for progressive calculations often needed on large projects (Case II).
See the section “Standard Drawings and Forms” for a copy of the PMP-EF-467
Form. (A sample of PMP-EF-467 is illustrated in Figure 100-4; Figure 100-5 illus-
trates a completed form.) The four parts of this form are described below:
• The Pump Service and Liquid Properties section includes service require-
ments and fluid properties. Space is provided to list maximum and minimum as
well as rated conditions. These limit values are important because they must be
allowed for in the design.
• The System Sketch is helpful in visualizing the system and helps ensure all
significant pieces of equipment are considered.
• The Pressure Calculations section is divided into three steps: Initial Estimate,
Pre-purchase, and Final Review. However, in most cases, only one calcula-
tion is necessary, as explained in cases I and II below.
• The Pressure Profiles section also helps visualize the system and ensure
consideration of important factors regarding system pressures.
133 PMP-EF-467 Instructions

Determine which of the following two cases applies to your situation:
Case I. The hydraulic system is well defined either by design specification or
because the hardware already exists. Go directly to Calculation II (“Pre-purchase”)
and complete it only.
Case II. The hydraulic system is not yet well defined (such as for a large plant
construction project). However, estimates must be made in order to initiate bidding
on long-delivery pumps. Complete Calculations I and II, (and III if appropriate).
PMP-EF-467 Form (Example)
Fig. 100-4
PMP-Completed EF-467 Form
Fig. 100-5
134 Calculation I (Initial Estimate)

This initial calculation is only performed in situations where the hydraulic system is
not yet well defined.
Calculation I is identical to Calculation II (Pre-purchase) except that several design
factors must be estimated. Refer to the detailed instructions in Calculation II for
completing Calculation I, but consider the following in this initial estimate:
1. Vessel, heat-exchangers, furnaces: design pressures and elevations must be esti-
mated. An estimate of the pump’s net positive suction head required (NPSHR)
will help determine the height of the suction vessel.
2. Discharge and suction piping sizes, lengths, and layouts must be estimated
unless a plot plan/system layout is available.
3. If no other estimates or numbers are available, Figure 100-6 shows values to
use for preliminary estimates.
Fig. 100-6 Table of Preliminary Estimates for Calculation
If no other estimates or numbers are available, use the following for preliminary estimates only:
Minimum Pressure
Drop (psi)
A. Reflux to Towers 5
Feed to Columns 10
To plot limits, or other plant connections 10
Fired Reboiler Loop 15
B. Exchangers: If no other information is available, assume a 10 psi pressure drop 10
per exchanger for liquids with a viscosity less than 10 cs, 15 psi with viscosities
between 10 and 25 cs and 25 psi for viscosities above 25 cs. These pressures
are based on fouled exchangers.
With air coolers, assume a 15 psi pressure drop for liquids with viscosities of 15
10 cs or less and 20-50 psi for liquids with viscosities between 10 and 100 cs.
C. Furnaces: Contact Process Engineering or refer to the specific unit’s Vendor 50
data if available.
D. Orifices: A nominal pressure drop of 2 psi (corresponds to 100-inch meter 2
orifice) is used for all orifice meters.
E. Others: For equipment in which pressure losses vary considerably (mixers, filter, 2
centrifuges, etc.), an assumed contingency is included here. Contingency is
usually not less than 2 psi, and only exceeds 10 psi in high pressure, complex
systems.
135 Calculation II (Pre-purchase)

Completing the system hydraulic design is actually a four-step process:
1. Enter the Pump Service and Liquid Properties
2. Complete the System Sketch
3. Complete the Pressure Calculations
4. Establish the Pressure Profile
1. Enter the Pump Service and Liquid Properties

When defining pump service conditions, the entire range of service conditions must
be considered. These alternatives should be shown as a range of values (minimum/
maximum) for specific gravity, viscosity, temperature, vapor pressure, and flow
rates as illustrated in Figure 100-7.
Fig. 100-7 Example: Pump Service and Liquid Properties
In addition to the effect on fluid properties from alternate service conditions, it is

also necessary to consider the effect of abnormal operating conditions: startup, shut-
down, pumpout, cold winter night vs. hot summer day operation, emergency situa-
tions, etc. It is important to develop the following numbers as accurately as possible
in order to achieve reliable operation:
• Corr./Erosion Caused by and Solids. Contaminants, such as water, chlorides,
H2S, and sand may affect overall performance and should be considered.
• Viscosity at Pumping Temperature (PT) (Centistokes). In addition to the
normal pumping viscosity, any long term viscosity changes, such as continuous
pumping in winter months, must be checked. Viscosity changes affect pump
differential head or delta P (DP), efficiency, and driver size.
• Vapor Pressure at PT - psia. The vapor pressure of pumped, boiling liquids is
the absolute pressure at the surface of liquid in the suction vessel. This is a
critical number. Obtain an accurate assessment; consult the Appendix of this
manual, Process Engineering, CRC, a lab for an actual test or a specialist, if
necessary.
• Specific Gravity (SG) at PT. See item B for similar concerns.
• Flow-normal - gpm. If the pump operates over a range of “normal” flows,
examine the flow extremes for piping, pump, and control valve sizing
(e.g., start-of-run/end-of-run).
• “Flow-rated-gpm” Most pumps should be rated for 110% of “normal” flow.
Level-controlled pumps and furnace-charge pumps should be designed to
handle 120% of normal flow. Minimum flow is particularly important for
centrifugal pumps.
2. Complete the System Sketch

This is a schematic drawing of all system components. It shows relative elevations
and any temperature changes that can effect viscosities, specific gravity, and pres-
sure drop. In some pipeline systems, the pump discharge pressure may be dictated
by the highest point in the system and not the pressure or elevation of the receiving
vessel.
Figure 100-8 illustrates a typical system sketch. Note there are three lines for listing
pressure drops of the respective pieces of equipment. These correspond to the three
calculation steps.
Fig. 100-8 Example: System Sketch
3. Complete the Pressure Calculations

Refer to Figure 100-9 under “Suction Pressure” for a sample calculation:
Suction Pressure
• Original Pressure - psia should be the lowest expected pressure under all
normal operating conditions at the point of origin at the surface of the liquid.
Absolute, not gage, pressure (tank, sump, vessel, etc.).
• Static HD (ft. × SG × .433) - psi is the suction static head, measured from
lowest liquid level to center of pump suction flange. If the lowest liquid level is
below the pump centerline, the number is negative.
• - Loss (Line) psi. For NPSH purposes, pump suction line sizing is based on the
rated flow (110 or 120% of normal). Enter the calculated friction loss, in psi.
The total friction loss in suction lines should not exceed 0.5 psi. An evaluation
of NPSHA may allow a greater loss, provided the pump NPSH requirements
are satisfied.
• Pump Suction Pressure - psia is the sum of the above values. It is used to
establish pump differential pressure, NPSHA, and head requirements.
Fig. 100-9 Example: Pressure Calculations
Net Positive Suction Head Available (NPSHA)

A pump’s NPSHA is usually set by the height of the lowest liquid level in the
column, vessel, or tank. With boiling liquids, static head is the only source of
NPSHA. (Elevating the vessel or column, or establishing a low liquid level which
satisfies pump NPSH requirements is preferred over the use of pumps in a pit or
vertical pumps in a suction can.)
The formula for calculating NPSH available (NPSHA) is:
NPSHA = S - F + H - Vp
(Eq. 100-1)
where:
NPSHA is measured in feet of the liquid pumped at the pump suction
flange.
S = Static Head. The vertical distance, in feet, between the surface of
the pumped liquid and the center of a horizontal pump suction
flange. The value is negative (-) when the pump is above the
liquid surface.
F = Line Loss. The friction head in the suction pipe and fittings,
measured in feet of liquid. (See the Fluid Flow Manual.)
H = Original Pressure. The lowest normal operating absolute pres-

sure, measured in feet of liquid, on the surface of the pumped
liquid.
Vp = The pumped fluid’s maximum true Vapor Pressure at pumping
temperature, in feet of liquid.
Pump suction systems should provide enough NPSHA to allow use of an 9000
suction-specific-speed (Nss) pump with at least a 3-foot margin. Refer to the
nomograph (Figure 100-10) to determine the 9000 Nss value for NPSH for the
correct pump flow. (Note: The nomograph assumes the flow rate used is equal
to the pump’s most efficient flow rate). Pump specific speed and NPSHA are
discussed further in Section 210.
• Acceleration Head - Feet. For reciprocating pumps, calculating NPSHA must
include the additional pressure drop that results from pulsating flow.
It is calculated in the same manner as for a centrifugal pump, except in the
NPSH required for a reciprocating pump some additional allowance must be
made for the reciprocating action of the pump termed acceleration head.
This is the head required to accelerate the liquid column on each suction stroke
so that there will be no separation of this column in the pump or suction line.
If this minimum condition is not met, the pump will experience a fluid knock
caused when the liquid column, which has a vapor space between it and the
plunger, overtakes the receding plunger. This knock occurs approximately two-
thirds of the way through the suction stroke. If sufficient acceleration is
provided for the liquid to completely follow the motion of the receding face of
the plunger, this knock will disappear.
Refer to Section 300, Figure 300-10 for estimated acceleration head values.
(These are not as accurate as the calculated values.)
If there is insufficient head to meet minimum acceleration requirements of
NPSH, the pump will experience cavitation resulting in loss of volumetric effi-
ciency; also, serious damage can occur to the plungers, piston, valves, and
packing due to the forces released in collapsing the gas or vapor bubbles.
The head required to accelerate the fluid column is a function of the length of
the suction line, the average velocity in this line, the rotative speed, the type of
pump, and the relative elasticity of the fluid and the pipe, calculated as follows:
LVnC
h a = ----------------
Kg
(Eq. 100-2)
where:
ha = Acceleration head in feet
L = Length of suction line in feet
V = Velocity in suction line in fps
Fig. 100-10 Specific Speed and Suction Specific Speed Chart
gpm × 0.321
V = ------------------------------------------------------------------
2
area of the suction pipe, in
(Eq. 100-3)
where:
n = Pump speed in RPM
C = .200 for duplex single-acting
= .115 for duplex double-acting
= .066 for triplex single or double-acting
= .040 for quintuplex single or double-acting
= .028 for septuplex single or double-acting
= .022 for nonuplex, single or double acting
Note The constant C will vary from these values for unusual ratios of
connecting rod length to crank radius.
K = A factor representing the reciprocal of the fraction of the theoret-
ical acceleration head which must be provided to avoid a notice-
able disturbance in the suction line (K = 2.5 for hot oil, 2.0 most
hydrocarbons, 1.5 amine, glycol, water, 1.4 deareated water,
1.0 urea and liquids with small amounts of entrained gases).
g = Gravitation constant (32.174 ft/sec2)
When the suction system consists of pipes of various sizes, calculate the accel-
eration head for each section separately. Add the acceleration head of all
sections to obtain the total.
If the calculated NPSHA, including acceleration head, is greater than the
suction system can provide, the system NPSH should be increased. This can be
accomplished by:
– Increasing the static head
– Adding a booster pump to the system
– Adding pulsation damper
A pulsation damper properly installed with a short, full-size connection to the
pump or suction pipe can absorb the cyclical flow variation and reduce the
pressure fluctuation in the suction pipe to that corresponding to a length of 5 to
15 pipe diameters, if kept properly charged.
There is a similar pressure fluctuation on the discharge side of every power
pump, but it cannot be analyzed as readily because of the greater influence of
liquid and piping elasticity and the smaller diameter and much greater length of
the discharge line in most applications. However, a pulsation damper can be
just as effective in absorbing the flow variation on the discharge side of the
pump as on the suction side, and should be used if pressure-fluctuation and
piping vibration is a problem.
Example: Given a 2" × 5" triplex pump running at 360 RPM and displacing
73 gpm of water with a suction pipe made up of 4 feet of 4-inch pipe and
20 feet of 6-inch pipe:
Average velocity in 4-inch pipe
0.321 × 73
V 4 = ------------------------- = 1.84 fps
12.73
(Eq. 100-4)
Average velocity in 6-inch pipe
0.321 × 73
V 6 = ------------------------- = 0.811 fps
28.89
(Eq. 100-5)
Acceleration head in 4-inch pipe
4 × 1.84 × 360 × 0.066 = 3.88ft

h a4 = ------------------------------------------------------
1.4 × 32.2
(Eq. 100-6)
Acceleration head in 6-inch pipe
20 × 0.811 × 360 × 0.066

h a6 = ------------------------------------------------------------ = 8.55ft
1.4 × 32.2
(Eq. 100-7)
Total acceleration head
h a = 3.88 + 8.55 = 12.4ft

(Eq. 100-8)
Discharge Pressure
• Delivery Pressure - psia is the absolute pressure in the receiving vessel mani-
fold or well at the desired destination of the fluid.
• Static Head (ft. × S.G. × .433) - psi is the difference in elevation between
pump discharge and highest point in the discharge system, in psi.
• Line Loss psi. Although pumps are specified to deliver rated flows, discharge
piping and equipment pressure losses should be based on normal flow require-
ments. If the calculated pressure drop is less than 5 psi, use 5 psi.
• Add up the losses for Exchangers, Furnaces, and Orifices and enter the sum
on the appropriate line.
• Enter any Contingency on the next line. Overly conservative contingencies
will waste energy and may cause unreliable operation.
• ΔP Control Valves. If the system’s flowrate is regulated by throttling a centri-
fugal pump’s discharge through a control valve, the pressure drop across the
valve must be estimated. This is important for two reasons: 1) the pump must
be designed with enough discharge pressure to allow the control valve to have
the pressure drop it needs to control flow; 2) allowing too much pressure drop
for the control valve will waste energy as the pump puts up more pressure than
needed, only to be let down (with no energy recovery) across the valve.
For background information, see control valve sizing and selection in the
Instrumentation and Control Manual. System curves (discussed in Method B)
are covered in Section 210 of this manual. The Fluid Flow Manual has infor-
mation on calculating pressure losses from flow. Two methods are presented,
and are repeated here in Figure 100-11 and Figure 100-12.
• Pump Discharge Pressure - psia is the sum of all discharge system pressure
losses at normal pump flow, plus static head and delivery pressure.
Fig. 100-11 Method A for Calculating Flow Pressure Losses
Method A - Quick Estimation

1. Add all the system pressure losses due to flow through lines, ΔP exchangers, ΔP furnaces, etc.
(Include suction as well as discharge losses). These are “dynamic losses.”
2. Using the chart below, enter the bottom scale at the sum of these dynamic losses, read up to the
line, then left to the scale marked “% of Dynamic Losses.”
The left scale is a percentage of the dynamic losses to allow for the control valve. For example:
calculated total pressure loss due to flow is 139 psi. Using the chart, the added percentage is
approximately 33%. Add 46 psi (.33 × 139=46) for “Control Valve ΔP.”
(This chart is a general, empirical chart for estimating common installations. It does not ensure
adequate pressure loss for all control valves, nor that energy losses will be minimized. In most
cases, it will be conservative (i.e. adequate pressure loss).)
Control Valve Pressure Drop Selection
Fig. 100-12 Method B for Calculating Flow Pressure Losses (1 of 2)
Method B - Detailed Approach

Control valves normally need a wide range of control in order to accommodate varying conditions in
the hydraulic system. For example, lowering a plant’s feed rate by 50% can lower the system’s pres-
sure losses by 300% or more. Other common changes include pumping through alternate mani-
folding, pumping from/to tanks at various elevations, pumping into higher-pressure systems due to
fouling in lines, reactors, etc. All these situations affect the ΔP across the control valve.
Method B accounts for these variations and helps ensure adequate control in most situations. At the
same time, it will help avoid excess energy waste across the control valve. It will not be accurate
where there are large variations in the suction static pressure. Consult a specialist in this situation.
Note that a pump curve must be available or estimated for this method. Vendors are the most likely
source of curves for proposed pumps.
1. Define the system curve which has the maximum pressure losses from flow (highest flow rate,
worst possible piping arrangement) combined with the highest static pressure on the pump
discharge. Calculate a system curve for this situation over the expected range of flow rates. (See
Section 210 and the Fluid Flow Manual, if necessary.)
2. Similarly, define the system curve with minimum pressure losses and the lowest static pressure
on pump discharge. Calculate a system curve for that situation.
3. Define the situation where the pump will experience the lowest suction pressure.
4. Construct a graph with the following three curves:
a. Plot the pump discharge pressure (not differential head) by adding suction pressure to
differential head. Use the lowest suction pressure from step 3. This is the plot of discharge
pressure shown below.
Combined System/Pump Curves
Fig. 100-12 Method B for Calculating Flow Pressure Losses (2 of 2)
b. Superimpose the two system curves from steps 1 and 2 as shown in

c. Define the maximum and minimum flows that the control valve would be expected to
handle. Draw vertical lines from the maximum- and minimum-flow situations as shown. F1
(min.) is the minimum flow, F2 (max.) is the maximum flow.
d. ΔP1 is the maximum pressure drop that would ever have to occur at the control valve. ΔP2
is the minimum ever across the control valve. If the system curve is above the pump curve
at F2 (max.), then the pump curve must be raised.
5. Check that the values from the graph (ΔP1, ΔP2, F1, F2) satisfy the following equation:
(F2/F1) × (ΔP1/ΔP2)1/2 ≤ 10
(Eq. 100-9)
If the pump curve satisfies this equation, then the necessary range of control will be achieved,
with minimum pressure losses, across a normal control valve.
If the equation is not satisfied, then the pump must be specified with more differential head in
order to raise the discharge pressure plotted on the graph. Raise the discharge pressure until the
equation is satisfied.
You should consult a specialist if:
– you cannot satisfy the equation with a reasonable discharge pressure;
– the flow-rate turndown exceeds 5:1; or
– the system is unusually complex.
Differential Pressure
• Total Pump Diff. Press. - psi is the difference between pump discharge and
suction pressures.
• Pump Head (psi × 2.31/S.G.) - Feet is total pump differential pressure,
converted to feet.
Horsepower
• Hydraulic Horsepower (gpm × psi /1715) is the pump power requirement at
“Rated Flow” and “Total Pump Diff. Press” assuming 100% efficiency.
• Efficiency %. Use an estimate initially. If necessary, consult a machinery
specialist or Vendor for assistance. Efficiency could range from 20 to 80%
depending on the selection and application.
Once pump quotes are received and a pump selected, the actual efficiency
should be checked for driver sizing.
4. Establish Pressure Profiles

Pressure profiles graphically describe the pressures in all components of the system.
A profile visually tracks pressure changes through the system for various operating
situations. Although in simple situations profiles may not be necessary, they are
valuable in complex or high-pressure systems where pump head determines the
design pressure of equipment and relief valve settings.
Note Since Calculation I profiles are often estimates, Calculation II and III
profiles may need to be adjusted. Confirm design pressures against system pres-
sures and relief valve settings.
The pressure profiles illustrated in Figure 100-13 are based on the following:
a. Maximum suction pressure equals the PSV setting of the suction vessel,
plus static head.
b. Pump shutoff pressure, based on maximum suction pressure.
c. The pressure profile, plotted through equipment and lines based on normal
and maximum suction pressures at rated differential pressure and
normal flow.
Fig. 100-13 Pressure Profiles
d. Downstream equipment design pressures, which determine minimum

design pressures. In the example, the design pressure is 10% or 20 psi
(whichever is greater) over system operating pressure.
Always check actual pump shutoff head against the design pressures of all
system components to help avoid equipment overpressure or undesirable
relief valve operation. This most commonly occurs in an installation of
high pressure multi-stage pumps.
136 Calculation III (Final Review)

The final calculation is the same as Calculation II. This is an opportunity to refine
the hydraulic design by substituting available information for any previous esti-
mates, which may be common on large projects where information becomes final-
ized in the latter design stages. Likewise, the “Final Review” column can also be
used to adjust for any design changes and contingent operation.
140 Application and Selection Criteria
141 Introduction
In specifying a pump installation, one of the first decisions is to determine the pump
category to use. As defined in this manual, “category” can be either centrifugal,
reciprocating, rotary, metering, sealless, or miscellaneous. This section discusses the
features, advantages, and disadvantages of each category. Once a category is
selected, refer to the respective section of the manual for information on selecting
the type of pump in that category.
142 Pump Selection Fundamentals

Note In addition to the following text, see Figures 100-14 through 100-18.
During the selection process, centrifugal pumps should always be considered first.
They are preferred because of their lower cost, simplicity, reliability, and smooth-
flow characteristics.
When selecting a pump for a given service, conflicting factors may make the choice
unclear. For example, if a pumped fluid is both highly viscous and contains abra-
sives, a rotary pump may be first choice for the viscous stock. However, depending
on the rotary pump chosen, it might be a poor choice.
The Pump Selection Guide in Figure 100-14 and the Head/Capacity Selection
Curve, shown in Figure 100-15, give additional information on choosing between
centrifugal, reciprocating, and rotary pump categories.
The Head/Capacity Selection Curve shows the category of pump most suitable from
a head/capacity standpoint. The Pump Selection Guide is then used to identify the
other factors affecting the selection. Completing the Local Priority column on the
Pump Selection Guide weighs factors by their importance in any given situation.
The user (usually an operating representative) should always be consulted
during pump selection, especially when the choice between centrifugal and posi-
tive displacement pumps is difficult. In addition to preference based on past perfor-
mance, the user may also wish to consider local maintenance and the convenient
availability of interchangeable parts.
Fig. 100-14 Pump Selection Guide
100-24

LOCAL PRIORITY RECIPROCATING CENTRIFUGAL ROTARY COMMENTS
SELECTION CRITERIA CRITICAL IMPORT NOT IMP GOOD FAIR POOR GOOD FAIR POOR GOOD FAIR POOR
FLOWRATE VS
---- SEE HEAD/CAPACITY SELECTION CURVE ----
DEVELOPED HEAD
AVOIDS EMULSIFYING
THE FLUID X X X
© 1999–2006 Chevron USA Inc. All rights reserved.
RELIABILITY X X X
SELF PRIMING CENTRIFUGAL DESIGNS
SELF PRIMING X X X ARE AVAILABLE
HANDLING "PROGRESSIVE CAVITY" ROTARY PUMPS
ABRASIVES X X X HANDLE ABRASIVES
HANDLING NPSHA ON RECIP'S MUST INCLUDE ALLOWANCE
LOW NPSHA X X X FOR "ACCELERATION HEAD"
HANDLING
ENTRAINED GAS X X X
PUMP VISCOUS
STOCK X X X
PUMP LOW
VISCOSITY STOCK X X X
CENTRIFUGAL EFFICIENCY VARIES WITH TYPE OF
ENERGY EFFICIENCY X X X PUMP, AND FLOW RATE RELATTING TO BEP
INSTALLED COST X X X
MAINTENANCE COST X X X
CONTROLS LEAKAGE (IF (IF AIR QUALITY PERMITS MAY DICTATE PUMP
TO ATMOSPHERE X X SEALED) X SEALED) SELECTION
TEMPERATURES VARIES WITH APPLICATION. REVIEW EACH
ABOVE 350° SPECIFIC APPLICATION
MANY CLIENTS HAVE STRON PREFERENCES
CLIENT PREFERENCES
BASED ON OPERATING EXPERIENCE
COMPATIBILITY WITH CONSIDER MAINTENANCE FACILITIES, PARTS, AND
EXISTING SYSTEM OPERATOR FAMILIARITY
VERY IMPORTANT IN OFFSHORE OR OTHER
WEIGHT AND SPACE
November 2006
LIMITED-SPACE APPLICATIONS. REVIEW
Pump Manual
EACH SPECIFIC APPLICATION.
Fig. 100-15 Head/Capacity Selection Curve
143 Abrasives
Abrasives can destroy any pump in a short time if the pump is not designed or the
pump type is not selected properly. Wear associated with abrasive particles is a
function of the velocity of the pumped fluid inside the pump to the "nth" power
where "n" ranges from 2.5 to 4 based on the hardness and smoothness of the partic-
ulate. Since centrifugal pumps operate on the principle of adding kinetic energy to a
flow stream, they rely on increasing the velocity of the stream inside the pump. This
velocity is then converted to potential energy at the discharge of the pump. Centrif-
ugal pumps can sometimes handle abrasive service if the pump is specially designed
as follows:
a. Run slowly, at 1200 RPM or less.
b. Have an impeller and case of hardened, hard coated, or hard materials.
c. Operate near its best efficiency point to minimize internal turbulence and
vibration.
d. Have pump-out veins on the pump’s impeller instead of balance holes.
e. Have a wear plate and wear rings.
f. Flush the mechanical seals with a clean, external liquid stream.
g. Have a case with a double volute.
Progressive cavity pumps, a type of rotary pump, are often used, but require
frequent liner replacements (see Section 400 of this manual). Twin-screw pumps are
often a good solution to pumping abrasives. In these pumps very little wear will
occur, even at high concentrations of abrasives, if the stock being pumped is highly
viscous. As with centrifugal pumps in abrasive service, twin-screw pumps should be
designed to run at slow speeds with all wetted surfaces fabricated of materials
harder than the abrasives.

Centrifugal pumps are selected more often than any other pump type. There are
many reasons why centrifugal pumps are the best choice. Some of these include:
• Low total cost of ownership (first cost, operating and maintenance cost)
• Good reliability
• Capable of a wide range of flow rates from 10 GPM to over 100,000 GPM
• Capable of developing heads from 20 feet to over 10,000 feet
• Able to operate with relatively low NPSH available
• Wide temperature and pressure capabilities
• Minimal environmental impact with proper mechanical seal designs
• Flow variation flexibility
• Easily controlled
• Available from many manufacturers
• Available in many different materials of construction

• Wide range of designs and types
• Low levels of vibration imposed on piping
• Low discharge pressure pulsations
• Capable of handling suspended solids when designed for this service
Poor Applications for Centrifugal Pumps

• When a self-priming pump is needed. Although centrifugal pumps are avail-
able in a self-priming design, the available sizes are limited and performance
penalties are substantial. Conventional centrifugal pumps must be primed and
supplied with adequate NPSH for proper operation.
• Low Net Positive Suction Head Available (NPSHA). Positive displacement
pumps are self-priming and can handle more gas than centrifugal pumps.
Because of this, people often think positive displacement pumps can also
tolerate a lower NPSHA. This depends on the type of positive displacement
pump. Rotary pumps are often a better choice for low NPSHA applications,
than centrifugals; however this is not true for reciprocating pumps where one
must subtract the Acceleration Head in the NPSHA calculation (see
Section 334 for the formula). The Acceleration Head is not part of the NPSHA
equation for centrifugal or rotary pumps.
• Low speed drivers. A centrifugal pump for normal oil service usually operates
at nominal speeds of 1800 or 3600 RPM. If the pump is driven by an internal
combustion engine operating at 600 RPM, a speed-increaser will be necessary.
A speed-increasing gearbox can easily cost more than the pump. In these cases,
a positive displacement pump may prove more economical.
• High efficiency requirements over a range of pressure conditions. This situ-
ation frequently occurs in pipeline transportation. A positive displacement
pump at any capacity remains reasonably efficient at pressures between 25%
and 100% of its rating. A centrifugal pump will not remain efficient in these
circumstances.
• Low flow, moderate to high head applications. Operating centrifugal pumps
far to the left of their best efficiency flow causes many problems. For example,
impeller eye recirculation and higher levels of vibration result when operating
below the recommended minimum flow. The effects of low flow operation are
shaft breakage, seal and bearing failures.
The point of “minimum flow” will vary for every pump. Some of the factors
which affect the minimum flow are fluid density, NPSHA, impeller design
suction specific speed, casing volute design, piping and control systems
designs.
If a centrifugal pump must be selected for a low flow application, refer to
Section 270, “Maintaining Centrifugal Pump Flow Rates Close to the Best Effi-
ciency Point (BEP) or Best Efficiency Flow Rate”, for pump protection alterna-
tives.
Sundyne vertical in-line pumps are designed for low-flow, moderate to high
head applications; however, these pumps will also have a minimum flow which
should be maintained for reliable service.
• Handling fluids with entrained gas. Centrifugal pumps can lose suction due
to an excess amount of gas accumulation in the impeller eye. The vapor or gas
accumulation can be caused by excessive cavitation, recirculation, or entrained
gas. A good rule of thumb for maximum entrained gas is 5% by volume.
The centrifugal pump impeller is a good centrifuge. The heavy material (liquid)
is expelled through the impeller while the light fluid (gas) will collect in the eye
of the impeller. If the gas or vapor volume fills the impeller eye, the pump may
lose suction.
A worst-case example of this would be a centrifugal pump selected for an
application with little NPSHA, some entrained gas, and operated at reduced
flow rates. At reduced flows, the fluid velocity in the suction pipe may not be
capable of pushing the entrained gas or vapor, through the impeller.
• Pumping viscous fluids. Although centrifugal pumps are capable of pumping
fluids with a viscosity of 4000 SSU and higher, the performance penalties are
substantial. As shown in Figure 200-18, pump efficiency can be drastically
reduced. The pump’s capacity and head capability are also reduced. For viscosi-
ties above 65 centistokes or 300 SSU, rotary or reciprocating pumps are recom-
mended.
• When fluid emulsification must be avoided. Centrifugal pumps are good
agitators and mixers. Rotary or reciprocating pumps are a better choice to mini-
mize fluid emulsification.

Reciprocating pumps are used most frequently for lower-capacity, high-pressure
services. The initial cost of small reciprocating pumps is competitive with centrif-
ugal pumps. However, larger reciprocating pumps (particularly over 200 gpm
capacity) are usually more expensive (initially and to maintain) than other pump
types. High-speed centrifugal pumps should not be overlooked when low-capacity
high-head services are involved.
When service requirements permit using either a centrifugal pump or a recipro-
cating pump, carefully consider both operating and maintenance costs. For most
services, the operating costs of motor-driven or turbine-driven centrifugal pumps are
less than the costs for reciprocating pumps. Maintenance costs usually exceed those
of a centrifugal pump because of the many moving parts, including valves and
sliding contacts.
Pulsating flow may limit the use of reciprocating pumps. However, pulsating flow is
usually not the decisive factor for determining if reciprocating pumps are the best
selection. The effects of pulsation can be minimized but not eliminated by using
pulsation dampers. Pulsating flow may cause problems in the application of auto-
matic control flow measurement or process. See Section 300 for more details about
reciprocating pumps.
Good Applications for Reciprocating Pumps

• Low to moderate capacity with relatively high differential pressure.
Figure 100-15 shows the head/capacity range for which reciprocating pumps
are normally considered in nonviscous services. The division shown is intended
as a guideline only.
• Relatively high viscosity. The efficiency of a centrifugal pump drops rapidly
with increasing viscosity. The most economical applications of these pumps are
normally limited to viscosities under about 500 SSU. Reciprocating pumps can
efficiently handle stocks up to about 8000 SSU.
• Higher viscosity oils can be delivered by reciprocating pumps operating at
slower speeds, but such applications usually are not economical. Rotary pumps
are more appropriate.
• Relatively constant capacity with widely varying discharge pressures.
Reciprocating pumps are particularly suited to this application.
• Highly variable capacity with either constant or varying discharge pres-
sure. Direct-acting, gas-driven pumps are well suited to this application
because the speed is easily controlled by the gas driver.
• Where a self-priming pump is needed.
Poor Applications for Reciprocating Pumps

• When pulsating flow is undesirable. An example of such a service is a fuel oil
feed to boilers. Ordinarily, rotary pumps are preferred for this service because
of their smooth discharge pressure and better efficiency at higher viscosities.
• Medium capacity and medium differential pressure with low viscosity.
Centrifugal pumps ordinarily are more economical. Examples of such services
are water or hydrocarbons pumping about 100 gpm with differential pressures
up to about 700 feet or 300 psi.
• High capacity. For capacities above about 200 gpm, reciprocating pumps are
seldom the best selection, regardless of the discharge pressure. In this range,
reciprocating pumps become so large that they are more expensive than centrif-
ugal pumps. An exception may be high-pressure water injection services.
• Minimum packing leakage required. Some hazardous or toxic services, such
as phenol or HF acid, require absolute minimum stock leakage. Reciprocating
pumps are subject to packing leaks and in such services must be fitted with a
double stuffing box to provide an enclosed leakage disposal system. Centrif-
ugal pumps with mechanical seals would be preferred.
146 Rotary Pumps

The most common rotary pump types are gear, multiple screw, and single screw.
Cam-and-piston and sliding vane pumps can be considered for special services.
The following discussion applies to all rotary pumps (also see Figures 100-16 and
100-17). See Section 400 for more details.
Fig. 100-16 Positive Displacement Pump Selection and Application

Maximum Maximum
Flow gpm Pressure psi Application
1. Plunger 300 6000 Used primarily as glycol pumps, steam-
generator feed pumps, condensate pumps,
drilling mud pumps, and additive injection
pumps, especially where solids are
present.
2. Piston 800 1500 Same as plunger, but without solids
present.
3. Diaphragm 10 1000 Used mainly in controlled volume applica-
tions as a metering pump or where low flow
rates and high solids concentrations are
present. Temperatures limited to less than
500°F maximum.
4. Cam-and- Vacuum services.
Piston
5. Rotary Gear 150 700 Used to transfer recovered oil from a drain
separator to a process oil/ water separator.
Also used as a diesel transfer pump and in
high viscosity applications, and for circu-
lating lubrication oil services for large
machines and other clean, high and low-
pressure services.
6. Single Screw 450 200 Used for high viscosity stocks (non-neuto-
or Progres- nian), stocks with up to 30% (entrained
sive Cavity gas, and often for abrasive services. Not
good for temperatures over 300°F and in
carbon dioxide entrained gas services.
7. Three-Screw 1000 4500 Used for low temperature, high viscosity
stocks (lower then progressive cavity)
without abrasives and circulating lubrica-
tion oil for large machines.
8. Two orTwin- 8000 4500 Used for low and high temperature liquids
Screw with and without abrasives (up to 600°F),
and as a multiphase pump (90% plus gas,
the rest liquid) with and without abrasives
with special design.
9. Sliding Vane Vacuum services.
(1) For reciprocating pumps, pumps 1,2 and 3 above. The recommended RPM is a function of
stroke length (inches). See API Standard 674.
(2) For reciprocating pumps, pumps 1,2, and 3 above, handling liquids with viscosities of 300
SSU at pumping temperature, the speeds are normally reduced to a percent of the basic
speed. See API Standard 674.
Fig. 100-17 Centrifugal Pump Selection and Application

Typical Capacity Typical Head Common
Range gpm Range Feet Applications
1. Vertical Inline (single stage) 20 to 1200 15 to 600 General processing and transfer service at
temperatures below 350°F. Minimum space
application.
2. End Suction, Frame Mounted 35 to 4,000 30 to 700 General processing and transfer service at
ANSI temperatures below 350°F.
3. Horizontal, Single-Stage, 100 to 20,000 40 to 900 Hydrocarbons in the low to moderate flow
Horizontal-suction Between and moderate head ranges. Cooling tower
Bearings API water circulation.
4. Horizontal, Multi-Stage Axially 200 to 1500 200 to 4500 Used to pump crude oil, high-pressure boiler
split feed water, sea water, gasoline, and other
hydrocarbons and also in waterflood opera-
tions.
5. Horizontal Barrel (Double Case) 200 to 1700 up to 9000 Used principally for process plant high pres-
sure reactor charge and water-flood applica-
tions. Pump speed may approach 7500
RPM.
6. Radially Split Vertical Can 20 to 2000 55 to 2000 Used principally for improving NPSHA when
pumping bubble point hydrocarbon mixtures.
Pumping end usually encased in a pressure
vessel (can).
7. Vertical Turbine Lineshaft 100 to 30,000 10 to 1500 Used for lift application such as sea water
and fire water.
8. Electric Submersible 100 to 30,000 10 to 2000 Used to eliminate long shaft lengths. Same
application as vertical turbines, and crude oil
production.
Good Applications for Rotary Pumps

• Viscous stocks. Rotary pumps can deliver high-viscosity stocks with a smaller
reduction in efficiency than other pump types. Rotary pumps can handle stocks
with viscosities varying from LPG (not recommended for continuous duty) to
very viscous greases. For stock more viscous than about 10,000 SSU, rotary
pumps are usually the most economical selection.
Under special conditions, centrifugal pumps can deliver viscosities up to about
5000 SSU, but their efficiency above 500 SSU is so poor that such applications
are not usually economical. Special reciprocating pumps operating at greatly
reduced speeds can handle viscosities as high as those handled by rotary
pumps, but these pumps must be so large, they become prohibitively expensive.
• Lubricating and hydraulic oils. Rotary pumps are most commonly used to
circulate lubricating oil in mechanical equipment or to provide pressure for
hydraulic operating systems. The oil used in these systems is usually cleaned by
filtering. The pumped stock lubricates the pump’s internal gears and bearings.
• Self-priming. Rotary pumps work well when services require self-priming in
moderate capacities, such as barrel, small tank, and sump unloading. However,
when these services involve high capacities and low-viscosity fluids, vertical
centrifugal pumps are usually used.
• Vacuum services. Rotary pumps, lubricated by special oils, are often used in
vacuum services to pump air or other gases and vapors. Low vapor pressure
oils are used to lubricate the pumps and seal the clearance spaces. Oil separa-
tors on the discharge of such pumps remove the oil from the gas. Cam-and-
piston-type and sliding-vane-type rotary pumps are frequently used in this
manner. Pressures as low as 2 × 10-4 mm mercury absolute are attainable with
the cam-and-piston-type.
With either the oil or water seal, the vacuum obtainable is limited by the vapor
pressure of the sealing liquid.
• Intermittent low-capacity services. Small internal-bearing rotary pumps can
sometimes be used economically in intermittent services where rotary pumps
might seem to be unsuited. In such cases, it is cheaper to periodically replace
inexpensive pumps than it is to buy pumps not subject to the same rate of wear.
An example of this application is pumping out small tanks or vessels where a
certain amount of scale and grit is expected. Another example is intermittent
handling of non-lubricating stocks, such as LPG and gasoline at small
marketing plants or on tank trucks.
• Non-pulsating flow. Hydraulic operating systems and fuel oil systems usually
require non-pulsating flow. Rotary pumps work well in these services, espe-
cially when high viscosity renders centrifugal pumps uneconomical.
• Handling “wet” oil. “Wet” oils (over 3% water by volume) should be used
only with pumps that will not cause the water and oil to emulsify. Rotary
pumps operated at slow speed (300 to 400 RPM) work well in such services.
Poor Applications for Rotary Pumps

• Non-lubricating stocks in continuous service. Internal parts of rotary pumps
must be adequately lubricated. Stocks with poor lubricating qualities, such as
LPG, gasoline, and water are not usually satisfactory for rotary pumps in
continuous service.
• High differential pressures and large capacity. Rotor deflection usually
limits the differential pressure produced by a rotary pump. For standard
designs, the larger the pump, the lower the maximum allowable differential
pressure.
• Medium-capacity and medium-head services. Except for high viscosities,
medium-capacity and medium-head services usually can be handled more
economically by centrifugal or reciprocating pumps than by rotary pumps.
Because of the rotary pump’s close clearances and the possibility of mechan-
ical damage, reciprocating and centrifugal pumps are usually recommended
unless a rotary pump promises significant savings.
• Abrasive material or possibility of running dry. Rotary pumps are ordinarily

not recommended for stocks containing appreciable quantities of abrasive mate-
rial. Often, however, rotary pumps can be used successfully for abrasives. See
Section 143. In any case, rotary pumps should never be allowed to run dry.
• Corrosive liquids. Rotary pumps do not handle corrosive liquids well. Pumps
constructed of brass and bronze are used occasionally. Stainless-steel rotary
pumps are not practical because of the possibility of galling or seizure.

Figure 100-18 lists pumps for special applications that do not fit into the typical
centrifugal or positive displacement categories.
Fig. 100-18 Miscellaneous Pumps

Pump Type Characteristics
Axial Flow Also called propeller pump. Used where large capacity and low head
are required. Generally with a vertical configuration for lifting waste-
water, effluent, etc.
Disc friction Also called regenerative turbine pumps. Similar to centrifugal pumps
except liquid is pressured by recirculation in the impeller vanes. A
low-capacity, moderate to high head pump that can handle large
amounts of gas or vapor. Pump efficiency is greatly affected by
internal clearances. These pumps are usually unsatisfactory where
abrasives are present. (See Section 700.)
Metering Small reciprocating plunger or diaphragm pumps used for accurate
pumping of chemicals and additives. Pumping rates are normally
measured in gallons per hour and are adjustable from zero to full
pump rate. Capable of high discharge pressures. (See Section 500.)
Diaphragm An air-operated, versatile, utility pump normally using compressed air
as the driving fluid. Useful in handling hazardous or abrasive mate-
rials, and in explosive environments. Smaller units are occasionally
used in metering service. (See Section 700.)
Jet Also called eductors or ejectors. Jet pumps have no moving parts
and use the venturi action of high velocity fluids through a nozzle to
create suction. The driving fluid and pumped fluid are mixed at the
discharge. Typical applications are moving granulated solids with
water, deep-well water pumping, and shipboard bilge pumping with
water. (See Section 700.)
Archimedes Screw Used in lifting effluent and waste water at relatively low flows where
agitation and mixing are undesirable. Limited to lifts of approximately
25 feet. Similar to a screw conveyor.
Peristaltic Also called hosepumps. Peristaltic pumps are used for pumping fluids
such as waste sludges, lime and cement mortar, adhesives and
shear sensitive fluids such as latex paints. In the petrochemical
industries use is limited to shear sensitive services. The pumps have
few moving parts, no seals, and can be run dry. Life is limited due to
the life of the elastomer hose.
150 Considerations for H2S Services

Pumps represent a potential leak source. They should be avoided as much as
possible where there are high concentrations of H2S. If pumps must be used,
centrifugal-types are more reliable and provide a better seal than reciprocating-
types.
When a centrifugal pump is required, select a high quality pump that has a low shaft
deflection. API 610 and Specification PMP-MS-983 are recommended design speci-
fications. The pump case should be constructed of steel or higher alloy, not cast iron
or bronze. Wherever possible, use a mechanical seal and inject a clean, non-H2S-
bearing fluid as a seal flush. The source of the flush should be reliable and available
whenever the pump is in operation. See Section 850 Mechanical Seals for further
information.
Reciprocating pumps are generally not recommended because of the number of
stuffing boxes on multiplex pumps, the greater difficulty of sealing reciprocating
motion, and packing is more difficult to positively seal than a mechanical seal on a
centrifugal pump.
When using a reciprocating pump, use steel material on the liquid end, including
steel rods. Provide a non-H2S-bearing flush to the lantern ring. Reciprocating
pumps that meet the requirements of API 674 are recommended, if a reciprocating
pump is selected.
Many copper alloys have poor resistance to H2S. Use them with care if it is likely
that they will make wetted contact with an H2S-containing process stream. Accept-
able copper-containing items include sleeve bearings, thrust bearings, and mechan-
ical-seal throttle bushings.
Avoid the use of alloys stronger than 90 ksi yield strength or harder than 235 BHN
(equivalent to about Rockwell C22). Exceptions to this hardness limit include
components such as reciprocating-compressor valves and piston rods where greater
hardness is required for reasonable service life. Materials with high yield strengths,
or hard materials, are very susceptible to stress-corrosion cracking.
160 Suction Piping Design Considerations
161 NPSHA and Fluid Velocity

NPSHA, Piping Symmetry and Parallel Operation
Suction piping should be as short and direct as possible without “pockets” or “high
spots” which could trap air or gas.
Suction piping should be designed to satisfy the NPSH requirements assuming
parallel operation of installed pumps.
Most pump installations include two 100% pumps, one operating and one spare. It is
a common operating practice to test run spare pumps at least once per month or
share operating duty between installed pumps on a monthly basis.
During monthly spare pump operation or switching between pumps, both pumps
will operate in parallel. With both pumps operating, flow through the suction piping
system may increase. An increase in flow in the suction line will increase the fric-
tion loss and reduce the NPSHA to both pumps.
Symmetrical suction piping (size and layout) will ensure equal pressure drop and
minimize operational problems caused by inadequate NPSHA to one or both pumps.
Refer to Section 225, “Suction-Stealing”.
Suction Line Fluid Velocity for Reciprocating Pumps

Suction lines should be sized for velocities below three feet per second. Refer to
Section 333 for more discussion of line sizing. The calculation of friction losses in
the suction line to reciprocating pumps must include “acceleration head” losses.
Refer to Section 135 for acceleration head equation and discussion.
Suction Line Fluid Velocity for Centrifugal Pumps

Generally, suction pipe velocities should not exceed the value which exists at the
pump suction flange. For most centrifugal pump designs, suction flange velocity
will vary between 8 and 15 feet per second. Suction line sizing is usually at least
one pipe size larger than the pump suction size to minimize friction losses and
maximize the NPSHA.
Boiler Feedwater Application

Boiler feed water pump suction piping design may be an exception to the normal
suction line velocity guidelines presented above. Under stable operating conditions,
the water in the deaerator is at saturation conditions. Unstable or transient condi-
tions in the deaerator can cause operational problems with the pumps including
“steam-binding.” Unstable conditions can occur when the condensate feed to the
deaerator is cooler than normal due to a change in the utility load conditions. The
pressure in the deaerator may drop slightly due to the cooling effects of the
returning condensate, while the water in the suction pipe may be at the original
higher temperature. Under these unstable conditions, the NPSHA to the feed pumps
will decrease and the water in the suction line may boil.
Reducing the volume of water in the suction pipe will minimize the problems
caused by unstable conditions in the deaerator. This reduces the temperature differ-
ence between the water in the suction pipe and the deaerator. “Reduced volume”
may be accomplished by using a smaller diameter suction pipe and separate suction
lines to each pump. In either case, fluid velocity in the suction line may be higher
than in most installations.
162 Adequate Venting

Suction Pipe Slope
Fluid velocity in suction lines is normally low so trapped air or gas may not be
swept out by the fluid flow. Selecting the proper pipe layout is critical to minimize
gas and vapor collection.
If the pump must lift the liquid from a lower level, the suction piping should be
designed with a continuously rising slope toward the pump suction flange. If the
liquid level in the suction vessel is above the pump suction, the pump suction piping
should slope continuously downward to the pump suction flange.
Trapped air or gas in the suction piping effectively reduces the diameter of the pipe,
increases suction line friction losses, and reduces NPSHA to the pump.
Pump Inlet Reducer Selection

Pumps with horizontal or side-suction flange orientation that require a reducer
should use an eccentric reducer with the flat side on the top.
End-suction pumps are the most common and require additional reducer selection
cautions because of the close proximity of the impeller inlet to the suction flange.
Reducers for end-suction pumps should be limited to one change in diameter of one
pipe size; however, more than one reducer can be installed in series. If the selected
reducer has a more abrupt reduction than one pipe size, then a run of straight pipe
between the reducer and the pump suction of at least 5 diameters is recommended.
End-suction pump installations where the suction pipe comes down from overhead
should use a reducing elbow or a combination of a long-radius elbow and a concen-
tric reducer.
163 Fluid Velocity Profile Entering Pump Suction

Providing a uniform velocity profile for fluid entering a pump should be the goal of
the inlet piping designer. A velocity profile is a snapshot of the fluid velocity
through a cross-section of the suction pipe perpendicular to flow direction. Uniform
velocity would indicate equal fluid velocity at all points in the cross-sectional plane.
For example, the velocity profile of fluid flowing through an elbow would typically
have higher fluid velocity at the outside diameter of the elbow and lower velocity
near the inside diameter (see Figure 100-19). The flow exiting the elbow would not
be a uniform velocity profile. Other piping components such as isolation block
valves and check valves can also disturb the fluid velocity profile at the inlet to the
pump.
Fig. 100-19 Uniform Velocity Profile Flowing into Elbow
The detrimental effect of a poor inlet velocity profile can be minimized by using
straightening vanes or length of straight pipe just upstream of the pump suction.
Other devices, such as Cheng Rotation Vanes, are also available and can be
specially designed to improve an inlet velocity profile. The use of straightening
vanes and lengths of straight pipe are most common and are explained in detail later
in this section.
Elbow Orientation for Double Suction Pumps

Single-stage, double-suction centrifugal pumps are particularly sensitive to suction
piping configurations which do not provide a uniform velocity profile. The use of an
elbow installed at the pump suction and parallel to the pump shaft will result in
higher fluid velocity and pressure on the outside of the piping turn (elbow) and
uneven flow to each side of the double-suction impeller. Standard elbows present a
bigger problem than long-radius elbows. Uneven flow distribution to each side of
the double-suction impeller can result in axial unbalance (shaft shuttling), cavita-
tion on the low flow side, premature bearing failure, and seal failure. There have
been many cases of premature failures due to improper suction piping layout.
The installation of an elbow where the plane of the elbow is parallel to the pump
shaft is unacceptable. See Figure 100-20.
The plane of an elbow should always be perpendicular to the pump shaft centerline.
This applies whether the fluid supply source is above or below the pump suction.
The use of long-radius elbows is recommended for suction lines to double-suction
pumps.
Inlet Swirl
High velocity and high pressure tend to develop on the outside of each piping turn
and successive turns made at right angles to each other produce a fluid swirl inside
the pipe. Pump impeller and inlet design is based on a uniform velocity profile, not
a swirling fluid. The result of swirling fluid at the inlet of the pump may be cavita-
tion damage, poor pump hydraulic performance (low head), short bearing life, and
noisy operation even if adequate NPSHA is provided.
Minimizing the Effects of Inlet Swirl and Fitting Turbulence

If the suction line (not pump suction nozzle) design velocity is five feet per second
or less, a straight run of pipe directly upstream of the pump suction of at least five
pipe diameters should be adequate to correct irregularities (swirl) in the velocity
profile created by one elbow or tee. A swirl generated by two or more fittings or
elbows may require straightening vanes at least two pipe diameters in length in the
straight length of inlet pipe or at least 10 diameters of straight pipe.
At suction velocities of ten feet per second, the straight section of pipe will have to
be at least ten diameters in length. Similarly, a swirl or disturbance created by two
or more elbows or fittings may require straightening vanes or a longer straight run
of pipe to correct the inlet velocity profile. Generally, the use of reducers at the
pump suction downstream of elbows, valves, and fittings will tend to improve the
inlet velocity profile.
Fig. 100-20 Elbow Orientation for Double Suction Pump Courtesy of the Hydraulic Institute
Abstract
This section describes how centrifugal pumps work, lists their limitations, and
explains how to select the right centrifugal pump for a given application. For infor-
mation on troubleshooting centrifugal pump problems, see Section 1100. For infor-
mation on mechanical seals, or installation or startup of centrifugal pumps, see those
sections.
Contents Page
210 Engineering Principles 200-3
211 Fundamentals
212 Head
213 Pump Curves
214 Series and Parallel Operation of Multiple Centrifugal Pumps
215 Effects of Changing Pump Speed (“Affinity Law”)
216 Effects of Changing Impeller Diameter (“Affinity Law”)
217 Cut-off Point
218 Specific Speed
219 Effect of Viscosity on Centrifugal Pump Performance
220 Suction Considerations 200-25
221 Pumping Liquids Near Their Boiling Points
222 Cavitation
223 Net Positive Suction Head Available (NPSHA)
224 Required NPSH (NPSHR)
225 “Suction-Stealing”
226 Horsepower
231 Factors in Pump Selection
232 Energy Efficiency for Centrifugal Pumps
April 2009  1999–2009 Chevron USA Inc. All rights reserved. 200-1
200 Centrifugal Pumps Pump Manual
233 Special Service Pumps

234 Application Guidelines
240 Centrifugal Pump Descriptions 200-48
250 Mechanical Components 200-77
251 Cases
252 Impellers
253 Wearing Rings
254 Shafts and Shaft Sleeves
255 Throat Bushings and Lantern Rings
256 Glands
257 Balance Drums and Bearings
258 Base Plates
259 Couplings and Coupling Guards
260 Centrifugal Pump Subsystems 200-90
261 Special Requirements for Hot Service
262 Vertical Turbine Pumps
270 Maintaining Centrifugal Pump Flow Rates Close to the Best
Efficiency Point (BEP) or Best Efficiency Flow Rate 200-92
271 General
272 Power Measurement
273 Flow Control Methods
274 Proportional Flow Control
275 Self-Contained Flow Control Valves
276 Economics of Flow Control
277 Variable Speed Devices (VSDs)
200-2  1999–2009 Chevron USA Inc. All rights reserved. April 2009
Pump Manual 200 Centrifugal Pumps
210 Engineering Principles
211 Fundamentals
Centrifugal pumps comprise a wide category of pumps which move liquid by the
rotational motion of one or more impellers. Their flow is uniform and normally
devoid of pulsations.
A centrifugal pump produces pressure by accelerating a fluid to a high kinetic
energy (velocity), then converting that energy to pressure.
Fluid flows into the “eye” of the impeller and is thrown outward by the vanes of the
spinning impeller, slowing as the velocity is converted to pressure in the “diffuser”
or “volute.” (See Figure 200-1). This momentum exchange provides an increase in
pressure or “head.”
Fig. 200-1 End View of a Centrifugal Pump From Centrifugal Pumps Design and Application
by Lobanoff and Ross, Copyright 1985 by Gulf Publishing Company, Houston, TX.
Used with permission. All rights reserved.
The incoming fluid is pushed into the low pressure area of the impeller eye by
higher pressure in the upstream system. Having enough upstream or “suction” pres-
sure to push adequate flow into the pump is a critical design consideration.
(Covered in Section 220.)
212 Head
The term “head” is used almost exclusively in the centrifugal pumping industry to
express pressure. All pump curves are calibrated to read “feet of head” as a
measure of pressure rise. Similarly, suction pressures and, often, friction losses are
also expressed as feet of head, not psi.
The concept of head is derived from the fact that a column of liquid will exert a
local pressure proportional to the depth of that liquid. For example, the pressure of a
column of water increases 0.433 psi for every foot of depth. In other words, at a
depth of ten feet, the pressure is 4.33 psi higher than at the surface; at 100 feet,
43.3 psi higher; at 1000 feet, 433 psi higher, etc.
The depth, or distance in feet, can therefore be used as a measure of pressure. For
water, the equivalent pressures are:
1 foot of head = 0.433 psi (for water at 60F and 1.0 specific gravity) or
1 psi = 2.31 feet of head (for water at 60F and 1.0 specific gravity)
Another example of measuring pressure by liquid depth is the barometric pressure,
reported as millimeters or inches of mercury.
14.7 psi = 760 mm Hg = 29.92 inches Hg
This relationship illustrates that normal atmospheric pressure (14.7 psi at sea level)
is the same pressure that would be exerted at the bottom of a column of liquid
mercury 29.92 inches high (assuming zero pressure—i.e., a vacuum—at the surface
of the mercury).
Similarly, visualize a centrifugal pump connected to a vertical pipe on its discharge.
The discharge pressure from the pump would push the liquid up the column to a
level where the pressure from the height would equal discharge pressure. This
height would be the feet of head noted by the pump manufacturer as “total head”
across the pump.
One reason the centrifugal pump industry has settled on head, or feet, as a measure
of pressure rise is that a pump will develop the same head regardless of the fluid’s
specific gravity. A pump that develops a column of water (S.G.=1) 1000 feet high
will also develop a column of hydrocarbon (S.G.= 0.7) 1000 feet high.
Of course, the actual pressure, in psi, would be quite different between water and
hydrocarbon. The pressure developed in a pump and the pressure at the bottom of a
column of liquid are both proportional to specific gravity. To convert from feet to
psi (and vice versa) use the following equation:
Pressure (psi) = feet  S.G.  0.433
(Eq. 200-1)
213 Pump Curves

Total Developed Head (TDH) is a measure of the energy a pump delivers to a fluid.
It is equal to the total discharge head minus the total suction head in feet of liquid.
The word “total” is used because each of these heads is composed of the pressure
head, velocity head, static head, and head loss. The Total Developed Head is
approximated by measuring the discharge pressure and suction pressure at the pump
nozzles, subtracting to determine the differential pressure, and converting to units of
head in feet. This approximation neglects the velocity head component, which
usually results in an error of 1% or less. A centrifugal pump’s Total Developed
Head depends on the impeller diameter, pump speed, fluid viscosity, impeller and
case design, and pump mechanical condition. It also varies with flow rate, largely
due to frictional losses in the impeller and casing. This relationship is plotted in a
“pump curve.” These characteristic curves are important to understanding the
performance of centrifugal pumps.
Typical Characteristic Curves for a Centrifugal Pump

Most characteristic curves show the relationship between Total Developed Head
(TDH), pump capacity (flow rate), brake horsepower, efficiency, and Net Positive
Suction Head Required (NPSHR) for a particular pump. Characteristic curves are
also known as “head-capacity” curves or, simply, “pump curves.” Two methods are
commonly used for plotting the characteristic curves of a centrifugal pump.
Figure 200-2 shows the method used to depict pump performance for a single speed
and impeller size. These curves result from a pump test at constant speed. Manufac-
turers commonly use these characteristic curves to predict and guarantee pump
performance.
Figure 200-3 shows the method used to express more fully the entire range of
performance of a pump, with various impeller diameters at constant speed. These
curves are commonly used in the selection of a pump for a specific service. The
curves in Figure 200-3 are generally made up from the average results of tests for
various diameter impellers plotted as shown in Figure 200-2.
Figure 200-4 shows a third method of plotting characteristic curves for a centrifugal
pump driven at variable speeds, with a fixed impeller diameter.
Note that practically all performance curves furnished by manufacturers are based
on water as the pumped liquid. If the pump is handling some other liquid, adjust-
ments must be made for viscosity and specific gravity before flow rate and
discharge pressure (psi) can be predicted.
Every centrifugal pump will operate on its characteristic curve if there is
enough Net Positive Suction Head Available (NPSHA) for a given S.G. and
viscosity. For any given capacity, there will be one total head rise, one efficiency,
one horsepower, and one NPSHR.
The slope and shape of the head-capacity curve is affected by individual pump
design. Head-capacity curves can take one of four typical shapes, as shown in
Figure 200-5.
• Steep-rise curve
• Steady-rise curve
• Flat curve
• Drooping curve (will have multiple flow points for a given head)
As a rule of thumb, curves that show a 140% increase in head between the capaci-
ties of peak efficiency and shutoff are called steep-rising curves; those showing a
10–25% increase are called steady-rising curves; and those with no more than a
5% increase are called flat curves. Rise to shutoff is a function of the following
parameters:
• Specific speed (Ns) design for the impeller
• Impeller-outlet-vane angle and volute diffuser area ratio
• Friction losses
Fig. 200-2 Typical Performance Curve for a 6-inch, Single-stage, Double-suction Centrifugal Pump—Speed and Impeller Diameter Fixed.
200-6

 1999–2009 Chevron USA Inc. All rights reserved.
Pump Manual
April 2009
Fig. 200-3 Typical Performance Curve for a 6-inch, Single-stage, Double-suction Centrifugal Pump—Speed Fixed, Impeller Diameter Variable
April 2009
Pump Manual

200-7
Fig. 200-4 Typical Performance Curve for a 6-inch, Single-stage, Double-suction Centrifugal Pump—Speed Variable, Impeller Fixed
200-8

Pump Manual
April 2009
Fig. 200-5 Four Typical Shapes of Head Capacity Curves
Pumps with drooping characteristic curves should be avoided because they may
exhibit unstable operating characteristics. In some cases, however, such as systems
with mostly dynamic loss and no requirements for parallel operation, drooping char-
acteristics could be acceptable.
Centrifugal pumps with steady-rise curves are most commonly used. Since the head
varies distinctly with a change in capacity, precise flow control can be maintained
with this type of curve. The rising curve is a stable curve; for every head, only one
corresponding capacity occurs.
System-Head Curves
Plotting the head vs. flow rate in a pumping system can be an aid in system design
and pump selection. Such a plot is called the system-head curve.
A system curve represents a complete piping system, i.e., the friction losses of all
the piping, elbows, valves, etc., and the total static head vs. flow rate. Each point on
the curve shows the head required to deliver that amount of flow through the piping
system.
A system-head curve (Figure 200-6) is obtained by combining the system friction
curve (Figure 200-7) with a plot of the total developed head. A system friction
curve is a plot of friction losses versus flow rate in a piping system.
Superimposing the pump characteristic curve on the system-head curve gives the
point at which a particular pump will operate (Figure 200-6, Point “A”). Changing
the resistance of the piping system by partially closing a valve changes the system-
head curve. Partially closing a valve in the discharge line produces a second system-
head curve, shown in Figure 200-6, shifting the operating point to higher head but
lower flow rate. The intersection of the pump characteristic curve and the new
system-head curve is the new operating point.
Fig. 200-6 Pump Characteristic Curve Superimposed Fig. 200-7 System Friction Curve
on System-Head Curve
Operating Point
It is important to understand that a centrifugal pump will operate at one point
(assuming the pump curve rises steadily to shutoff). This point is the intersection
of the system curve and the pump curve. This is an important concept both for
sizing pumps and troubleshooting problems.
This concept also illustrates the most common basis for centrifugal-pump control:
discharge throttling. As a control valve in the discharge line varies the total pres-
sure drop in the system, the system curve varies. This variance in the system curve
causes the operating point to shift right or left on the pump curve, with a resulting
increase or decrease in flow rate.
Unstable Head-Capacity (Drooping-Curve) Characteristics

Under certain conditions, a portion of the head-capacity curve of a low-specific-
speed pump is unstable, causing fluctuations in the pump head, capacity, and power
input. Figure 200-8 shows the type of head-capacity curve (a drooping curve) that
can cause unstable operation.
In Figure 200-8 the system curves OB, OC, OD, OE and OF correspond to different
settings of a pump discharge throttle valve. Point F represents the normal operating
condition of the pump. As system resistance is increased (by throttling the discharge
valve, for example) the pump is able to supply the additional head until point C is
reached on the pump head-capacity curve. Additional system resistance causes the
operating point to move into the part of the pump curve where the head decreases as
the flow decreases. Operation in this region of the head-capacity curve may result is
an unstable surging discharge pressure.
It is not good practice to install drooping-curve centrifugal pumps in parallel. One
pump may operate at a lower flow rate than the other and could fail if operating
below the manufacturer’s recommended minimum flow rate.
Fig. 200-8 Typical Head-Capacity Curve that May Indicate Unstable Operation
(Drooping Curve)
214 Series and Parallel Operation of Multiple Centrifugal Pumps

Centrifugal pumps may be operated in series or in parallel. The combined head-
capacity curves for series or parallel operation of two or more centrifugal pumps are
obtained as follows:
• Series: Add heads for each pump at any given capacity.
• Parallel: Add capacities for each pump at any given head.
Figure 200-9 illustrates both series and parallel operation for two pumps under
various discharge conditions. Two pumps, P-1 and P-2, have head-capacity curves
as shown and are to pump through pipe systems with characteristics shown by
system curves I, II, III, IV, and V. The intersections of the pipe system characteris-
tics with the pump head-capacity characteristics show the quantities and heads at
which the pumps will operate either singly, in series, or in parallel. Adequate
suction pressure is assumed.
200-12

Fig. 200-9 Typical Series and Parallel Operation of Two Centrifugal Pumps Pumping Through a Pipe System Throttled at the Discharge End
Pump Manual
April 2009
Figure 200-10 is an example of the difficulty with series pump operation. Two iden-
tical pumps, P-5 and P-20, operate in series. The suction and discharge pressures
are noted on the diagram. Both pumps should develop the same differential head.
Actually, P-5 develops a differential of 20.5 psi and P-20 develops a differential of
72.0 psi. Average capacity is 543 GPM, which is well below the anticipated flow
rate. The performance curve for the two pumps, Figure 200-11, shows that P-20 is
developing its rated head but P-5 is not. The difficulty is that Pump P-5 is “losing
suction” and “cutting off” at about 543 GPM as shown on Figure 200-11.
In Figure 200-10, the actual differential developed by P-5 is shown by AC. The
differential head developed by P-20 is shown by DG. The sum of these two
produced the head required at H for a flow of 543 GPM. If P-5 had been provided
with adequate suction pressure, it would have developed a differential head equal to
AE. The total pressure which both pumps would have developed is shown by BI.
215 Effects of Changing Pump Speed (“Affinity Law”)

Knowing the effects of varying a centrifugal pump’s speed is helpful in many situa-
tions, such as adjusting to new service requirements, sizing a new driver, turning
down to avoid excessive flow or pressure, etc.
The following affinity law holds for any corresponding points on the head-capacity
characteristic curve when the speed is changed:
1. Flow rate (quantity) varies directly with the ratio of change in speed.
2. Head varies with the square of the ratio of change in speed.
3. Horsepower varies with the cube of the ratio of change in speed.
In all three cases, the efficiency remains relatively constant. Efficiency tends to rise
very slightly as speed increases, because neither hydraulic nor mechanical losses
increase as fast as the square of the speed.
The characteristic curve of Figure 200-4 is marked to show a set of corresponding
points for the same impeller at different speeds.
The affinity law for speed change holds with considerable accuracy when speed
changes do not exceed a two-to-one ratio and flow is not limited by suction
conditions.
Fig. 200-10 Analysis of Performance of Two Identical Centrifugal Pumps in Series When Suction Pressure at First
Pump is Too Low
Fig. 200-11 The Effect of Abnormal Suction Conditions on Centrifugal Pump Performance
216 Effects of Changing Impeller Diameter (“Affinity Law”)

The curves in Figure 200-3, except the underfiled curve, may be approximated from
a single curve by the following rules, which apply to reducing impeller diameter to
the stated design minimum without other changes in design. They are applicable to
minor changes (5-15%) in impeller diameter.
The following rules may be applied for any corresponding points on the character-
istic curves when the impeller diameter is changed:
1. Flow rate (quantity) varies directly with the ratio of change in impeller diam-
eter.
2. Head varies with the square of the ratio of change in impeller diameter.
3. Horsepower varies with the cube of the ratio of change in impeller diameter.
These rules are essentially the same as the affinity law for speed change, but do not
apply with the same accuracy over as wide a range.
For (1), (2), and (3) all to be true, the efficiency must remain constant for the corre-
sponding point. Since this is not exactly what happens, the head calculated by the
above rules will be too low. The efficiency will usually drop. The table in
Figure 200-12 will aid in estimating how much deviation from the simple rule
should be expected. Both columns give impeller diameter, in percent, of original
diameter.
Fig. 200-12 Impeller Diameters (% of Original)
% to Reduce Impeller, as
Calculated by the Affinity Law Actual % Impeller Reduction
65 71
70 75
75 79
80 83
85 87
90 91.5
95 91.5
When the cut becomes so great that the overlap of the vanes is destroyed, proper
guidance or “control” of the liquid is lost and the performance becomes unpredict-
able. When possible, the correct diameter for new conditions should be obtained
from the manufacturer.
Conservative practice limits the diameter after cutting to not less than 75% of the
full diameter. The pump manufacturer can readily determine the allowable
minimum diameter from the impeller drawings.
The affinity law for impeller diameter applies not only to the point of best effi-
ciency, but to any corresponding points on the original and calculated new head-
capacity characteristics, provided they are not affected by suction conditions.
The combined effects described above are summarized in the affinity law equa-
tions shown in Figure 200-13.
Fig. 200-13 Affinity Law Equations From Centrifugal Pumps Design and Application by Lobanoff and Ross, Copyright
1985 from Gulf Publishing Company, Houston, TX. Used with permission. All rights reserved.
Diameter Change Only Speed Change Only Diameter and Speed Change
Q2 = Q1 (D2/D1) Q2 = Q1 (N2/N1) Q2 = Q1 (D2/D1  N2/N1)
H2 = H1 (D2/D1)2 H2 = H1 (N2/N1)2 H2 = H1 (D2/D1  N2/N1)2
BHP2 = BHP1 (D2/D1)3 BHP2 = BHP1 (N2/N1)3 BHP2 = BHP1 (D2/D1  N2/N1)3
where:
Q1 = Initial flow rate Q2 = New flow rate
H1 = Initial differential H2 = New differential head
head
N2 = New rpm
N1 = Initial rpm
D2 = New diameter
D1 = Initial diameter
BHP2 = New brake horsepower
BHP1 = Initial brake
horsepower
Effects of Changing Liquid Specific Gravity

Specific gravity (S.G.) has the following effects on pump performance, assuming
constant rpm and impeller diameter:
1. Flow rate (quantity) is unchanged by S.G. (although the flow reading on a
differential-pressure flow meter varies.)
2. Pressure varies directly with S.G. (Although pressure varies, head is constant.)
3. Horsepower varies directly with S.G.
These relationships are important when converting a pump to another service or if
significant changes to fluid gravity are anticipated. For example, converting from a
light hydrocarbon service to water service may significantly overload an existing
driver.
Increasing the Capacity of a Given Pump

Increasing the capacity and head of a pump within its design limits is usually
accomplished by increasing impeller diameter or driver speed.
Small increases can be obtained by underfiling the impeller vanes without changing
impeller diameter. This means that the exit end of the vanes are filed back, without
cutting the shroud, as shown in Figure 200-14. (Figure 200-3 shows the effect on
the pump curve of “underfiling” the impeller.)
Fig. 200-14 Underfiled Vanes on a Centrifugal Pump Impeller
In Figure 200-3 the head-capacity curve for the underfiled condition is for the full
diameter vanes. Similar effects are obtained by underfiling any other usable diam-
eter. Underfiling is adopted only in cases where the standard impeller does not attain
the required rating and changing the impeller or using a larger pump is not
warranted.
Figure 200-3 shows a set of typical characteristic curves for a 6-inch, single-stage,
double-suction pump running at 1770 rpm. Total pumping head, efficiency, and
horsepower are plotted against capacity for impeller diameters from 15½ to
18¼ inches using the standard vane, and also for full diameter with underfiled
vanes. Note that the underfiled curve is unstable. Underfiling pumps with flat curves
can lead to unstable (drooping) curves; this would not happen on pumps with steep
curves. This is a good example of why underfiling should be carefully considered.
217 Cut-off Point

Figure 200-11 shows that the greatest possible capacity obtainable with this pump is
about 1100 GPM, which may be obtained at a head of 150 feet. This point is known
as the cut-off point and is the maximum quantity of liquid that the available suction
head can force into the impeller. The cut-off point depends on the relationship
between required and available NPSH. See Section 220 for a complete discussion of
NPSH and Figure 200-21 for an example of NPSH limiting capacity.
Pumps should not be selected with a cut-off close to the required rating. Pumps
operating above cutoff will vibrate excessively and fail prematurely.
218 Specific Speed

Specific speed is a dimensionless term used to compare the performance and shape
of impellers, regardless of their size. Specific speed (usually designated Ns) is the
speed, taken in revolutions per minute, at which a geometrically similar
impeller would run if it were of such size as to discharge one gallon per minute
against one foot of head.
In practice, specific speed is used to relate the three main parameters (GPM, head,
and rpm) to the performance of the pump:
0.5
Q
N s = n -------------
0.75
H
(Eq. 200-2)
where:
Q = U.S. gallons per minute
H = Feet per stage
n = Revolutions per minute
Low-specific-speed impellers have high heads and low flow capacities. Impellers
for low heads and high flow rates have high specific speeds.
Figure 200-15 gives the general relationships between impeller shape, efficiency,
and capacity. It also shows that each impeller design has a specific speed range for
which it is best adapted. These ranges are approximate, without clear-cut demarca-
tions between them. Most petrochemical pumps are designed with impellers that
have specific speeds between 800–1500 (as calculated using Equation 200-2).
Fig. 200-15 Relationship of Impeller Shape, Efficiency, and Capacity From Pump Handbook,
(1976) Edited by Karassik, Krutzch, Fraser, & Messina. Used with permission from
McGraw Hill.
Specific speed is a pump design tool, but it may be used in the pump selection
process to compare the curve shape and stability. It can also be used in evaluating
new pump bids. (See Section 231.)
In general, low specific speeds indicate flat head-capacity curves, with peak effi-
ciency over a wide range of capacity, and brake-horsepower decreasing as the pump
is throttled. High specific speeds result in steep head-capacity curves, sharply
peaked efficiency curves, with brake-horsepower increasing as the pump is throt-
tled.
219 Effect of Viscosity on Centrifugal Pump Performance

Since requirements often call for pumping liquids with a viscosity greater than
water (while most manufacturers’ curves are for pumping water), it is important to
have a method for estimating the effect of viscosity upon water performance curves.
In general, because of the increased internal fluid friction, the head, efficiency, and
flow of centrifugal pumps are reduced when pumping a fluid with a higher viscosity
than water.
Figure 200-16 shows the effect of viscosity on pump performance. Figure 200-17
(1 and 2) provides viscosity corrections to pump performance. These data are also
available from the Hydraulic Institute Standards, 14th Edition. The curves convert
the pump’s water performance to that of the viscous fluid.
These correction curves do not apply to mixed-flow or axial-flow pumps, nor to
pumps handling non-Newtonian liquids. Slurries and similar non-Newtonian liquids
may produce widely different results depending on their characteristics. Also, the
correction curves cover only single-stage performance using the best efficiency flow
rate for the impeller. If viscous performance for a multi-stage centrifugal pump is
required, the head per stage should be used to obtain the proper correction factors,
which should then be verified with the original equipment manufacturer.
It is worth noting that, at 100 GPM, Figure 200-17 (1 and 2) gives somewhat
different results, indicating they are compiled from separate tests and that either
chart is only an approximation of the actual results for a viscous liquid.
The correction curves provide factors to be applied at the best-efficiency-point to
arrive at the viscous performance curve. Efficiency is the parameter affected most
severely by viscosity, followed by capacity, then head. In practice, since efficiency
has the greatest effect, power cost should be evaluated as it may impact the pump
selection.
Positive-displacement reciprocating screw or gear pumps are very efficient in
viscous fluids. They should be considered when fluid viscosity exceeds 200 to
500 SSU and when there are very few suspended solids present.
Fig. 200-16 Effect of Viscosity on Centrifugal Pump Performance. Note: In Figure 200-17 (both parts 1 and 2, overleaf), enter the chart at GPM, read vertically to
200-21

Head, then Horizontally to Viscosity, then vertically to Head/Capacity/Efficiency, then left to the Correction Factor.
Pump Manual
April 2009
Fig. 200-17 Viscosity Corrections for Centrifugal Pumps Handling Viscous Fluids 100 GPM and Over (1 of 2) From
Standards 14th edition, Hydraulic Institute. Used with permission.
Fig. 200-18 Viscosity Corrections for Centrifugal Pumps Handling Viscous Fluids Under 100 GPM (2 of 2) From Stan-
dards 14th edition, Hydraulic Institute. Used with permission
Sample Problem: Viscosity Effects. Given the pump performance obtained by test
on water, plot the performance of this pump when handling oil with a specific
gravity of 0.9 and a viscosity of 1,000 SSU, both at pumping temperature.
On the performance curve, Figure 200-16, the best-efficiency-point when pumping
water is 750 GPM at 100 ft of head (Point A) with an efficiency of 82% (Point B).
Using 750 GPM, 100-ft head, and 1,000 SSU, read Figure 200-17 (1 of 2) and deter-
mine the correction factors:
Capacity correction factor: CQ = 0.95

Head correction factor: CH = 0.92
Efficiency correction factor: CE = 0.635
Multiplying the water capacity, head, and efficiency by the correction factors gives
the best-efficiency-point as follows:
Viscous capacity:
750 GPM  0.95 = 712 GPM
Viscous head:
100 ft  0.92 = 92 ft
Viscous efficiency:
82%  0.635 = 52%
The point for viscous capacity and head can now be located below the water curve
(Point C, Figure 200-16). The viscous head-capacity performance curve is drawn
from the water head at zero capacity (Point D) through the viscous head-capacity
point (Point C) with approximately the same shape as the water curve. The effi-
ciency at the best-efficiency-point for viscous performance can be plotted as Point E
and the viscous efficiency curve plotted from zero (Point F) through Point E; the
shape of the curve is similar to that obtained for water efficiency.
The horsepower (BHP) for any capacity can now be calculated from the head and
efficiency at the capacity desired. The best-efficiency-point for viscous perfor-
mance is:
712GPM  92 ft.  0.9 S.G.

BHP = -------------------------------------------------------------------- = 28.6
3960  0.52 eff
(Eq. 200-3)
This horsepower can now be plotted as Point G and the horsepower curve for
viscous performance drawn through Point G approximately parallel to the brake
horsepower curve for water.
220 Suction Considerations

One of the most important aspects of successful pump operation is to have enough
suction pressure to push liquid into the pump without flashing or boiling. This
requirement is particularly critical where liquids are already near their boiling points
(reflux, boiler feedwater, flash separators, furnace circulation, etc.). Failure to
assure adequate suction pressure will lead to numerous operational and
mechanical problems, up to and including destruction of the pump.
221 Pumping Liquids Near Their Boiling Points

Pumps should be selected with inlet velocities sufficiently low to prevent vapor
formation in the entering liquid. This may call for (1) oversized inlet piping,
(2) pumps operating at low speed, (3) pumps designed for such conditions, or
(4) use of vertical pumps installed in a suction can.
The design requirement is that the pressure at the pump inlet be adequate to accel-
erate the liquid to the required velocity at the impeller entrance without the pressure
in the pump falling below the fluid’s vapor pressure. Boiling or flashing of the fluid
in the pump suction eye is called cavitation and can significantly affect pump
performance.
222 Cavitation
The formation of vapor bubbles in the impeller suction eye due to fluid flashing or
boiling, with subsequent collapse of the bubbles as the pressure rises, is called cavi-
tation. Cavitation may cause vibration, pitting damage, and impaired performance.
Cavitation may or may not be serious depending on the pump, HP/stage, impeller
design, and the fluid being pumped. In small pumps with low differential head per
stage, the energy of collapsing bubbles is much less than in larger, high-head-per-
stage pumps. Cavitation is more severe in a single-boiling point fluid (like water)
than with a mixture (like petroleum stocks) that have a broad boiling range.
Recirculation
Recirculation is a flow reversal at the inlet eye or discharge tip of an impeller.
Recirculation at the inlet eye is called suction recirculation. Discharge recirculation
occurs at the impeller tip. Recirculation usually occurs when operating centrifugal
pumps at flows below their best efficiency flow.
Refer to standard drawing GA-G1097-2, “Minimum Continuous Flow for Centrif-
ugal Pumps,” to help predict the flow at which a pump will begin to demonstrate
problems related to suction recirculation. Section 270 describes several ways to
prevent pump operation below the recommended minimum flow.
All impellers will begin to recirculate at a certain flow rate. The point recirculation
begins may not be the same for suction and discharge. Suction recirculation usually
will begin at a higher flow than discharge recirculation.
The capacity at which recirculation occurs is determined primarily by the impeller

design. Most of the problems associated with recirculation can be avoided by
selecting pumps with impellers of low suction specific speed (Nss) designs. Recom-
mended limits for Nss are:
Horsepower Per Stage < 250 to 300 > 300
Nss limit 11,000 9,000
The effects of recirculation can be impeller and casing damage, bearing failures, and
seal or shaft failures. Symptoms associated with recirculation are listed below.
Suction Recirculation:
• Cavitation damage to the pressure side of the impeller vanes at the inlet of
the vane.
• Cavitation damage to the stationary or splitter vanes in the suction side of the
pump casing.
• Random crackling or gravel pumping noise. (Inadequate NPSH will sound the
same except the noise will be constant not random.)
• Surging pressure in the suction pipe.
Discharge Recirculation:
• Cavitation damage to the pressure side of the impeller vane and exit shroud at
the discharge of the impeller. This may be seen as impeller failures at the
impeller exit vanes or shroud.
• Higher-than-normal axial vibration or shaft movement. This may be accompa-
nied by thrust bearing damage.
• Cavitation damage to the “cut water” (casing tongue) or diffuser vanes in
the case.
223 Net Positive Suction Head Available (NPSHA)

NPSHA is a critical factor in pump performance. It is a result of the suction system
design. In practical terms, NPSHA is the differential pressure between (1) the actual
pressure at the lowest pressure point in the pump, and (2) the pressure at which the
liquid begins to vaporize (flash or boil). NPSHA is the “available” pressure above
the liquid’s vapor pressure that prevents vaporization (or cavitation). Remember that
as the liquid accelerates into the spinning impeller eye, its pressure drops. If the
pressure falls below the vapor pressure, cavitation occurs.
NPSHA is technically defined as the total suction pressure (in psia) at the
suction nozzle less the true vapor pressure of the liquid (in psia) at the pumping
temperature. For centrifugal pumps, NPSHA is always expressed in feet of the
liquid pumped.
Net Positive Suction Head Required (NPSHR)

NPSHA must exceed the NPSH required by the pump. NPSHR depends on the
impeller design, operating speed and flow rate, and, to a lesser extent, on the charac-
teristics of the liquid handled. NPSHR represents the frictional losses and initial
pressure-to-velocity energy conversions occurring between the suction flange and
the point where the impeller begins to do work on the fluid.
During NPSH testing and NPSH curve development, the pump manufacturer oper-
ates the pump at a constant flow rate while closely monitoring the pump head as
suction pressure is reduced. During the process of lowering suction pressure, cavita-
tion begins. When the volume of the vapor bubbles impairs pump performance by a
reduction in head of 3%, the pump manufacturer defines that NPSH value as the
required NPSH for that particular flow. This is repeated at several flow points to
develop an NPSHR curve.
NPSH testing is done using cold water as the pumped fluid. The values of NPSHR
determined from cold water tests are conservative and are practical to use for virtu-
ally all services.
NPSHR, Suction Specific Speed, and Minimum Flow

The NPSHR by a pump is largely dependent on the impeller “eye area” and inlet
“vane angle” design. These relatively complicated and proprietary design features
can easily be evaluated by comparing each pump’s Suction Specific Speed (Nss).
Nss is a design number which relates the best-efficiency flow and NPSHR for the
maximum diameter and pump rpm. This value provides a great deal of information
about pump performance. To calculate Nss, use the following formula:
0.5
Q N
Nss = ----------------------------------
0.75
 NPSHR 
(Eq. 200-4)
where:
Q = pump best efficiency flow in GPM for the maximum diameter
impeller. Q divided by 2 is used for double suction impellers.
N = pump rotating speed in rpm
NPSHR = net positive suction head required in feet at flow point Q
Typical values for Nss range between 7,000 and 14,000 as determined by pump
design. However, conservative impeller designs will have a Nss value less than
11,000. Multistage, high-energy pumps which operate above 3600 rpm should have
a first-stage impeller Nss value of less than 9000.
The following is an example of the relationship between Nss, NPSHR, and pump
minimum flow.
PUMP #1 PUMP #2
Manufacturer ABC Co. XYZ Co.

Model 328 328
Speed (rpm) 3,600 3,600
NPSHR (feet) 10 6
Nss 8,750 11,500
Minimum flow (GPM) 30 60
Pump #1 with the lower Nss requires a higher NPSHR and has a lower minimum
flow. Therefore: (1) Pump #2 probably has a larger impeller inlet eye area and less
conservative inlet vane angle design; and (2) due to the less conservative design of
Pump #2, the stability of flow in the impeller is reduced at lower flow rates resulting
in a higher minimum flow.
In summary, as Nss increases, the pump NPSHR decreases, and the pump minimum
flow increases.
Company experience has shown that pump reliability is directly related to the pump
Nss. Pumps with Nss values above 11,000 are less reliable. The lower reliability
usually manifests itself as high vibration and shaft deflection due to flow instability
in the impeller eye. The shaft deflection and vibration results in reduced mechan-
ical seal and bearing life.
Refer to Figure 200-19 for a nomograph to help determine NPSHR or Nss values
for pumps without the need for calculation. The nomograph along with basic knowl-
edge of pump performance requirements can (1) assist in the selection of a conser-
vatively designed pump by establishing design parameters for new or retrofit of
existing pump suction systems; and (2) help diagnose problems with existing pump
suction systems.
Refer to GA-G1097-2 to help determine the stable operating range for the selected
pump based on its Nss. This figure can also be used to compare minimum flow
quotes from various vendors, as they often will not consider the Nss of the pump
when quoting the stable minimum flow.
Fig. 200-19 Specific Speed and Suction Specific Speed
NPSHR for Liquids Other than Cold Water

Manufacturer’s test data for NPSHR are published based on cold water and
are normally included on pump performance curves. When liquids other than
cold water are handled, the actual NPSHR becomes uncertain. Tests, however, indi-
cate that cavitation starts at nearly the same NPSH for all liquids, but that some
liquids (primarily high-vapor-pressure liquids such as propane and butane) do not
require as much NPSH as does cold water.
Three factors cause the NPSHR for some liquids to be less than for cold water:
1. Vaporization removes heat from surrounding liquid, reducing its vapor pres-
sure, and suppressing further vaporization. The magnitude of this effect
depends on the thermodynamic properties of the liquid at the suction condi-
tions.
2. The volume of vapor bubbles in the impeller eye determines the extent to which
performance is impaired. The volume of vapor formed depends on the pressure
and temperature at which vaporization takes place and on the molecular weight
of the stock. To make the same volume of vapor, more weight of a high-vapor-
pressure stock must be vaporized than of a low-vapor-pressure stock. The
higher molecular weights of hydrocarbons compared to water require
more hydrocarbon than water to be vaporized for the same volume of
vapor formed.
3. Multi-component liquids have light ends that vaporize first. These may be
small enough in proportion to the total fluid so that some vaporization can
reduce the vapor pressure before pump performance is seriously impaired. This
effect will vary with changes in the composition of the hydrocarbon. Some
hydrocarbons require almost as much NPSH as cold water if the fractions of the
stock first evaporating are significant in relation to the whole NPSH for a given
service condition. The use of any NPSH correction factor which supposedly
allows less NPSHR than cold water is not recommended.
Vapor Pressure and NPSH

A primary factor in calculating the NPSHA for a pump is the vapor pressure of the
liquid handled. One commonly used method, Reid vapor pressure, requires a certain
amount of liquid to be evaporated in the measuring apparatus before the vapor pres-
sure is indicated. Such vapor pressures are too low for determining when gas evolu-
tion will start (the point that will affect pump performance). This error is variable,
being small for fractioned stocks and greater for wild crudes. The true vapor pres-
sure (TVP) at the pumping temperature should be used for NPSHA calcula-
tions rather than vapor pressure by the Reid method.
In determining true vapor pressure, do not overlook the possibility of dissolved
gases in the liquid. A frequent cause of NPSH trouble is dissolved or entrained air
or gas in the liquid pumped. When tested by the bubble-point method, water which
has been aerated has a higher “vapor pressure” than water which has not been
aerated. The same is true for hydrocarbons or other liquids. When the pressure of a
liquid containing dissolved gases is reduced, the gas dissolved in the liquid may
evolve and cause an effect similar to cavitation.
You must consider the effect of temperature changes on vapor pressure in

determining the NPSH available for a pump. Vapor pressure is a function of
temperature alone for any given composition of liquid. For some fluids, a small
increase in temperature causes a relatively large increase in vapor pressure. When
selecting a pump for such a fluid (water, for example), see that the NPSHA is calcu-
lated at the highest probable fluid temperature.
The same precaution applies to pressure changes. The NPSHA must take into
account any reduction in suction pressure that might result from pressure variations
in the system. This is of particular importance in applications such as boiler feed
pumps, where you should always make reasonable allowance for variation in deaer-
ator pressure and its effect on pump suction.
NPSHR Quotations
Since most pumps are tested by the manufacturer on cold water only, quotations by
the supplier will usually provide the cold water NPSHR.
Calculation of NPSHA
NPSHA can be calculated as follows:
NPSHA = H + S - F - Vp
(Eq. 200-5)
where:
NPSHA = feet of head of the pumped liquid, at the pump impeller-eye eleva-
tion and suction flange face.
H = minimum absolute pressure on the surface of liquid pumped, in
feet of the liquid.
S = static head, or vertical distance between the surface of the liquid
and the center of the impeller, in feet. S is negative (-) when the
pump is above liquid surface, and positive (+) when the pump is
below.
F = friction losses, in the suction pipe and fittings, in feet of the
liquid.
Vp = True vapor pressure of the liquid, in feet of liquid, at pumping
temperature. For water this may be determined from the steam
tables. For hydrocarbons refer to ETC technical data books,
process designs, or other sources. (Also see the Appendix.)
H and Vp are calculated from pressures in absolute, not gage units. (Absolute
pressure = gage pressure plus atmospheric pressure).
Sample Calculation: Static Head (S)

Gasoline is to be pumped at a rate of 300 GPM from a tank having atmospheric
pressure on the surface of the gasoline. What is the minimum required static head,
S, to satisfy the pump NPSH requirements?
Since we want to obtain Static head (S), Equation 200-5 can be rearranged to:
S = Vp + F + NPSHA - H
A check on the gasoline shows that the true vapor pressure is 10 psi absolute, and
the specific gravity is 0.75. Therefore:
Vp psia  (2.31 feet/0.75) = 30.8 ft.
From the size and length of the line, fittings, and quantity to be pumped, the friction
head loss of the suction line is found to be:
F = 10 ft.
To calculate NPSHA for the specified pump flow of 300 GPM, Figure 200-19
shows the pump requires a NPSH of 10 feet, assuming 3600 rpm operation and a
Nss of 11,000.
Since the objective is to find the necessary static head (S) to satisfy the pump NPSH
requirements, we can substitute the 10 feet required from Figure 200-19 and add an
operating margin of 4 feet, for the minimum necessary NPSHA.
In other words, we must provide:
NPSHR from Figure 200-19 10 feet
Operational margin 4 feet
System NPSHA by design 14 feet
(The minimum recommended operational margin is 2 feet, a margin of 4 feet is

preferred.)
H is the atmospheric pressure, or 14.7 psia:
14.7  (2.31/.75) = 45.4 feet of gasoline
Substituting in the equation,
S = Vp + F + NPSHA - H
S = 30.8 + 10 + 14 - 45.4
and
S = 9.4 ft.
The positive value of S indicates that the center of the impeller must be below the
surface of the gasoline; the example shows that the center of the impeller should be
at least 9.4 feet below the lowest level of the gasoline in the tank.
Figure 200-20 shows variations of the equation for calculating NPSHA, depending
on whether the liquid surface is above or below the pump centerline, and open or
closed to atmospheric pressure.
Fig. 200-20 Calculations of System Net Positive Suction Head Available (NPSHA) for Typical Suction Conditions Cour-
tesy of Goulds Pumps, Inc.
Legend:
S = Static head, feet absolute
Vp = Vapor pressure of the liquid at maximum pumping temperature, in feet absolute
H = Pressure on surface of liquid in feet absolute
F = Friction losses, feet absolute
224 Required NPSH (NPSHR)

NPSHR is a function of pump design, varying with the capacity and speed of any
given pump. While NPSHA is easily calculated for a given set of conditions, the
NPSHR for a particular pump must be obtained from the manufacturer (determined
by the actual testing of a similar pump) or estimated from Figure 200-19.
If NPSHR is particularly critical for the pump application being considered, an

NPSH test can be specified for the actual pump being furnished. This test is recom-
mended if the difference between NPSHR and NPSHA is less that 4 feet for a
centrifugal pump.
For a properly designed pumping system:
NPSHA  NPSHR + OM
(Eq. 200-6)
where OM (operating margin) is the extra margin to suppress cavitation.
Values of OM may be selected from experience or in consultation with a specialist.
For most centrifugal pump applications, a 2-foot margin should be considered as a
minimum, with values from 3 to 5 being recommended. Any margins less than
4 feet should be demonstrated by an NPSH test in the manufacturer’s shop. New
pump or impeller designs should also be NPSH and performance tested.
Limit of Capacity Due to NPSH

Figure 200-21 shows a method for computing the capacity limitation imposed by
the NPSH on a given pump.
225 “Suction-Stealing”
When two or more pumps are connected to the same suction header and operated in
parallel, the total volume pumped is often much less than proportional to the num-
ber of pumps used. One pump seems to take all the liquid from the other pump or
pumps. This effect, called “suction-stealing,” arises from unequal suction pressures
at the impeller inlets of the various pumps. It is most pronounced where the pres-
sure in the suction header is low, so that the inequalities in friction between the inlet
to the header and inlets to the various pump impellers greatly influence the volume
of flow into the pump. The remedy is to provide equal head losses between the inlet
to the header and the inlets to the pump suction nozzles and adequate NPSHA to
both pumps at the total flow rate. Independently matched pump curves give the
same effect, especially if they are “flat,” permitting minor inlet piping variances to
produce major effects. Actual cases of suction-stealing can usually be traced to flat
or unstable curves.
It is equally important that pumps in series have adequate suction pressure. Occa-
sionally, pumps in series operation have not developed the anticipated total differen-
tial head. This is usually the result of one pump operating under cavitating
conditions because of insufficient NPSHA. Figure 200-21 shows how capacity is
limited when adequate NPSHA is not provided.
Fig. 200-21 Limit of Capacity Due to Net Positive Suction Head (NPSH)
226 Horsepower
The hydraulic horsepower (HHP) for a centrifugal pump is a theoretical value calcu-
lated from the rated capacity and differential head, assuming a 100% efficient pump.
It can be calculated as:
HHP = ( Q  H  S.G. ) / 3960
where:
HHP = hydraulic horsepower
Q = rated capacity in gpm
H = differential head at rated capacity in feet
S.G. = fluid specific gravity
(Eq. 200-7)
Once the pump efficiency is known, the rated power (BHP) can be determined. The
rated power is the power which the pump driver must transmit to the pump shaft at
the rated pump capacity. It can be calculated as:
BHP = HHP / eff
where:
BHP = rated power in horsepower
HHP = hydraulic horsepower
eff = pump efficiency at rated capacity as a fraction
(Eq. 200-8)
Pump efficiency is determined empirically from the pump’s factory performance
test, and appears on the pump’s characteristic curve.

This section discusses the criteria for selecting a centrifugal pump for a specific
service. It is assumed that a centrifugal pump has been selected rather than a posi-
tive displacement pump. This material provides background information on
selecting a pump configuration for most applications in the petrochemical industry.
While this section provides general information for pump selection, engineering
judgement and user preferences must always be considered in the final decision.
Keep in mind that you are trying to minimize the sum of first cost, operating cost,
and maintenance cost for every selection. Also note the potential flexibility required
in operations and changes in environmental laws (which might require multiple
seals where a pump cannot accommodate it).
231 Factors in Pump Selection

General Pump Quality
Recommended practice is to specify that any pump in heavy-duty or critical service
be manufactured to API Standard 610. This includes all continuous-duty, process-
plant, hydrocarbon pumps and all other pumps in critical services (i.e., boiler feed-
water, off-plot charge pumps, high-pressure waterflood, etc.). Light duty pumps
(smaller than 150 HP and in noncritical services) are often purchased to meet ANSI
Standards or as general purpose pumps to supplier standards.
In practice, most pumping needs are met with single-suction, single-stage,
3600/1800 rpm centrifugal pumps. These are the work horses of the industry and are
generally the best choice for a given service. Historically, these have been hori-
zontal pumps. In recent years, however, single-stage, vertical, in-line pumps have
often proven to be as reliable and usually less expensive to purchase and install.
In all cases, the user should be consulted on proposed selections. There may be local
preferences based on past performance. Availability of maintenance and stocking of
interchangeable parts can also be significant factors.
ANSI versus API

There are significant construction and design differences between ANSI and API
pumps. These differences will impact the pump selection. A tabulation of major
differences is shown in Figure 200-22.
There are two major differences: pressure rating and materials of construction.
ANSI pumps are limited to 150# ratings. Also, ANSI pumps are not readily avail-
able with carbon steel casings or impellers. Cast iron or ductile iron are ANSI stan-
dard materials.
There are two limitations with use of cast or ductile iron.
• Cast or ductile iron castings (case and impeller) cannot be repaired by welding.
• Cast iron materials are susceptible to cracking due to thermal shock. When a
hot cast iron pump is exposed to cold extinguishing fluids it may crack. If the
pump was pumping a flammable or hazardous fluid, it could feed a fire or cause
other environmental hazards.
If ANSI pumps meet the required service conditions but cast or ductile iron mate-
rials are not acceptable, consider using 316 SS.
Fig. 200-22 Comparison of ANSI and API Pump Designs (1 of 2)

ANSI API
Type Pump and ANSI B73.1 for horizontal end suction top API 610 for horizontal single and multistage pumps,
Specification discharge pumps. vertical in-line, vertical single and multistage centrif-
ANSI B73.2 for vertical in-line pumps. All are ugal pumps.
single stage.
Maximum Allowable 275 PSIG Minimum 700 PSIG

Working Pressure Some API pumps are designed for
(MAWP) pressures above 5000 PSIG.
Hydrostatic Test 415 PSIG Minimum 1050 PSIG

Pressure API pump hydrostatic test pressure will be 1.5 times
the MAWP.
Flange Rating 150# flat faced is standard. 150# raised face is 300# raised face is standard. 600, 900, 1500, and
available. higher ratings are available if required by the service.
Maximum Temperature 250F 800F

Pump casing is foot mounted which limits allow- Pump casing is centerline mounted. No casing thermal
able thermal growth. growth limitations.
Materials of Ductile Iron Carbon steel casing is standard;

Construction (Casing and 316 SS stainless steel is also available.
impeller) Alloy 20 Impeller materials are cast iron, carbon steel, and
A carbon steel casing or impeller is not stainless steel.
commonly available.
Maximum Head 550 to 600 feet Practical limit is 10,000 feet.
Differential ANSI pumps are only single stage. Maximum Horizontal API pumps can have as many as 14
impeller diameter is about 13 inches. stages.
Impelller Design and Open impellers are common. Some enclosed All are enclosed design. Some open designs are avail-
Attachment impellers are available. able for special coke crushing services. Impellers must
No standard for attachment to the shaft. Most be key driven with a lock nut attachment.
are threaded on the end of the shaft.
Fig. 200-22 Comparison of ANSI and API Pump Designs (2 of 2)

Standard Dimensions ANSI pumps are built for interchangeability No standard dimensions apply.
between manufacturers.
Shaft Sleeves Not required but are available. Fit to the shaft Are required to prevent shaft damage in the seal or
and extension past the gland are not ANSI spec- packing area. Sleeve and stuffing box design is part of
ification requirements. the API 610 specification.
Lubrication Can be grease or oil lubricated. Oil lubrication is required. Usually ring oil system is
provided.
Thrust Bearing and Life Antifriction bearings only. Antifriction ball bearings must be duplex, single-row,
B-10 bearing life of 17,500 hours at design load 40-degree angular-contact type, installed back to
is required. back.
L-10 bearing life must exceed 25,000 hours at rated
conditions, or 16,000 hours at maximum axial and
radial loads at rated speed.
Wear Rings Not required and not available in most designs Case and impeller, front and back wear rings are
due to the use of open impellers. required. Wear ring clearances, attachment, and hard-
ness differential are specified.
Head/Capacity Considerations
The head-capacity requirement is a significant factor in selecting pumps. Proper
definition of these parameters requires considerable thought to be sure all possible
operating conditions have been considered. This is discussed in detail in
Section 130, “System Hydraulic Design.”
The performance of centrifugal pumps over a range of Heads and Capacities is a
function of the pump impeller and case design. There are three general impeller
designs: radial-flow, mixed-flow, and axial-flow (or propeller). These designs and
their relative performance are noted in Section 210, “Engineering Principles.”
Figure 200-15 indicates the general shape of the characteristic curves for radial,
mixed flow, and axial (propeller) pumps. It shows the head, brake horsepower, and
efficiency plotted as a percent of their values at the design, or best efficiency, point
of the pump.
The head curve for a radial flow pump is relatively flat, and the head decreases
gradually as the flow increases. Note that the brake horsepower increases gradually
over the flow range with the maximum normally at the point of maximum flow.
Mixed flow centrifugal pumps and axial flow or propeller pumps have considerably
different characteristics. The head curve for a mixed flow pump is steeper than for a
radial flow pump. The shut-off head is usually 150% to 200% of the design head.
The brake horsepower remains fairly constant over the flow range. For a typical
axial flow pump the head and brake horsepower both increase drastically near
shut-off.
The distinction between the above three classes is not absolute, and there are many
pumps with characteristics falling somewhere between the three.
Head-capacity ranges, and other pump features are shown in the Application
Guidelines (Figures 200-23 and 200-24) and on the Pump Description sheets in
Section 240.
Fig. 200-23 Horizontal Centrifugal Pump Application Guidelines
Fig. 200-24 Vertical Centrifugal Pump Application Guidelines
Although 3600/1800 rpm, single-stage pumps are the most popular selections in the
centrifugal pump family, the following factors may preclude their use.
High Head
When an installation calls for a high head combined with a low-flow rate (outside
the typical range of single-stage pumps), a high-speed, single-stage, vertical-in-line
pump should be investigated. If requirements exceed the limits provided by this
pump, a multi-stage centrifugal or positive displacement pump may be suitable.
Axially-split, horizontal, multi-stage pumps should be limited to approximately
2000 psig discharge pressure. Higher heads require double case or barrel pumps,
which are inherently more expensive. In special cases such as high-pressure pipe-
lines with limited NPSH available, pumps in series may be considered, but shaft
sealing becomes increasingly difficult as pump inlet pressures increase.
Some situations require vendors to develop a design for a particular service. For
example, the feed pumps in the Richmond Refinery ISOMAX TKN units were
designed to pump 1425 GPM of light hydrocarbons against an 8900 ft head at
300F. These pumps are radially-split, horizontal, 14-stage, 6600 rpm, and stretch
the vendor’s experience in design and operation for proven machinery. However,
“prototype” pumps are definitely not recommended. Consult a specialist in such
situations and always check the user’s list carefully when in doubt.
Low Head/High Flow

If a requirement calls for low head (50-200 ft) combined with a high pumping rate
(greater than 5000 GPM) that does not fall within the parameter range provided by
horizontal or in-line pumps, high-capacity pumps should be investigated.
There are also many double-suction pumps available that provide higher heads than
mixed-flow or axial-flow pumps. These are designed to move large quantities of
liquid without the usual high NPSH required by high-capacity suction pumps.
Typical services include transfer and loading pumps, ballast pumps, and cooling
water pumps.
Another pump type for very low heads in water service is the “Archimedes Screw
Pump.” The Company has almost no experience with these.
Physical Installation
In some cases, the physical arrangement of the installation is a significant factor in
pump selection. This is especially true when adding to existing facilities or retrofit-
ting a plant. For example, there may be limited space available, resulting in the
installation of a vertical, multi-stage, barrel pump where a horizontal pump with
fewer stages would be the first choice. This is also true for offshore platforms where
deck space is at a premium.
NPSH
Suction considerations often dictate pump selection. Cavitation can be of prime
concern if there is limited NPSH available or if suction lift is required. Pumps which
operate at low speed, have high Nss (suction specific speed), or have double suction
impellers require less NPSH. In certain cases, vertical-turbine barrel or self-priming

pumps may be the most reasonable solution. Vertical sump pumps can be used when
suction lift is required, if the head requirement is not too high.
Operating Temperature
Most pump installations operate at 250F or less, and pump design temperature is
normally not a problem. In high temperature situations (greater than 450F), atten-
tion must be given to pump materials and mechanical design, as they relate to the
stock and severity of service. Auxiliary cooling of bearings and seals is recom-
mended in most pumps starting at 300F, plus pedestal cooling at temperatures
above 500F. Some process pumps operate above 800F. Suggested bearing, seal,
and pedestal cooling arrangements are shown in API-610.
Three special design features needed for hot service:
1. An arrangement that permits piping and pump thermal expansion without
moving bearings out of line or imposing undue loads on them.
2. Corrosion-resistant materials suitable for the pumping temperature.
3. A design that minimizes leakage and confines it to avoid ignition and hazard to
personnel. Mechanical seals are used in almost all centrifugal pump services.
See Section 800, “Mechanical Seals.”
Hazardous Stocks
Special care must be given to installations handling toxic or hazardous stocks (H2S,
LPG, Ammonia, chlorine, HF, other acids, etc.) or hydrocarbons above their flash
point. In such cases, pumps that can take dual mechanical seals, or seals with
external flush should be considered. Pump materials must be carefully selected for
compatibility with toxic, hazardous, or corrosive stocks. Suggested seal flush
arrangements are also shown in API 610 and Section 800. Canned seal-less, and
hydraulic-seal pumps are available for low head/low HP applications. See
Section 150 for H2S considerations.
Dirty Fluids
Depending on the pumped fluid and its contaminants, some pumps will require
more frequent maintenance than others. This can be due to entrained solids (as in
crude oils, FCC cycle oils, sandy water, sludges, etc.) or the corrosivity of the fluid
itself.
Pumps with replaceable liners in the pump case are also available. Centrifugal
pumps in abrasive service should operate near the best-efficiency point to avoid
imbalanced hydraulic forces that accelerate wear.
When selecting pumps for such service, consider access to bearings and seals and
the pump itself. In such cases, consider pumps that can be disassembled without
disturbing connected piping (back pull-out feature), or that allow seal replacement
in place (cartridge seals).
Intermittent Operation
Centrifugal pumps are normally designed for continuous operation. If frequent shut-
downs are possible, the pump should remain flooded. If this is not possible, or
suction lift is needed, the seals must be flushed at startup. Canned pumps with
stock-lubricated bearings and pumps with close internal clearances must never be
run dry. Intermittent operation is generally harder on a pump than continuous
operation.
232 Energy Efficiency for Centrifugal Pumps

Operating costs account for a major portion of the total cost of ownership of pumps.
Small increases in efficiency (1–2%) can result in company-wide energy savings
amounting to several million dollars per year. Selection of the proper impeller size
and the proper number of stages can significantly affect pump efficiency. For all
centrifugal pumps, wear ring design, materials, and running clearances may improve
efficiency.
Impeller Considerations
Impeller disc friction is a major factor affecting overall efficiency. The outer
surfaces of a rotating impeller are subject to friction with the surrounding fluid.
Some of this friction is recovered as contribution to pump head if the rotating flow
induced by disk friction freely enters the pump casing. Wear ring leakage, on the
other hand, causes a radial flow which tends to reduce disk friction.
Disc friction effects are more evident in low specific speed (Ns) pumps. (Refer to
Section 218 for discussion of specific speed.) These pumps tend to have large diam-
eter, narrow shaped impellers as shown in Figure 200-15. Figure 200-25 shows the
typical variation of pump losses with Ns. For low Ns impellers (Ns < 1000), disc
friction accounts for 15% or more loss in efficiency.
Disc friction horsepower losses can be estimated as follows:
HP = 1.83(U/100)3 (D/10)2 (S.G.)(N)
Fig. 200-25 Factors Affecting Overall Pump Efficiency
where:
U = Peripheral velocity of impeller, Ft/sec
D = Outside diameter of impeller, inches
S.G. = Specific gravity of fluid at pumping temperature
N = Number of impellers
(Eq. 200-9)
Other calculation methods are available for determining disc friction losses but none
are precise because of the effect of other pump design details. For example, disc
friction losses increase as impeller-to-casing side clearances increase and as
impeller sidewall roughness increases. Losses are also affected by fluid viscosity.
For most pumps, this is generally an insignificant effect since fluid viscosity is typi-
cally low. (Refer to Section 219 for services where fluid viscosity is greater than
water.)
When pump suppliers offer a different number of stages for a specific pump appli-
cation, disc friction can clearly account for differences in quoted efficiency. Pump
supplier’s quoted number of stages will vary most often when the rated capacity is
less than 200 gpm or the head is more than 500 feet. Adding a stage or stages and
reducing impeller diameters may reduce losses and increase overall efficiency. The
addition of stages is not desirable from first cost and maintenance standpoints but
the operating cost incentive may more than offset maintenance aspects.
Wearing Ring Considerations

Similarly, wearing ring (also commonly called wear ring) clearances can signifi-
cantly affect efficiency. Figure 200-26 shows the effect of increasing wear ring
clearance on pump horsepower (efficiency). Most petrochemical pumps are
designed with impeller specific speeds in the range of 800–1500. As shown in
Figure 200-25, wear ring losses for a new pump in this Ns range typically average
only 3–4%. For low Ns impellers (Ns < 800), wear ring losses can account for much
larger losses (up to 15%) in efficiency. Generally there is little incentive to reduce
new wear ring clearances to a minimum. The likely efficiency savings is only 1–2%
with an increased risk of reduced reliability. (See Section 253.)
Fig. 200-26 Effect of Wearing Ring Clearance on Pump Horsepower
In service, wear ring clearances gradually increase due to corrosion, erosion, abra-
sion, etc. Consequently, efficiency decreases. Clearance increases of 100% or more
over as-built (new) clearances typically occur in a 2 to 3 year operating period. This
100% increase results in about a 5% decrease in pump efficiency. Sustaining as-new
clearances over long operating periods is much more beneficial from an efficiency
standpoint than reducing clearances to minimize losses when the pump is new.
Selection of proper wear ring materials is critical to minimizing efficiency losses
and maintaining long-term pump reliability. Section 253 discusses metallic and non-
metallic materials available for use in today’s pumps.
Trimming Impellers for Efficient Operation

Section 216 discusses the “affinity law” for changes in impeller diameter. This law
provides a reasonable estimate as long as impeller diameter changes are within 15%
of the original impeller diameter.
When the head developed by a single stage pump with constant speed driver is
higher than that actually required, the impeller diameter can be reduced. For multi-
stage pumps with constant speed drivers, one or more impellers can be removed.
This assumes that the lower head requirement is not a short-term operational condi-
tion. The required BHP is reduced directly with a reduction in head.
If the pump is driven by a steam turbine or variable speed motor, the speed can be
reduced to obtain the lower head. However, caution should be used since driver effi-
ciency may decrease and offset the benefit of the lower pump head.
There are two ways to trim impellers to achieve best efficiency. One way is to trim
only the vanes, leaving the shrouds (disc and cover) untrimmed. The second way is
to trim both the vanes and the shrouds to the same diameter. In addition to effi-
ciency considerations, machining costs, stress levels in unsupported shrouds, stress
levels at the vane-to-shroud joint, the effect on the shape of the performance curve,
thrust loads and seal cavity operating pressure need to be considered.
Industry practice for both enclosed and semi-open impeller designs is to trim both
vanes and shrouds to the same diameter. Exceptions to this practice include high
capacity pumps, mixed flow pumps, multistage diffuser pumps and certain pump
designs with pumping vane construction on the back shroud (disc).
For multistage diffuser pumps (typically double case types), suppliers often trim
only the impeller vanes. Leaving shrouds untrimmed helps guide the flow exiting
the impeller as it enters the narrow diffuser passage. There are stress limits which
set the amount of unsupported shroud which can be left untrimmed. A typical limit
for steel impellers running at 3600 rpm is ½-inch.
For volute pump designs (typically single stage and multistage, axially split case
types), industry practice is to trim both vanes and shrouds to the same diameter. In
general, there is no clear cut efficiency advantage to leaving the shrouds untrimmed
or to trimming them. Efficiency improvements afforded by added flow guidance
provided by the shrouds is approximately offset by the efficiency decrease due to
added disc friction. From a manufacturing standpoint, it is easier and less costly to
trim vanes and shrouds to the same diameter. Much more care needs to be taken if
only the vanes are trimmed. For example, profiling the vane-to-shroud intersection
to reduce stress concentrations is important when only the vanes are trimmed. (See
Figure 200-14.)
In high capacity, low speed volute design pumps, suppliers sometimes taper the
impeller trim from the front to the back shroud. This is done to reduce pressure
pulsations due to vanes passing the volutes. Vanes of double suction impellers
sometimes are profiled in a “V” shape for the same reason.
In a few specific cases, it may be advantageous to leave shrouds untrimmed for
other reasons. One reason may be the stability of the pump’s head-capacity curve.
(See Figure 200-8.) Also, having the shroud permits vanes to be restored to their
original diameter should future head requirements require it.
233 Special Service Pumps

“Magnetic Drive” or “Canned” Pumps for Hazardous Stocks
Stuffing boxes have been eliminated in designs called “magnetic drive” or “canned”
pumps.
Canned pumps have a special electric motor operating under pressure in a liquid-
filled chamber adjacent to the pump case. The motor chamber is filled with the
liquid pumped. The bearings are usually carbon, lubricated by liquid pumped. These
pumps are available in sizes up to 150 HP, 1500 GPM and 600 feet of head;
however, they cost considerably more than pumps with stuffing boxes or seals.
Magnetic drive pumps utilize standard horizontal electric motors which are coupled
to the pump bearing housing which supports a rotating magnet. The rotating magnet
rotates or “pulls” the impeller rotor supported by product-lubricated, carbon bush-
ings inside a sealed case. Like canned motor pumps, these are available in sizes up
to 200 HP, 2000 GPM, and 600 feet of head and cost considerably more than
conventional centrifugal pumps with seals.
The advantages of completely eliminating stuffing box or seal leakage have led to
many installations of these pumps in the Company, primarily in acid and hydrogen
sulfide services. However, performance has often been unsatisfactory, primarily
because of bearing wear from grit or lack of lubrication. Use these pumps only
where the liquid pumped is clean and lubricating, and the pumps are never run dry.
Propeller (Axial-Flow) Pumps

These pumps are used in high volume/low head services. Although available with
2 or 3 stages, most are low-speed, single-stage, vertical pumps. Typical applications
are sewage, waste-water lifting, and sump pump out. Lifting 30,000 GPM against
20 ft of head is typical.
Slurry Pumps
These units are in common use and handle abrasive slurries, sand, chemical sludges,
plant wastes, and similar products. They are generally low-speed and often are
rubber-lined, or cast from very hard materials.
Non-Metallic and Lined Pumps

Non-metallic and lined centrifugal pumps are available as a lower cost alternative to
pumps constructed of more expensive metallic alloys. Both types are horizontal
end-suction pumps designed to ANSI standards (ANSI/ASME B73.1M). They are
primarily used in acid, deionized water, and other highly corrosive chemical
services.
The wetted components of non-metallic pumps are generally manufactured of glass
filament reinforced plastic (FRP). The wetted components of lined pumps are gener-
ally manufactured of ductile iron and steel lined with Teflon (PTFE). Both types of
pumps are available in capacities to about 800 gpm and head to about 450 feet.
Non-metallic and lined pumps can be considered when the material class goes
beyond Alloy 20 (when metals such as nickel, hastalloy, or titanium are required).
They should only be considered when there are significant savings over the cost of
metallic pumps, or when there is no other practical pumping solution.
234 Application Guidelines

Figures 200-23 and 200-24 show several factors to consider in selection and appli-
cation of horizontal and vertical centrifugal pumps. As in selecting the pump cate-
gory, there is no straightforward, general procedure to follow in all cases. The
design factors are too numerous and often conflict. Consider the design factors most
important to your location and refer to the Application Guidelines for information
on those factors.
240 Centrifugal Pump Descriptions

This section illustrates and describes the most commonly used types of centrifugal
pumps.
Horizontal Centrifugal Pumps

1. Single Stage, API, top/end suction and discharge.
2. Single Stage, ANSI, end suction, top discharge.
3. Single Stage, ANSI, end suction, top discharge, self priming.
4. Single Stage, Double suction, axially split.
5. Multi-stage, API, axially-split case.
6. Multi-stage, API, radially-split case.
Vertical Centrifugal Pumps

1. Single Stage, In-line, ANSI, rigid coupling.
2. Single Stage, In-line, ANSI, integral-shaft.
3. Single-Stage, In-line, ANSI, flexible coupling.
4. Single Stage, In-line, high-speed
5. Single Stage, Sump, bearing supported.
6. Single Stage, Sump, overhung impeller.
7. Multi-Stage, Vertical-Turbine, barrel.
8. Multi-Stage, Vertical-Turbine, deep well.

Pump Description Centrifugal – Horizontal – single-stage (top/end suction and top
discharge) typical API 610 class pump (See Figure 200-27.)
Typical Service Continuous-duty refinery process and critical water service.
Typical Head/Capacity Range 50-800 ft/100-10,000 GPM
Max Allowable Temperature 350F without cooling
500F with Bearing Cooling
800F with Bearing Cooling and Pedestal Cooling
Typical Speed Range Up to 3600 rpm
Construction Features Cast steel and alloy available. Available single or double suction.
Normally closed impellers. Oil lubrication. Packed, single or multi-
seals. Radially split. Centerline mounted. Back pullout for maintenance
with single suction. Ductile iron or cast iron casings are not available.
Typical Control Method Throttled discharge on flow, level, or pressure control.
Advantages More rugged and reliable than ANSI or Industry Standard pumps.
Available in a wide range of pressures and capacities. Lower operating
costs since efficiency is usually higher. Available in overhung design up
to 900 HP.
Disadvantages and Limitations Most expensive standard centrifugal pump.
Specification PMP-MS-983/API 610.
Data Sheet API 610, Appendix B.
Fig. 200-27 Horizontal, Single-stage, Top/end-suction, Top-discharge, API 610 Class Centrifugal Pump Courtesy of
Peerless Pump Co.
Pump Description Centrifugal – Horizontal - single-stage. ANSI B73.1 (end suction,

top discharge) (See Figure 200-28.)
Typical Service Chemical. Water. Noncritical hydrocarbon. General purpose.
Typical Head-Capacity Range 50-600 ft/50-3500 GPM
Max Allowable Temperature 250F recommended
Construction Features Standard material options for the pump casing and impeller are cast
iron or ductile iron, 316 series stainless, and Alloy 20. Carbon steel is
not standard or readily available. Always end suction/top centerline
discharge with overhung impeller. Open or closed impellers available.
Ball bearing grease or oil lubricated single, tandem, or double seals
available. Foot-mounted casing. Back pullout for maintenance.
Advantages For each size, ANSI pumps are dimensionally interchangeable from
any manufacturer. Less expensive than API pumps. Wide variety of
alloy construction materials available.
Disadvantages and Limitations 150 HP maximum recommended. Carbon steel case is generally not
available. Pressures limited to 275 psig @ 60F.
Specification ANSI B73.1. See also PMP-PC-1241 in this manual.
Data Sheet PMP-DS-1241-H.
Fig. 200-28 Horizontal, Single-stage, End-suction, Top-discharge ANSI Class Centrifugal Pump Copyright 1995
Ingersoll Dresser Pumps. Worthington is a trademark of Ingersoll Dresser Pump Company.
Pump Description Centrifugal – Horizontal – single-stage. ANSI B73.1 (end

suction, top discharge) self-priming (See Figure 200-29.)
Typical Service For vertical lift when non-pulsating flow desired. Sump pumpout.
Tank car unloading.
Typical Head/capacity Range 150-250 ft/0-1000 GPM
Max Allowable Temperature 250F Recommended
Construction Features Same as ANSI Horizontal
Typical Control Method Throttled discharge, on/off level control.
Advantages Up to 20 ft effective static lift. Eliminates need for foot valve.
Dimensionally interchangeable with all ANSI pumps. More reli-
able than submerged vertical sump pumps.
Disadvantages and Limitations Less efficient than standard nonself-priming pumps. May take too
long to prime on large suction lines. A mechanical seal may run dry
without an external flush.
Company Specification ANSI B73.1. See also PMP-PC-1241 in this manual.
Company Data Sheet(s) PMP-DS-1241-H.
Fig. 200-29 Horizontal, Single-stage, Self-priming, ANSI Class Centrifugal Pump Courtesy of Goulds Pumps, Inc.
Pump Description Centrifugal – Horizontal – single-stage. (double suction, axially

split) (See Figure 200-30.)
Typical Service Cooling water circulation. Fire pump. Cargo loading. Crude transfer.
Construction Features Typically cast iron or bronze case (steel case for HCs) and bronze
trim. External sleeve or anti-friction bearings. Horizontal inlet and
outlet. Closed impellers. Also available with stainless steel impellers
for higher cavitation resistance
Typical Control Method Throttled discharge, system back pressure (cooling water).
Advantages Balanced thrust on shaft. Can maintain pump in place. Low NPSH
requirement. Wide range of sizes and capacities.
Disadvantages And Limitations More expensive than single suction, overhung pump design. Suction
lines must be carefully designed to avoid nonsymmetrical flow that
would channel to one side, resulting in unbalanced thrust and possibly
cavitation.
Specification PMP-MS-983/API 610 (hazardous, flammable, and special purpose
services).
See also PMP-PC-1241 in this manual (general purpose services).
Data Sheet API 610, Appendix B (hazardous and flammable services).
Fig. 200-30 Horizontal, Single-stage, Double-suction, Axially (Horizontally)-split Case, Centrifugal Pump
Courtesy of Goulds Pumps, Inc.
Pump Description Centrifugal – Horizontal – multi-stage. API 610 axially split

(See Figure 200-31.)
Typical Service Crude feed. Waterflood. Boiler feedwater. Process. Pipeline.
Typical Head/Capacity Range 200-7000 ft/100-5000 GPM
Max Allowable Temperature 250F without cooling
400F with Cooling
Construction Features Carbon steel case. CI, steel, stainless steel, or bronze impellers.
Between bearings. Horizontal nozzles, both suction and discharge
nozzles located in bottom half casing.
Advantages Ease of in-line assembly and inspection. Can be designed with
balanced axial thrust. Eliminates multiple in-line series pumps.
Disadvantages and Limitations API 610 limits the axially-split case design to applications below
400F and pumped fluids with specific gravity above 0.70. More
complex than single-stage pumps. However, note that pressures to
2000 psig are common in producing water flood applications.
Fig. 200-31 Horizontal, Multi-stage, Axially (Horizontally)-split Case Centrifugal Pump

Courtesy of Flowserve Corporation
Pump Description Centrifugal – Horizontal – multi-stage. API 610 radially split

double case (high pressure, high temperature)
Typical Service High pressure process feed pumps. Boiler feedwater. Crude pipeline.
Typical Head/Capacity Range 0-10,000 ft/100-5000 GPM
Max Allowable Temperature 850F w/pedestal, bearing and seal cooling
Typical Speed Range 1800 - 7000 rpm
Construction Features Usually top suction/discharge; however, nozzle location may vary
with installation requirements. Radially split. Double casing. Carbon
steel cases. Water-cooled pedestals, bearings and seals available.
Typical Control Method Spillback on external-flow control.
Advantages Pressures possible without series pump operation. Double casing
allows in-line assembly/disassembly.
Disadvantages and Limitations Clearances extremely sensitive to differential temperatures in pump.
Slow pump startup mandatory with hot pumps. Proper assembly
difficult with many stages. Very important not to run with blocked
discharge.
Fig. 200-32 Horizontal, Multi-stage, Radially (Vertically)-split, Double Case, Centrifugal Pump Copyright 1995
Ingersoll Dresser Pumps. Pacific is a trademark of Ingersoll Dresser Pump Company
Pump Description Centrifugal – Vertical – single-stage - in-line. ANSI B73.2– Rigid

Coupling (See Figure 200-33.)
Typical Service Chemical. Water. Noncritical hydrocarbon. General purpose.
Typical Head/Capacity Range 600 ft/3000 GPM
Construction Features DI/CI, stainless steel or alloy available; steel not available. Motor
supported by pump. Suction/discharge flanges with common center-
line which intersects shaft axis. Open or closed impellers. Motor
bearings carry pump loads.
Advantages Can remove seal and impeller without disturbing motor. Unit is inter-
changeable with all other vertical ANSI designs. Simpler and cheaper
to install than horizontal. Occupies less floor space. No field align-
ment of pump and motor needed (as long as fits remain within toler-
ance).
Disadvantages and Limitations 150 HP maximum recommended. Typically cannot install dual
mechanical seals. Vapor or gas in liquid tends to collect at mechanical
seal faces, promoting failure unless properly vented during startup
(needing a vent), and flushed during operation. Always single-stage.
Rigid couplings are troublesome to keep in alignment (causes short
bearing and seal life typically).
Company Data Sheet(s) PMP-DS-1241-H.
Fig. 200-33 Vertical, In-line, Single-stage, Rigid-coupled, ANSI Class Centrifugal Pump
Pump Description Centrifugal – Vertical – single-stage - in-line. ANSI B73.2 Inte-

gral Shaft (no Coupling) (See Figure 200-34.)
Typical Service Chemical. Water. Non-critical Hydrocarbon.
Typical Head/capacity Range 600 ft/3000 GPM
Construction Features Motor shaft is integral with pump shaft. All bearings are in the
motor—none in the pump. DI/CI, 316 stainless steel, and alloy 20 are
standard materials; carbon steel is not available.
Typical Control Method Throttled discharge on flow, level, or pressure control
Advantages Unit is interchangeable with all other vertical ANSI designs. Simpler
and cheaper to install than horizontal pump. Occupies less floor
space. No field alignment of pump and motor needed. Provides better
seal and bearing life than rigidly coupled in-line.
Disadvantages and Limitations Must remove motor for access to seal or impeller. Cannot accommo-
date dual mechanical seals.
Data Sheet PMP-DS-1241-H.
Fig. 200-34 Vertical, Single-stage, In-line, Integral Shaft, ANSI Class Centrifugal Pump
Pump Description Centrifugal – Vertical – single-stage - in-line. ANSI B73.2

flexible coupling (See Figure 200-35.)
Typical Service Chemical. Water. Noncritical Hydrocarbon. General purpose.
Typical Head/Capacity Range 600 ft/3500 GPM
Construction Features Pump has own bearings. Otherwise, same as rigid coupling pump.
Advantages Field alignment of pump and motor shafts is maintained by register
fits. Hydraulic loads not carried by motor bearings. Can remove seal
and impeller without disturbing motor. Interchangeable with all other
vertical ANSI designs. Simpler and cheaper to install than hori-
zontal. Occupies less floor space. No field alignment of pump and
motor needed.
Disadvantages and Limitations Complete bearing bracket/pump rotor must be sent to shop for seal
repairs. More expensive than rigid coupling or integral shaft pumps;
otherwise, same as rigid coupling pump. Taller and heavier installed
height than other vertical in-line options.
Data Sheet PMP-DS-1241V.
Fig. 200-35 Vertical, In-line, Single-stage, Flexible-coupling, Centrifugal Pump Copyright 1995 Ingersoll Dresser
Pumps. Worthington is a trademark of Ingersoll Dresser Pump Company
Pump Description Centrifugal – Vertical – single-stage – in-line. High-speed

Typical Service High head/low flow service for water and HC.
Typical Head/Capacity Range 0-4500 ft/0-800 GPM
Max Allowable Temperature 400F with cooling
Typical Speed Range Up to 15,000 rpm
Construction Features Integral gear box with self-contained lube system. Available in
carbon steel, stainless steel or alloy. Built-in seal flush, dual seals
available. Usually open impellers. Suction/discharge flanges with
common counterline intersecting shaft axis.
Typical Control Method Minimum flow bypass with flow control.
Advantages Less expensive to purchase and install than comparable moderate
high-pressure horizontal, centrifugal, and plunger pumps. Field align-
ment of pump/motor not required. Occupies less floor space than
equivalent horizontal or P.D. pumps.
Disadvantages and Limitations Special prelube system for higher suction-pressure applications.
Separate minimum flow bypass with controller for each pump. High
speed creates seal face problems. Vapor collecting at top of case can
cause seal failure if not flushed. Accidental reverse rotation can
loosen impeller and cause failure. Must dismantle to replace seals.
May have unstable performance curve at low flows. 400 HP upper
limit. More NPSH(R) and much less efficient than equivalent hori-
zontal pumps. Better metallurgy required for impeller/diffuser due to
sensitivity of performance vs. internal clearances. There are
numerous ports (seal flush, vents, etc.) which are complex, and must
be carefully piped up.
Fig. 200-36 Vertical, In-line, Single-stage, High-speed Centrifugal Pump

Courtesy of Sundstrand Fluid Handling Company.
Pump Description Centrifugal – Vertical – single-stage. Sump - bearing supported

Typical Service Sump pumpout. Sewage. Nonabrasive solids. Sludge.
Typical Head/Capacity Range 20-250 ft/50-2000 GPM (@ 30 ft)
Max Allowable Temperature 250F
Construction Features Typical C.I. Plastic and 316 stainless steel available. Optional line
shaft bearing flush. Open or closed impellers.
Typical Control Method On/Off level control and throttled discharged.
Advantages Simple mounting; no foundation. No stuffing box or seal leakage.
Submerged impeller.
Disadvantages and Limitations 20 ft shaft is practical limit. Less reliable than self-priming hori-
zontal or vertical cantilever pump. Line shaft bearings require lubri-
cation from one of the following: (1) grease, (2) Continuous
Water/pumped Fluid Injection, (3) Pumped Fluid.
Data Sheet PMP-DS-1241-V
Fig. 200-37 Vertical, Sump, Single-stage, Bearing-supported Centrifugal Pump Courtesy of Goulds Pumps, Inc.
Pump Description Centrifugal – Vertical – single-stage. Sump - Pump Cantilever

Impeller and Shaft (See Figure 200-38.)
Typical Service Sump pumpout. Sewage. Abrasive solids, sludge, slurry
Typical Head/Capacity Range 0-200 ft/0-5000 GPM (@ 100 ft)
Speed Range Up to 1800 rpm
Construction Features Typical C.I. Plastic and 316 stainless steel available. Open or closed
impellers. No bearings in pumped liquid. Large-diameter shaft to
support cantilevered impeller.
Typical Control Method On/off level; throttled discharged
Advantages No bearing/pumped liquid contact. More reliable than bearing
supported vertical sump pumps in abrasive or sludge service. Simple
mounting; no foundation. No stuffing box or seal. Submerged impeller.
Pump can run dry for short periods.
Disadvantages and Limitations 10 ft is practical shaft limit. Requires rigid, large diameter shaft. More
expensive than bearing supported vertical sump pump.
Fig. 200-38 Vertical, Sump, Single-stage, Cantilever Impeller and Shaft Centrifugal Pump Copyright 1995 Ingersoll
Dresser Pumps. Worthington is a trademark of Ingersoll Dresser Pump Company
Pump Description Centrifugal – Vertical-turbine – multi-stage. Barrel

Typical Service Low NPSHA applications. Boiler feed water. Flashing liquid.
Condenser hotwells.
Speed Range Up to 3600 rpm; however, 1800 rpm is the preferred maximum speed
for improved reliability.
Construction Features Open or closed impellers. Steel barrel and steel or C.I. head with
typically C.I. bowls and C.I. steel or bronze impellers. Barrel
designed for discharge pressure. Can be installed in a sump without
the barrel. Weight of pump and pump thrust taken by motor thrust
bearing.
Typical Control Method Throttled discharge, flow, pressure or level control.
Advantages Little floor space required. Low NPSH required. Typically high effi-
ciency.
Disadvantages and Limitations Shaft sleeve bearings exposed to pumped liquid. Must remove pump
for all maintenance except mechanical seal changes.
Specification PMP-MS-983/API 610 (hazardous, flammable and special purpose
services).
See also PMP-PC-1241 in this manual (general purpose services).
Data Sheet API 610, Appendix B (special purpose); PMP-DS-V (general
purpose).
Fig. 200-39 Vertical, Multi-stage, Barrel, Centrifugal Pump Courtesy of Flowserve Corporation
Pump Description Centrifugal – Vertical-turbine – multi-stage. Deep well line-shaft

type (See Figure 200-40.)
Typical Service Potable and irrigation water wells. Platform. Seawater. Firewater.
Typical Head/Capacity Range 0-3500 ft/60,000 GPM
Construction Features Column pipe and shaft in 10 ft lengths can be provided with open or
enclosed oil lubricated shaft. Open lineshaft arrangement is preferred;
less expensive and easier to assemble/disassemble. Semi-open or closed
impellers. Weight of shaft and hydraulic thrust supported by vertical
motor bearings. Can be provided with engine driver with right angle
drive. Same materials as vertical turbine (barrel).
Typical Control Method Throttled discharge. Level control for sumps.
Advantages Typically high efficiency. Can be installed in wells or wet-pit sumps.
Disadvantages and Limitations Size limited to diameter of well casing. Practical maximum setting
depth 1000 ft. 1800 maximum rpm. Bowl bearings are process-fluid
lubricated. Abrasives will shorten pump life.
Available With Submersible Motors To Eliminate Long Drive Shafts;
However, Submersible Motor Installations Are Less Reliable And Are
Not Recommended Above 50 Hp.
Fig. 200-40 Vertical, Multi-stage, Deep-well (Vertical Turbine) Centrifugal Pump Courtesy of Flowserve Corporation
250 Mechanical Components

This section considers the mechanical design of the principal parts of centrifugal
pumps. See Section 260 for related subsystems.
The principal parts of centrifugal pumps include:
1. Cases
2. Impellers
3. Wearing Rings
4. Shafts and Shaft Sleeves
5. Throat Bushings
6. Lantern Rings
7. Glands
8. Balance Drums
9. Bearings
10. Base Plates
11. Couplings
12. Coupling Guards
Figure 200-41 is a cross-section showing most of the principal parts of a typical
centrifugal pump.
Fig. 200-41 Cross-section of a Typical Centrifugal Pump Courtesy of Peerless Pump, Inc.
251 Cases
The case is that part of the pump in which energy imparted by the impeller is
converted into pressure. Pump cases are either axially (horizontally) split or radially
(vertically) split. Although axially-split cases were common on all types of pumps
for many years, they are now used principally in high-flow and multi-stage designs.
Pumps with the most common head and capacity ranges are radially split.
Diffuser vs. Volute Construction

Single-stage centrifugal pumps are usually “volute” type. Multi-stage pumps are
either diffusion-vane or volute. The diffusion-vane or diffuser type incorporates in a
cylindrical case a stationary ring of vanes around the periphery of each impeller.
Diffusion-vane pumps are widely applied in boiler feed and in high-head, hot oil
services. In general, however, volute construction is preferred.
The vane angle for either volute or diffuser, if properly designed, is correct for only
the capacity at the best efficiency point. If the pump is operated at some other
capacity, the diffuser may act as a hindrance rather than as an aid to efficient
operation.
In recent designs, the efficiency of the volute type is equal to or better than the
diffusion-vane type. Further, the diffusion-vane type is more difficult to reassemble
after dismantling for maintenance. In some services, however, diffusion-vane pumps
are preferred because of space considerations. One such use is for pumping deep
wells of small diameter. All centrifugal deep-well pumps are “turbine” type pumps
with diffusion vanes as an integral part of the case.
The volute-type pump presents one problem not found in a diffusion-vane pump:
radial force against the shaft caused by unbalanced pressure conditions in the volute.
The radial force is greatest at shutoff and least at maximum efficiency. This radial
force must be compensated for by using a stiff shaft or placing a second volute
throat on the opposite side of the shaft. This “double-volute” construction is
provided on many heavy-duty process-type pumps, 3- to 4-inch discharge size or
larger.
Centerline vs. Bracket vs. Foot Mounting

In horizontal single-stage, centrifugal pumps (commonly used for process services),
two case arrangements have been available – the bracket-mounted type shown in
Figure 200-42 and the centerline-mounted type shown in Figure 200-43. Points in
favor of centerline-mounted pumps are:
1. Piping stresses are transmitted more directly to the foundations and are less
likely to cause misalignment and distortion of the pump.
2. Piping and driver can be left in place while the complete rotating element,
including the bearing housing and stuffing box, is removed for repairs. This is
called the “back pullout” feature.
3. Some pumps are designed with larger impeller eye areas which need less
NPSHR than bracket-mounted pumps for the same operating conditions.
4. Centerline-mounted pumps in accordance with API 610 generally have heavier

construction with greater case thickness, heavier shafts, heavier bolting, and
high design pressures.
Fig. 200-42 Bracket-mounted Centrifugal Pump Fig. 200-43 Centerline-mounted Centrifugal Pump Cour-
tesy of Flowserve Corporation
While centerline mounted pumps provide superior support for heavy-duty service
and high temperatures (greater than 250F), foot-mounted pumps are less expensive
and suitable for low temperature, lighter-duty service.
Typical foot-mounted pump casings are shown in Figures 200-28 and 200-29. These
pumps are not suitable for temperatures above 250F because all the casing thermal
expansion is from the casing base or feet toward the discharge nozzle. The thermal
growth will cause misalignment between the rotor/impeller and the case.
Vertical In-line
Single-stage, single-suction centrifugal pumps are also made in a vertical in-line
design as shown in Figure 200-44. The “in-line” pump is increasingly used in a
variety of services including process plant services. The pump case is flanged
directly in the line and a vertical motor is supported by the pump. The in-line pump
offers the following advantages over a comparable horizontal pump:
1. Lower initial cost because there are fewer parts, no fabricated base plate, no
pump bearing housing on some designs (bearings are in the motor) and no flex-
ible couplings or coupling guards.
2. Lower installation costs because the foundation is smaller or not needed at all
and the piping is simplified.
3. Lower maintenance cost because the pump has fewer parts and is permanently
aligned with its driver.
4. Occupies less plot or deck space.
In-line pumps are generally applicable for temperatures to 250F, flows to
3000 GPM, and heads to 600 feet. They usually have mechanical seals to seal the
shaft but can also be obtained with packing. A special high speed in-line, the
Sundyne pump (discussed below), is rated for temperatures to 400F and can
produce 4500 feet of head at lower flow rates.
Fig. 200-44 Vertical In-Line Pumps Courtesy of Flowserve Corporation
There are three basic types of shaft coupling designs for in-line pumps:
1. Flexible spacer-coupling
2. Integral or close-coupled
3. Axially-split rigid coupling
Flexible coupling allows changing the mechanical seal without removing the motor.
Alignment of motor and pump shafts are maintained by register fits. In addition, the
pump has its own bearings which minimizes shaft deflection at the seal for
improved seal life.
The close-coupled type has the significant advantage of built-in alignment and a
short stiff shaft. The close-coupled pump has one disadvantage: the motor and
impeller assembly must be lifted and removed to change the seal. Unless being able
to change the seal without removing the driver is of primary importance (as in some
remote locations with larger drivers) and operation is always near peak efficiency,
the close-coupled or integral shaft design is recommended.
The bearings for rigidly coupled and integral shaft-type in-line pumps are in the
motor. These motor bearings should be checked to see that their design life meets
Company specifications. This is particularly important on high suction pressure
services where up thrust may be quite high.
Axially-split rigid-coupled in-line pumps are the least reliable due to the inability of
getting and maintaining proper alignment between the pump and motor shaft,
resulting in vibration. Company experience with these pumps is very poor.
In addition, in-line pumps with rigid couplings have greater shaft deflection at the
seal than integral or flexibly-coupled pumps, leading to reduced seal life. The
advantage of rigid-coupled in-line pumps is the motor does not need to be removed
for seal maintenance. In-line pumps are made by almost all major centrifugal pump
manufacturers. The Company has had good experience with most of these pumps,
except rigid-coupled pumps with mechanical seals.
The Sunstrand Corporation manufactures the Sundyne in-line pump designed to
give high heads at relatively low flows. This pump uses a single impeller rotating at
speeds up to 15,000 rpm by using a gear box between the motor and the pump.
Currently, maximum brake horsepower is limited to 400 BHP. This pump will
deliver 1500 feet of head at 400 GPM and about 4000 feet of head at 100 GPM.
By using speed changes and various diffuser and impeller configurations to satisfy a
wide variety of operating conditions, Sunstrand has been able to achieve a high
degree of standardization. This allows rapid delivery of new pumps and replace-
ment parts. The head-capacity curve is usually very flat and drooping with a rapid
falloff at the cutoff point. In some sizes an internal hydraulic instability may cause
vibration at low flows. The manufacturer’s limit for minimum allowable flow
should be considered as it may be higher than for conventional centrifugal pumps.
Many problems have been experienced with these pumps because they have oper-
ated at low flows without having minimum flow protection. The cost of minimum
flow bypass facilities should be taken into account during bid evaluation.
252 Impellers
The shape and size of the fluid channels in the impeller and casing give the liquid
the required velocity, speed and direction. Impellers require more attention to
hydraulic design than mechanical design. This discussion will be limited to mechan-
ical considerations.
Petroleum industry pumps have impellers usually cast of iron, steel, or bronze. The
impellers come from the foundry rough, out of balance, and not strictly alike,
although made from the same patterns. To reduce friction losses, all impeller
surfaces should be as smooth as possible, particularly those of the fluid channels
bounded by the shrouds and vanes. Since the fluid channels are of irregular shape,
the impeller may require more hand work than the remainder of the pump. The
strength of the material used and speed of rotation will limit the maximum safe
diameter of the impeller. This limit is determined by the pump manufacturer.
Impellers can be open or closed type, or an intermediate type, usually referred to as
“open,” which is really a “semi-open” impeller. An open impeller consists of a hub
and several straight or curved vanes. A closed impeller has plates or shrouds on
each side of the vanes in the planes of rotation. The semi-open impeller, as well as
the one usually called an “open” impeller, has a shroud on one side only. This type
of impeller is recommended when the liquid pumped may plug a closed impeller.
These impellers ordinarily are not as efficient as closed impellers.
Impellers are fastened to the shafts in various ways. Some are threaded to the shaft
so that the rotation of the shaft tends to keep them tight. Most are keyed to the shaft
with either a cylindrical or a taper fit and are secured with a nut. Keyed construc-
tion is preferred over screwed, because screwed impellers are more susceptible to
coming loose and causing damage if the pump is run backwards. This is common
during startup, following repairs where the electrical leads were disconnected.
Balancing is the final step in manufacturing an impeller. Until recently, manufac-
turers were content to balance impellers by the static method. To be completely
balanced, however, an impeller should usually be dynamically balanced. While
dynamic balancing is preferred, the extra cost may not be justified for small pumps
and pumps operating at lower speeds. Dynamic balancing should be considered in
the following cases:
• All multi-stage pumps.
• Critical pumps operating at 1700 rpm and above, with impellers 8 inches or
more in diameter.
Some pump specifications call for dynamic balancing of the complete rotating
element after individual balancing of each impeller and balance drum. Except for
large multi-stage pumps, this requirement is expensive and usually unnecessary.
Since elements of the rotating assembly, except the impeller are concentric shapes,
if the impellers are dynamically balanced there is little to cause dynamic unbalance.
Dynamic balancing should be done with a half key for key driven impellers.
253 Wearing Rings

Wearing rings are usually in pairs, one stationary, one rotating. The rotating ring is
attached to the impeller; the stationary ring is concentric with the impeller wearing
ring but seated in the casing. The primary purpose of these rings is to minimize
internal leakage from the discharge back to suction. In well-designed pumps of
moderate size, this leakage is about 5% of the total liquid passed through the
impeller. (See Figure 200-41.) The less the wearing ring clearance, of course, the
less the internal leakage and the higher the pump efficiency. (See Section 231.)
However, wearing ring and pump seizure can result from too close clearances.
Wearing ring trouble may be due to any of the causes listed below. Extra wear ring
clearance may prevent these problems:
1. Distortion of pump case from pipe stresses or from improper warmup proce-
dure, causing contact between the wearing rings.
2. Lodging of hard foreign bodies between wearing rings.
3. Deflection of the shaft, causing contact between the wearing rings.
4. Unbalance in the rotating element.
5. Eccentric fit due to improper machining and/or assembly.
6. Thermal transients which cause loosening of the fit and eventual wear ring
movement.
7. Galling due to improper wear ring material combination.
Be cautious of manufacturers who claim unusually high efficiencies. They achieve
these (claimed) efficiencies by very tight wear-ring clearances. Unfortunately, expe-
rience shows many rubs and seizures occur during run-in of new pumps.
Shaft deflection is due in most cases to unbalance of the rotating element caused by
hydraulic side thrust in the volute, unbalanced impellers, or both.
Important wearing ring material properties include machinability, wear resistance,
gall resistance, corrosion resistance, and thermal expansion. Ring stretch due to
centrifugal forces is also a consideration. Wearing rings are designed with running
clearances and normally do not touch. In many pumps, however, on occasion
wearing rings do come in contact. Continuous heavy contact will likely lead to over-
heating, galling, and seizure. Light occasional contact will not cause serious trouble
if the rings are of non-galling materials that can operate to a limited extent as bear-
ings when lubricated by the stock pumped.
Metallic wearing rings have been used in pumps almost exclusively. Very recently,
non-metallic materials have been installed, usually in combination with a mating
metallic ring. When the mating rings are both metallic, they should be of different
hardnesses and preferably of different materials. A minimum of 50 Brinell hardness
difference is recommended for wearing rings of the same material. Otherwise, the
similar metallic materials may gall and seize. Exceptions to this hardness difference
requirement are mating cast iron rings and mating bronze rings. These materials are
gall resistant.
Figure 200-45 gives wearing ring clearances recommended by API Standard 610.
These clearances are for process pumps operating at temperatures below 350F with
metallic wearing rings of materials that are non-galling, such as cast iron, bronze,
and 11–13% Chromium steel with adequate hardness difference.
Note For pumps operating above 350F with metallic rings, an additional
0.002 inch clearance for each 100F above 350F should be added to the values
shown in Figure 200-45. Also, add 0.005 inch clearance for pumps with wearing
ring materials with galling tendencies.
Fig. 200-45 API-610 Minimum Wearing Ring Clearance for Metallic Rings
Courtesy of the American Petroleum Institute.
Wearing Ring Diameter, in. Diametral Clearance, in.
<2 0.010
2.000 – 2.499 0.011
2.500 – 2.999 0.012
3.000 – 3.499 0.014
3.500 – 3.999 0.016
4.000 – 4.999 0.016
5.000 – 5.999 0.017
6.000 – 6.999 0.018
7.000 – 7.999 0.019
8.000 – 8.999 0.020
9.000 – 9.999 0.021
10.000 – 10.999 0.022
11.000 – 11.999 0.023
Note: For non-galling materials and pumps operating below 350°F.
Use of some of the newer non-metallic wear ring materials offers opportunities for
improved reliability, especially in services where frequent startup occurs, dry
running occasionally occurs, or rotor radial deflection is high.
Some of the newer non-metallics in use in pump applications include thermoplas-
tics (PEEK, Torlon, and composites) and metallized graphite. Advantages of these
materials include good wear resistance combined with excellent anti-galling charac-
teristics. Several are self-lubricating. These characteristics often allow new radial
clearances to be set tighter than conventional metallic wear rings. However, careful
consideration needs to given to thermal expansion characteristics of these materials.
In most cases, thermal expansion is anisotropic. This means that the thermal expan-
sion coefficient is different in each direction (along its width, length, and thick-
ness.) Published materials data frequently lists properties in one direction only.
Anisotropic characteristics must be carefully considered when setting wear ring
clearances to avoid heavy contact between mating rings and subsequent ring failure.
Also, the effect of centrifugal stress on wear ring strain (expansion) must be consid-
ered where non-metallic materials are used for impeller wear rings.
Disadvantages of non-metallics include higher material costs (typically 1½ to 3
times that of conventional metallic wear rings), special handling to avoid breakage
during installation, difficulty keeping rings secured in place, and limited tempera-
ture capability. For improved reliability and reduced first cost, the usual choice is to
use non-metallic case wear rings running against a conventional metal impeller wear
ring. Use of proper methods to secure wear rings to pump casings and to pump
impellers is also important to success when using non-metallic wear ring materials.
Thermal properties and temperature limitations of non-metallic materials require
careful consideration as well. Many thermal plastics are limited to temperatures
below 350°F. Materials such as PEEK are suitable up to 500°F. Metallized graphite
is the only material suitable for temperatures above 500°F. Non-metallics may also
have limited dry running capability due to poor thermal conductivity. Finally,
thermal diffusivity differences between non-metallics and metals make non-metal-
lics a poor choice for services where significant process thermal transients are likely
to occur.
254 Shafts and Shaft Sleeves

A shaft must be large enough to transmit the necessary energy to the liquid being
pumped. It must also have strength to resist deflection from hydraulic thrust of the
liquid in the pump case. Horizontal shafts must also be able to carry the weight of
rotating parts. These factors all call for a large shaft diameter consistent with the
strength of material, the distance between the bearings, and the distance from the
bearings to the impeller. However, increasing the diameter may increase the
entrance velocity at the impeller eye and result in more NPSHR. Further, the larger
the shaft diameter, the larger the bearings, stuffing box area and other parts,
resulting in higher pump costs. The desirable design is one which has adequate shaft
stiffness, but not an uneconomically large shaft.
Packing and Shaft Sleeves

The only part of the shaft subject to wear lies in contact with the packing in the
stuffing box. Since the packing must be kept in close contact with the shaft and
sides of the stuffing box, there is rubbing pressure on the shaft which in time will
cause wear. On all but the smallest pumps, the shaft is an expensive part of the
pump, so it is common practice to fit removable sleeves over the shaft in the
packing area. These sleeves may be readily replaced when worn.
A shaft sleeve must be hard, smooth and, often, corrosion-resistant. The sleeve
should prevent leakage between the shaft and the sleeve. The sleeve should extend
beyond the gland so that if leakage does occur, the operator can distinguish leakage
under the sleeve from leakage through the packing. Manufacturers are usually
willing to meet this requirement, and it should rarely be necessary to accept a pump
with a short shaft sleeve.
A gasket is usually provided in the design to prevent leakage between the shaft
sleeve and the shaft. Sometimes a lap joint used in place of a gasket is satisfactory if
given proper care during installation.
Many sleeves in severe services have their wearing surfaces covered with Stellite #1
or Colmonoy, with a hardness of 450 to 500 Brinnell, or covered with tungsten
carbide (Wallex 55).
255 Throat Bushings and Lantern Rings

Throat bushings as shown in Figure 200-46 are mechanical restrictions between the
stuffing box and impeller, installed to raise the pressure in the stuffing box and to
prevent packing from extruding into the pump case.
Throat bushings are desirable because, when wear occurs, they can be replaced
quite easily to renew the clearance between the shaft (or shaft sleeve) and the
stuffing box. From the standpoint of safety and fire prevention, throat bushings must
be so designed that they will not blow out of the stuffing box if the gland fails.
Lantern rings are generally used in packed stuffing boxes to distribute the sealing
and lubricating liquid around the shaft. They are also used for “leak-off” in conjunc-
tion with a throat bushing to reduce the pressure on the packing itself. Both types of
lantern rings are shown in Figure 200-46.
Take special care in packing pumps with lantern rings to make sure that the rings are
properly located opposite their connecting passages. If the lantern ring is displaced,
no sealing or lubricating medium can enter; if the lantern ring is for “leak-off”
purposes, no liquid can leave the pump.
A lantern ring used for “leak-off” in the center of the packing is usually not satisfac-
tory. Bleeding the lantern ring to allow pressure point under these conditions does
not relieve the mechanical compression of the packing, and it may reduce leakage
through the remaining packing to less than that required for proper lubrication.
Fig. 200-46 A Conventional Stuffing Box with Leak-off-type Lantern Ring at the Bottom of the
Box
256 Glands
A pump gland (identified in Figure 200-46) compresses the packing rings in the
stuffing box. Sometimes the term “gland” is used incorrectly to refer to the packing
rings or to the stuffing box in general.
Packing glands can be made in two pieces so they can be removed entirely from the
shaft to provide adequate clearance for working on the packing. In a “quench-type”
or “cowl-type” gland, the outer portion of the gland is made in the form of a hood or
cowl. The gland has a pipe connection on top for a smothering fluid when required
and an open drain on the lower side. The cowl prevents spraying packing drips or
quench liquid out the side of the pump. One-piece glands are also used to retain
mechanical seals (see Section 800).
257 Balance Drums and Bearings

In some multi-stage pumps where the impellers all face in one direction, the axial
thrust is cumulative and may reach very high values. Such pumps could require
excessively large thrust bearings. However, to balance the thrust, these designs
provide a balancing drum which has discharge pressure on one side and a lower
pressure (usually suction pressure) on the other side. The drum area (the cross-
sectional face area) is determined to approximately counter-balance the hydraulic
thrust from the impellers.
Insofar as possible, it is better to face half of the impellers in one direction and half
in the other to provide axial balance without a balancing drum. Balance drums also
reduce the pressure in the stuffing box on the discharge end.
Bearings
Two types of bearings are used on pumps: anti-friction bearings and sleeve (hydro-
dynamic) bearings. The first type includes ball, roller, and needle bearings; the
second type includes radial-sleeve bearings and thrust bearings of the disc and
tilting-pad types.
Ball Bearings
Except for large pumps, ball bearings are used in most pumps. When properly
selected and installed, they are usually satisfactory and can take both radial and
thrust loads. However, in plants with predictive maintenance, anti-friction bearings
give more warning of impending failure than do sleeve bearings. Ball bearings do
permit shorter shafts and less expensive pumps, and they have been so standardized
in uniformity and interchangeability that practically all pump manufacturers have
adopted them for most pumps.
Single and double-row ball bearings are made in various classes of internal clear-
ance. Loose internal fit (AFBMA internal fit class 3) bearings are highly recom-
mended for all single- and double-row ball bearings. This is standard for Company
specifications and in API 610. Pumps in hot services require such bearings because
the inner bearing race may be expanded by heat conducted through the shaft. The
outer race is usually cooler. Expansion of the inner race without equal expansion of
the outer race can squeeze the balls, causing early failure if internal clearance is not
adequate. Cold service pumps can operate equally well on loose clearance bearings;
for standardization and to reduce the chance of putting standard internal clearance
bearings in hot pumps by error, the use of only loose clearance ball bearings is
recommended.
Some single- and double-row ball bearings are made with filling slots cut in the
races so that a maximum number of balls can be inserted. These are called max-type
bearings. Single- and double-row bearings without filling slots in the races have a
lesser number of balls. They are called Conrad or deep-groove-type bearings. Max-
type bearings, because they have more balls, are rated to handle more load (or give
longer life for the same load) as compared to the Conrad-type.
It is doubtful, however, that this theoretically better life is actually realized in prac-
tice. Single-row max-type bearings, especially with loose internal clearance, cannot
handle any appreciable thrust load without the balls hitting the filling slot. For this
reason, single-row bearings with filling slots should not be used.
For double-row max-type bearings, manufacturers claim that the balls will not hit
the filling slots within the allowable thrust rating. Some of the Company’s refin-
eries, on the other hand, believe that filling-slot interference has been the cause of
failures. The manufacturers’ claim is probably true provided the bearings have been
made to the correct tolerances. It is likely, however, that the Conrad-type bearings
can stand higher overloads. More important, bearings usually fail because of
contaminated or inadequate lubricant or are changed out before failure because the
pump is dismantled for other reasons. As a result, there is little or no reason to
specify double-row max-type bearings just because they are rated for a slightly
longer life than the Conrad-type.
For high thrust loads, manufacturers often supply angular-contact-type bearings.

They have the advantage of the maximum number of balls without the disadvantage
of a filling slot. An angular-contact bearing is designed to handle high thrust load in
one direction only. For this reason, angular-contact bearings are usually supplied in
matched pairs to handle thrust in either direction.
Sleeve Bearings and Thrust Bearings

Sleeve bearings are of simple construction, permitting fairly easy reconditioning.
They are easy to lubricate and sudden failures are rare. Consequently, a ball thrust
bearing is sometimes used with sleeve radial bearings. For heavy thrust loads, the
tilting-pad type bearing is required. The tilting-pad type bearing consists of a
rotating disc (keyed to the shaft) with tilting pad/self-adjusting bearing shoes on one
or both sides to take the thrust loads. It is more expensive than a ball thrust bearing
and has the disadvantage of requiring very careful adjustment. It usually requires
external cooling from circulating oil. However, when properly installed and oper-
ated, the tilting-pad type bearing is quite reliable and able to handle larger varia-
tions in thrust load in both directions.
Closures for Bearing Housings

Except for pumps located in clean indoor areas, the bearing housings should be
equipped with special closures where the shafts emerge. Careful sealing of bearing
housings is required to prevent entrance of abrasive material or water, which might
cause rapid bearing failure.
Two common types of closures for bearing housings are lip seals and labyrinth
seals. Lip-type bearing housing seals are standard equipment for general service and
ANSI pumps. Labyrinth seals are standard equipment for API pumps.
Pumps equipped with oil mist lubrication of the bearings should always be provided
with Labyrinth-type bearing housing seals.
258 Base Plates

Base plates are used to supply a support and reference plane from which to accu-
rately line up the pump and driver. The space under the base plate requires filling
with grout to give it rigid support and more mass. Provide holes of adequate size
(4 inches where practical) to facilitate grouting.
Base plates should have a drain lip or a sloping top to catch the drips of both pump
and driver. The lip or top should be well sloped to a liberal drain hole (1-inch
minimum) and tapped for a drain pipe. Base plates are made of either cast iron or
fabricated steel; either is acceptable depending on the specification. Fabricated steel
base plates have tended to replace cast iron base plates because they frequently cost
less to manufacture.
259 Couplings and Coupling Guards

See the General Machinery Manual for additional information on couplings.
Couplings connecting the pump shaft and driver shaft are usually of the flexible
type. A few are rigid, but they are primarily used in vertical pumps. Flexible
couplings are required to permit operation with slight misalignment of the pump and
driver shafts and to isolate axial forces to each respective machine.
Most couplings for horizontal pumps are flexible disc-type, or gear-type. (Of these,
the flexible disc-type is generally preferred for pump drives because it does not
require lubrication.) A Company-wide study of couplings indicated that a primary
cause of coupling failures was lack of lubrication. However, this remains an area of
considerable controversy. For small pumps (less than 50 HP), rubber couplings are
used in several OPCO locations.
Couplings are highly stressed moving parts. Proper alignment is essential as
misalignment adds to the cyclical stresses. Excessive misalignment will cause
fatigue or wear-related failures dependent upon the degree of misalignment.
Flexible disc couplings should have stainless steel discs to resist corrosion. Spacers
are recommended to allow less stringent alignment tolerances and to facilitate
maintenance.
One disadvantage of flexible disc-type couplings is the danger posed from flying
debris when a failure occurs on some older designs. Newer types, such as the Meta-
stream and Thomas 71, prevent the danger of flying debris in the event of a failure.
Spring-grid-type couplings are much more limited in their allowable misalignment
than are the couplings listed above and so usually are not preferred. They do have a
certain amount of resiliency in the spring-grid, which may justify their use when
calculations indicate a need to lower the torsional critical in a reciprocating system.
Limited-end-float couplings are required with motor drivers having sleeve bear-
ings. These are large motors (ordinarily over 250 HP at 3600 rpm) in which the
sleeve bearings are designed to permit the rotor to move axially ¼ to ½ inch. The
limited-end-float coupling keeps the motor sleeve bearings within their axial limits
and lets any electrical thrust from the motor transmit to the pump thrust bearing.
The motor thrust is usually small in comparison with the design load for the pump
thrust bearing. Thomas disc-type couplings are inherently limited-end-float. Gear-
type couplings must have limit stops added.
The allowable end-play of limited-end-float couplings is given in NEMA Specifica-
tion MGI-6.11, as specified in Figure 200-47.
Fig. 200-47 Recommended End-play of Limited-end-float Couplings Per NEMA Specification

MG 1-6.11(1) Courtesy of NEMA
Minimum Motor Rotor Max. Coupling
Motor HP Syn. Speed, rpm End-Play, In. End-Float, In.
125 - 200 3600 1/4 3/32
250 - 450 1800 and less 1/4 3/32
250 - 450 3600 1/2 3/16
500 and larger All Speeds 1/2 3/16
(1) Per NEMA Specification MGI-6.11
Spacer-type couplings have a center section which can be removed without shifting
either the pump or driver. Spacer-type couplings should be specified whenever their
use will permit the removal of a mechanical seal or the disassembly of a pump
without disturbing the alignment of the pump and driver.
A spacer coupling has one other major advantage: it will accept greater actual
misalignment between pump and driver shafts.
Couplings are rated for a maximum amount of misalignment, measured in degrees.
The greater the distance between the hubs, the greater the measured misalignment
can be without exceeding the maximum limit.
This feature will make alignment easier and reduce the chance of vibration or other
alignment related pump problems. Spacer couplings are recommended for all flex-
ible coupling applications.
All couplings should be protected by substantial guards. These are best made of
steel angles and either expanded sheet metal or coarse wire mesh. Guards should
ordinarily be in accordance with the OSHA Safety Orders, which require that the
guard cover the rotating parts to within ½-inch of the pump and driver housings.
Guards should, of course, also comply with any local requirements.
260 Centrifugal Pump Subsystems
261 Special Requirements for Hot Service

Three special design features are needed for hot service: (1) an arrangement that
permits expansion to take place without moving bearings out of line or imposing
undue loads on them; (2) materials suitable for the temperature and resistant to
corrosion and erosion; (3) a design that minimizes leakage or confines it in order to
avoid ignition and hazard to personnel.
The case should be vertically split to allow confined gaskets. It should be supported
as near the center as possible in order to prevent expansion from causing casing and
shaft misalignment. Sometimes the supports are water-cooled. In any case, the
alignment should be adjusted to be correct at the operating temperature. If neces-
sary to support the casing at two points fairly far apart, provision must be made for
one support to move axially either by sliding or by elastic yielding of the support.
(The latter method of taking up case expansion is common in steam turbine
application).
The flexible coupling should be able to absorb adequate axial expansion without
developing excessive end thrust, and it should not deteriorate because of heat.
Lubricating oil to the bearings should be properly cooled, ordinarily with water to
jackets or cooling coils. Tests have been run at Richmond Refinery to determine if
we could safely omit water cooling on pumps handling hot stocks, but we have
found that some cooling is required to prevent excessive lube oil temperature. Some
pumps will require cooling because of high thrust loads regardless of stock tempera-
ture. Since water will usually be required at the mechanical seal area, it takes no
more water and only costs a small amount of piping to run the water through the
bearing jacket first. Some hot-oil pumps may develop extra end thrust from coke
formation, so it is especially necessary to provide a thrust bearing of adequate load-
carrying capacity and to cool it properly.
262 Vertical Turbine Pumps

A vertical turbine pump is a vertical shaft centrifugal pump with a multi-vane
diffuser bowl in which each impeller operates. It may be single-stage, although most
are multi-stage. Usually the pumping element and line-shaft column are suspended
from the discharge head on which the driver is mounted. These pumps are some-
times known as deepwell or turbine-well pumps.
These pumps most commonly are installed in water wells or sumps. Sometimes they
are installed in a barrel with inlet and outlet pipe connections to make a vertical
barrel pump. The latter type is usually used when NPSHA is too low for a hori-
zontal pump. The vertical barrel allows the first stage impeller to be located as low
as necessary to match the NPSHA.
With openline shaft designs, the bowl and line-shaft bearings are lubricated by the
liquid pumped. Bearing wear may be excessive because of grit in the liquid.
Selection of bowl and line-shaft bearing material is important. The most generally
suitable bearing material is high-lead bronze, although it cannot be run dry. Where
bronze may be subject to corrosion, rubber bearings may be satisfactory. Rubber
bearings are often used in water pumps and should be satisfactory provided the
water does not contain hydrocarbons or other chemicals to make the rubber swell,
and provided the rubber is never run dry.
These pumps require special attention to thrust loads. Thrust bearings are normally
in the driver. Factors affecting thrust include:
1. Weight of the rotating assembly.
2. Axial hydraulic thrust of the impeller. Thrust is usually downward; however, at
high flow rates (well beyond peak efficiency) the thrust force may reverse
direction.
3. Suction pressure acting on the area of the shaft.
The design and operation of the pump should be to keep the thrust down so the shaft
will always be in tension. This will minimize shaft whip which would cause exces-
sive bearing and seal wear. If even momentary up-thrust is possible under any
startup, operating, or shutdown condition, the thrust bearing should be designed and
locked to take the up-thrust.
Company experience has been poor with deepwell vertical turbine pumps operating
in excess of 1800 rpm. Line-shaft failures and bearing failures have been excessive.
Impellers for vertical turbine pumps may be semi-open (no shroud on the suction
side) or closed (shroud on the suction side). Open impellers have caused problems
at some deepwell installations because of the difficulty of obtaining the close clear-
ance required between the impeller and the bowl. In one case when the clearance
was set close enough to meet the rating conditions, the shaft stretch allowed the
impellers to rub. Closed impellers, on the other hand, are not usually sensitive to
end clearance. Company specifications recommend closed- type impellers in deep-
well pumps.
270 Maintaining Centrifugal Pump Flow Rates Close to the

Best Efficiency Point (BEP) or Best Efficiency Flow Rate
271 General
All centrifugal pumps operate best when flow is maintained in the preferred oper-
ating region—the region surrounding the pump’s best efficiency flow rate. The best
efficiency point (BEP) or best efficiency flow rate is defined as the flow rate where
the pump has it highest efficiency. If the pump’s flow rate is too far on either side of
the best efficiency flow rate, vibration will increase and probably exceed the limits
of API Standard 610. As a general rule, most pumps can typically operate as low as
40-50% of their BEP or best efficiency flow rate.
API Standard 610 defines the term minimum continuous stable flow as the flow
rate above which a pump can operate and not exceed the vibration limits specified
in API Standard 610. At minimum continuous stable flow, circulation of liquid
occurs in and out of the impeller’s eye. This causes the liquid’s pressure to fall
below its vapor pressure, creating gas bubbles, similar to cavitation. The bubbles
implode, causing impeller pitting that is usually more severe with heavier stocks
like water than with lighter stocks like hydrocarbon.
API Standard 610 defines the term minimum continuous thermal flow as the flow
rate at which the pump can operate without the pump’s operation being impaired by
the temperature rise or vaporization of the pumped fluid. At minimum continuous
thermal flow, the pump becomes very inefficient and most of the power transmitted
from the driver to the pump heats the liquid. The liquid can vaporize, especially if it
has a high vapor pressure like propane.
A pumping system needs to be designed so that the pump always operates above
both the minimum continuous stable flow and the minimum continuous thermal
flow. Suppliers will supply both flow rates when completing the API 610 data sheet.
Centrifugal pumps also have an upper limit flow rate, called runout or stone-
walling. For most pumps, this is approximately 120% of the BEP, but depending
upon the pump’s design, it can be closer to the BEP.
Figure 200-48 shows this operating range (minimum flow rate and maximum flow
rate). Operation at or below minimum flow rate is especially critical to high speed
pumps (such as those manufactured by Sundyne) because vibration is likely to
quickly cause gearbox damage.
Fig. 200-48 Effect of Wear on Pump Performance
Refer to Standard Drawing GA-G1097-2 in the Standard Drawing section of this

manual to determine the recommended minimum flow rates for specific pump selec-
tions. This drawing covers the minimum continuous stable flow only and is usually
conservative.
Several methods of pump control are used to prevent pump operation outside a
preferred range: pressure control, flow control, and less commonly, electronic
control based on electric power consumption. See Figure 200-49 and Figure 200-50
for schematics.
The use of variable speed devices (VSDs) to control pump speed is another method
to keep the pump operating in the desired range. The VS approach is discussed in
Section 277.
Pump controls circulate fluid from pump discharge back to a suction vessel or tank
to maintain a minimum flow rate, or they impose backpressure on a pump to prevent
runout. Runout is defined as operating beyond a pump’s maximum recommended
flow rate. Runout is most likely to be a problem when discharge lines are short (no
friction loss) or when pumping into a system with low backpressure (e.g., an empty
tank).
Fig. 200-49 Pressure Control Methods
Operation of a centrifugal pump against a closed block valve can cause over-
heating, vibration, and eventual pump failure, and should be avoided for any signifi-
cant length of time. It is normal operating procedure, however, to start centrifugal
pumps with the discharge block valve “cracked open” (i.e. nearly closed) and the
suction valve wide open. The discharge valve should then be gradually opened as
discharge pressure increases. This promotes quick build-up of pressure and prevents
cavitation, which can cause pump failures.
Fig. 200-50 Effect of Variable Suction Head on Pressure Control
When using a recirculation bypass, never return fluid directly back to pump
suction—this will cause swirling and heating problems which may raise vapor pres-
sure and affect NPSHA. Instead, route the bypass line back to a tank, vessel, or heat
exchanger.
Pressure Control
Common methods of pressure control include using a bypass controlled by one of
the following: a globe valve (or orifice), a mechanical relief valve, a proportional
pressure controller with control valve, or a pressure switch alarm. Self-contained
back-pressure regulators are used to prevent runout. These devices are discussed in
detail:
1. Fixed recirculation through a globe valve (or orifice) (See (a), Figure 200-49.)
A globe valve or orifice is mounted with a pressure indicator in a bypass line
(from pump discharge back to the storage vessel or tank). Proper recirculation
is established by keeping the discharge pressure below that corresponding to
the manufacturer’s minimum continuous flow rate. This is usually an inexpen-
sive approach but operating costs may be high due to energy losses across the
valve or orifice. Also, continuous recirculation should be specified at the time
the pump is purchased to insure sufficient capacity for both process and bypass
flow rates. This is the preferred approach for low energy pumps (less than
10 HP).
2. Variable recirculation through a mechanical safety/relief valve (See (b),

Figure 200-49.)
A relief valve is installed to bypass discharge back to a storage tank or vessel.
The set pressure is the pressure corresponding to the manufacturer’s minimum
continuous flow rate. As the discharge pressure rises (and flow decreases), the
relief valve opens, circulating flow back to suction, maintaining the required
minimum flow rate.
Relief valves reseat below their set point and this may not be acceptable. After
it pops, a relief valve may stay open or simmer, even as the pump returns to
operation above minimum flow. A relief valve is therefore not recommended
except for positive displacement pumps, and then not for minimum flow
protection, only pump casing relief protection.
3. Proportional pressure control (See (c), Figure 200-49.)
A pneumatic or electronic pressure controller is used to sense pump discharge
pressure and control a bypass control valve. As discharge pressure rises, pump
flow decreases towards the manufacturer’s recommended minimum. The
controller senses the pressure rise and opens the bypass control valve, main-
taining the flow rate above minimum.
4. Pressure alarm (See (d), Figure 200-49.)
Pressure switches installed in the discharge piping identify high discharge pres-
sures (an indication of low flow), and also low discharge pressure (loss of
prime or surging). Pressure switches are usually inexpensive and can be
connected to an alarm or automatic shutdown of the pump.
5. Back-pressure regulator (See (g), Figure 200-49.)
Runout can be prevented with a backpressure regulator (BPR). BPRs are gener-
ally self-contained and easily field adjustable. Set point changes are made
through spring replacements or by changing spring preload.
When Not to Use Pressure Control

Pressure control is usually cheaper than flow control. However, pressure control
does not work well under the following situations:
1. Pumps with Variable Suction Conditions
If suction pressure changes, the discharge pressure will also change. Thus, if
the suction tank level is high, the discharge pressure will be higher than if the
tank is nearly empty. Figure 200-50 shows this by plotting pump discharge
head vs. flow for high- and low-suction tank conditions. The control valve,
however, tries to maintain a steady pressure. Consequently, the flow rate varies
as the suction pressure varies. In some cases, the flow rate may drop to below
the recommended minimum at very low head conditions. For this reason, pres-
sure controls may be used only when the pump suction pressure is stable, or
when suction pressure variations cannot move the flow rate below the manufac-
turer’s recommended minimum.
2. Pumps With Flat Pump Curves

Pressure control works best when a large change in pressure affects flow. The
opposite occurs with flat pump curves, common for multi-stage pumps. In this
case, using pressure to control flow is inaccurate, because a small pressure
change results in a large change in flow (see Figure 200-51).
Fig. 200-51 Problems with Pressure Control on Flat Pump Curves
3. Pumps in Parallel, Each With a Dissimilar Curve

Dissimilar curves can occur from different pumps or from identical pumps,
when one is worn internally (Figure 200-48). The differences in the perfor-
mance curves can allow one pump to move more liquid than the other, with the
result that one pump may operate below the manufacturer’s recommended
minimum flow rate.
4. Pumps With Drooping Head Curves
A drooping pump curve has a dome at the top of the curve where one differen-
tial head value corresponds to two flow rates (Figure 200-52). The pump may
oscillate between the two flow rates while maintaining the same differential
head. When two or more drooping curve pumps are connected in parallel, one
pump could operate below minimum flow rate while the other operates well out
on its curve.
In the above conditions, flow control is usually the best choice, even though it is
more expensive.
Fig. 200-52 Drooping Pump Curves
272 Power Measurement

Since the majority of our pump/motor power requirements constantly increase from
shutoff to runout conditions, motor KW draw can be used to control pump opera-
tion. The motor will shut down if the pump encounters low flow (KW draw too low)
or high flow (KW draw too high).
There are several reasons why monitoring motor KW is preferred over motor
current for pump low-flow protection.
• KW = (voltage) (amperage) (power factor) (3.5). Except for small changes in
motor efficiency, KW varies directly with load. The equation applies to three
phase power and voltage measured line-to-line.
• Voltage in most electrical distribution systems fluctuates due to changes in
system loading and the reliability of purchased power. Because of this, motor
amperage draw will also vary even with a constant load.
• Power factor rapidly decreases as motor load decreases below 50 or 60 percent
of nameplate rating. As motor load and power factor decrease, amperage draw
will not vary directly with motor load. Current draw at motor loads below
50 percent may not be accurate enough to protect pumps against low-flow
related problems.
At the same time, sensors monitor phase reversal amperage power factor and
voltage fluctuations. Inductive flow controls cost approximately $1000 (1990) for a
460 volt AC, 200 HP, 3-phase device.
These devices suffer some of the same drawbacks as pressure control. Like pres-
sure control, sensing electrical power to control flow is indirect and can be inaccu-
rate.
273 Flow Control Methods

Direct flow measurement and control is typically used to maintain minimum flow
rates and to prevent pump runout. Advantages are:
1. Better accuracy than pressure control because pump flow is measured directly.
2. Flow control is unaffected by fluctuating suction pressure. It is effective on
pumps with flat or drooping curves or parallel pumps with dissimilar perfor-
mance curves.
Flow control usually costs more than pressure control, unless flow is to be measured
for other process reasons.
Methods of flow control include using a bypass with one of the following: a globe
valve (or orifice), an automatic recirculation (ARC) valve, proportional control
using a flowmeter, proportional controller/control valve, or a snap-acting solenoid
valve. Runout can be prevented by either a self-contained (pilot operated)
diaphragm valve or proportional controller/valve arrangement, mounted in the pump
discharge piping.
1. Fixed recirculation through a globe valve (or orifice). See (a), Figure 200-53.
A globe valve or orifice is mounted with or without a meter in a bypass line
from pump discharge back to the storage vessel or tank.
This is an expensive approach. Energy losses occur across the orifice or valve.
Continuous recirculation should be specified when pumps are purchased to
ensure sufficient capacity for both the process flow rate and the bypass flow
rate. If continuous recirculation is added as a retrofit, the pump may tend to
operate too far near the right end of its curve.
Due to the energy costs, this is only recommended for low-energy services (less
than 10 HP). Note that the globe valve could be completely closed, leaving the
pump unprotected.
2. Automatic Recirculation (ARC) valves. See (b), Figure 200-53.
ARC valves are often used for flow control, primarily with Sundyne high-
speed, integral-gear pumps. These valves discharge a side stream back to
suction when flow falls below the recommended minimum. Because ARC
valves recirculate only when they have to, they minimize energy loss. However,
ARC valves are only available in ANSI classes 150 and 300 and are costly.
ARC valves are most economical with high bypass flow rates and larger differ-
ential pressures. Under these conditions, recirculation with either a fixed
restriction or globe valve bypass becomes too expensive due to energy losses.
ARC valves can also be used when a pump curve will not allow the additional
flow required to satisfy the process and a fixed-recirculation bypass.
Fig. 200-53 Flow Control Methods
274 Proportional Flow Control

A flow controller receives a signal from a flow sensor/transmitter and positions a
control valve located in the bypass piping. The control valve stays closed as long as
the pump operates above minimum flow. It begins to open at minimum flow and
bypasses just enough to keep the pump running at exactly minimum flow.
This type of proportional control is more expensive than using an ARC valve
because of the instrumentation involved, but may be feasible when a meter run is
already installed for other reasons.
Proportional controls can also be used to prevent runout (see (f), Figure 200-53).
The control valve is mounted in the pump discharge line rather than the recircula-
tion line. The flowmeter/controller senses excess flow and positions the control
valve to reduce flow to an acceptable rate.
On-off control with a flow switch and solenoid valve. See (d), Figure 200-53.
A flow switch in the discharge piping detects low flow and opens a bypass-mounted
solenoid valve. Flow is routed back to suction as long as the process flow rate is low
enough to keep the flow switch tripped. If the bypass line allows too much flow to
be recirculated, a fixed restriction orifice should be installed.
The flow switch opens only when minimum flow occurs, avoiding the potential
problem of operating near the end of the curve and energy losses associated with a
continuous recirculation.
Flow switches can also be used for high and low flow alarms and shutdowns.
275 Self-Contained Flow Control Valves

A self-contained diaphragm valve can be used to prevent runout. This type of valve
works by sensing differential pressure and using this pressure to modulate the posi-
tion of the trim and diaphragm. A disadvantage is that the valve has small diameter
passages that may become plugged with sand and scale. When this happens, the
valve will operate erratically, and can worsen the pumping problems. For this
reason, do not use diaphragm valves in sandy or dirty service (such as on water
wells), or in services that could lead to corrosion or scaling of the internal trim.
Avoid hot services that could lead to failure of the elastomer diaphragms.
Self-contained flow control valves can be used to prevent runout as shown in (e),
Figure 200-53 or can be used in a bypass line back to suction to keep flow above
minimum.
276 Economics of Flow Control

The economics of minimum flow systems depend on the bypass flow rate, differen-
tial head, and the equipment already in place:
1. Fixed-orifice recirculation is inexpensive and probably the best method at low
flow rates and low differential pressures. It uses only a valve or orifice with or
without an indicator, pressure gage or rotameter. However, it constantly wastes
energy across the valve or orifice and can be expensive to operate with high
head pumps or with pumps that require a large minimum flow rate.
2. Flow control through power measurement has the same problems inherent in
pressure control. Except for applications involving very low flow rates and
pressures, this method may be the least expensive. This method is suited for
retrofitting existing equipment because no piping changes are required and
pump operating conditions do not change.
3. On-off bypass control using flow switches and a solenoid valve is simple and
relatively inexpensive. It does not waste energy and does not allow the pump to
operate at the end of its curve.
4. ARC valves and proportional controls are costly. Restrict their use to critical,
unattended, unspared, high-head pumps or pumps with large minimum required
flow rates.
277 Variable Speed Devices (VSDs)

Another approach to keep a pump operating inside the desired flow range, is to use
a variable speed device (VSD) where the VSD is controlled by a downstream flow
transmitter. The transmitter senses the flow, sends a signal to the VSD that changes
the speed accordingly to keep the pump operating at that flow. Flow does not
change so a bypass controller is not necessary; however, a low flow alarm, set above
the minimum continuous stable or thermal flow, is recommended.
Abstract
This section discusses engineering principles, pump types, application and selection
criteria, and describes two commonly used reciprocating pumps. See Section 1100
for troubleshooting information.
Contents Page
320 Pump Types 300-4
321 Single and Double Acting Pumps
331 Gas (Steam) Driven Pumps
332 Power Pumps
333 Sizing of Suction Lines
334 Selecting a Reciprocating Pump
May 2007 © 1999–2007 Chevron USA Inc. All rights reserved. 300-1
300 Reciprocating Pumps Pump Manual

A reciprocating pump is a positive displacement machine. It traps a fixed volume of
liquid at near-suction conditions, compresses it to discharge pressure, and pushes it
out the discharge nozzle. The basic principle involved is that a plunger or piston will
displace a quantity of liquid equal to its swept volume. In Figure 300-1, plunger A is
lowered into the container, displacing liquid which flows into container B. The
volume of liquid in container B is equal to the product of the cross-sectional area
of plunger A and the depth of immersion. In a reciprocating pump, the action of
plunger A is accomplished by a reciprocating piston, plunger, or diaphragm.
Fig. 300-1 Reciprocating Pump Principles From Pump Handbook, (1976) Edited by Karassik,
Krutzch, Fraser & Messina. Used with permission from McGraw Hill.
The fluid-handling section of a reciprocating pump is commonly called the liquid

end. The liquid end has a piston or plunger that displaces the fluid being pumped; a
close-fitting cylinder in which the piston travels; and suction and discharge valves
to admit and discharge the pumped fluid. Packing prevents liquid from leaking past
the rod attached to the piston, or, in a plunger pump, past the plunger.
Figure 300-2 depicts the suction stroke of a plunger pump. When the plunger moves
away from the head end of the cylinder, the discharge check valve is held closed by
the higher pressure in the discharge pipe compared to the lower pressure in the
liquid cylinder. This lower pressure in the liquid cylinder also causes the suction
valve to be opened by the higher pressure in the suction line. Fluid then flows into
the cylinder until the plunger reaches the end of its travel.
300-2 © 1999–2007 Chevron USA Inc. All rights reserved. May 2007
Pump Manual 300 Reciprocating Pumps
Fig. 300-2 Plunger Pump Liquid End During Suction Fig. 300-3 Plunger Pump Liquid End During Discharge
Stroke From Pump Handbook, (1976) Edited Stroke From Pump Handbook, (1976) Edited
by Karassik, Krutzch, Fraser & Messina. by Karassik, Krutzch, Fraser & Messina.
Used with permission from McGraw Hill. Used with permission from McGraw Hill.
Figure 300-3 depicts the discharge stroke of a plunger pump. As the plunger moves
toward the head end, the increasing pressure in the cylinder closes the suction valve.
The pressure in the cylinder continues to rise until it exceeds the pressure in the
discharge line and the discharge valve opens, releasing the volume of fluid
displaced by the plunger.
Unlike the centrifugal pump, which is a kinetic machine, the reciprocating pump
does not require velocity to achieve pressure. This is one of the reciprocating
pump’s advantages, particularly for abrasive, slurry, and high-viscosity applica-
tions. High pressures can be obtained at low velocities, and fluid capacity varies
directly with pump speed.
The discharge pressure of a reciprocating pump is only that required to force the
desired volume of liquid through the discharge system. Within the constraints of
pump construction, the maximum pressure developed for gas-driven pumps is
limited only by the differential gas pressure available; for crank-driven pumps, the
driver torque is the only limit.
The flow of liquid from a reciprocating pump pulsates, varying both in flow rate and
pressure. As the piston or plunger moves back and forth in the cylinder, alternately
opening and closing the suction and discharge valves, a cyclic pulsation is set up in
the suction and discharge lines of the pump. Figure 300-4 shows the changes in flow
rate as a function of crank angle for duplex, triplex, and quintuplex single-acting
pumps. These changes become less severe as the number of stages increases.
Fig. 300-4 Flow Rate Per Stage Permission granted. “Chemical Engineering”,
September 21, 1981.
320 Pump Types

Reciprocating pumps can be classified as power pumps or direct-acting pumps.
The driver for a power pump has a rotating shaft such as a motor, engine, or turbine.
Power pumps reciprocate the pumping element with a crank or camshaft. Direct-
acting pumps are driven by pressure from a motive gas. Direct-acting pumps were
originally known as steam pumps because steam was the motive fluid.
Reciprocating pumps can also be classified as piston, plunger, or diaphragm pumps.
Piston and plunger pumps are shown in Figure 300-5. A diaphragm pump is shown
in Figure 300-6.
Fig. 300-5 Typical Single- and Double-acting Pumps Courtesy of The Hydraulic Institute
Double-Acting Cylinder Pump
Vertical Single-Acting Plunger Power Pump
Horizontal Single-Acting Plunger Power Pump
Horizontal Double-Acting Piston Power Pump
Fig. 300-6 Diaphragm Positive Displacement Pump (Shown here as Double Diaphragm posi-
tive Displacement Pump)
Reciprocating pumps are typically classified by:

1. Type of drive
a. Direct-acting, gas-driven
b. Crank-driven (power pumps)
2. Cylinder orientation
a. Horizontal
b. Vertical
3. Liquid end arrangement
a. Plunger (outside packing)
b. Piston (inside piston rings and packing on the piston rod)
c. Diaphragm (plunger or air pushing a flexible diaphragm)
4. Number of pistons or plungers
a. Simplex
b. Duplex
c. Triplex
d. Quintuplex, etc.
5. Type of action
a. Single-acting (delivers on either forward or backward stroke, not both)
b. Double-acting (delivers on both forward and backward strokes)
Figure 300-7 illustrates these classifications. (Metering pumps that use a recipro-
cating motion are described in Section 500 of this manual.)
Fig. 300-7 Reciprocating Pump Sub-types Permission granted. “Chemical Engineering”,
September 21, 1981.
321 Single and Double Acting Pumps

Single-acting pumps discharge on either the forward or return stroke of the piston or
plunger; every cycle of the pump displaces only one volume of liquid. In double-
acting pumps, liquid is discharged on both the forward and return stroke of the
piston. Plunger pumps are only single-acting; piston pumps can be either single- or
double-acting. Figure 300-5 illustrates this pump action.
Simplex, Duplex, and Multiplex Pumps

The terms simplex, duplex, and multiplex refer to the number of piston-and-rod
assemblies in a pump. Simplex pumps have one piston-and-rod assembly; duplex
pumps have two; multiplex pumps have three or more.

This section discusses selection of reciprocating pumps. With the accompanying
Pump Applications Guidelines, it will allow an individual to select pumps for most
services.
There is normally little problem in choosing between the two basic types of pumps,
direct-acting gas driven pumps and crank-driven power pumps. Gas driven pumps,
once the workhorse of the industry, are generally limited to utility functions by the
availability of compressed gas such as steam, air, or field gas. Power pumps, which
are motor, turbine, or engine driven, are available in a wide spectrum of capacities
and heads.
331 Gas (Steam) Driven Pumps

Gas-driven pumps are commonly used for sump pump, transfer, low pressure boiler
feed, or relief drum pump-out. Although their speed and capacity are directly
affected by system pressures, gas driven pumps are of particular value when:
1. Electric power is not readily available or is unreliable.
2. A standby pump is required for use during electric power failures.
3. A wide capacity range, easily achieved by varying pump speed, is required.
4. Steam for pumping is available at little or no cost (when, for example, exhaust
steam from pumps is used to heat the pumped product).
Duplex gas pumps are more generally used than simplex because of their larger
capacity, smoother discharge, and simpler valve mechanism. Duplex pumps are also
made in a greater variety of sizes and types than are simplex pumps, which were
developed largely for vacuum or other low-pressure service.
The simplex is usually slightly more efficient than the duplex and has one less set of
packing. Simplex pumps are usually preferable in vacuum or other services where
gas or vapor must be handled. Duplex pumps may “short-stroke” and fail to clear
themselves of vapor. In fact, they may vapor lock and come to a complete stall.
Because the simplex valve mechanism prevents short-stroking, close-clearance
pumps, designed especially to handle gas or vapor, are commonly made only in the
simplex type.
Reciprocating gas pumps range in size from a small 3 × 2.75 × 3-ft. pump, rated to
handle 23 gpm water, up to a 25 × 12 × 24-ft. pump capable of handling 1150 gpm
with a maximum liquid-end working pressure of 750 psi. Simplex pumps are ordi-
narily not made for capacities over about 500 gpm.
Selection of gas driven pumps requires attention to pump speed as it relates to the
required capacity in any given service. Figure 300-8 shows the maximum recom-
mended piston speeds and corresponding revolutions per minute for direct-acting
gas pumps in various services. This figure represents the manufacturer’s recommen-
dation for maximum speed in these services. These speeds are acceptable for
standby or infrequently operated pumps, but should be reduced for pumps in contin-
uous service. For best operation, continuous duty pumps should be sized to run from
50 percent to 75 percent of the maximum speed shown on Figure 300-8.
See Figure 300-9 for additional application guidelines.
Sizing of Steam Cylinders

When possible, steam-cylinder diameters should be selected based on the required
work. Oversizing steam cylinders permits overspeeding the pumps with greatly
increased wear and high maintenance costs. Steam consumption is increased and
there is the possibility of overpressure. If a steam cylinder of the proper size is
selected, overspeeding will be minimized and it will not be necessary to place a
relief valve on the pump discharge.
Maximum Recommended Speed and Capacity of Direct-Acting Gas Pumps
Fig. 300-8
Fig. 300-9 Reciprocating Pump Application Guidelines

Pump Description Direct Acting Power Plunger Power Piston
Self Priming Y Y Y
Can Run Dry-Short Time Y Y Y
Will Emulsify N N N
Field Alignment Req’d N Y Y
Good for Some Entrained Gas Y Y Y
Good for Abrasives Y N Y
Parallel or Series Recommended P P P
Brgs Lub. (Oil, Grease, Stock) S O O
Coupling Rigid or Flexible N/A F F
Legend:
Y = yes, N = no
P = parallel
O = oil
S = stock
F = flexible
N/A = not applicable
Note: Pumps are commercially available outside the parameters shown. These pumps should be
avoided or, if they are used, special care should be taken to maintain reliability.
On the other hand, steam cylinders should not be too small. An adequate allowance
should be made for tight packing, leaking valves, and other bad service conditions.
Such an adequate allowance will result if steam inlet pressure is taken as the
minimum pressure actually available at the inlet to the pump after making proper
allowance for piping and valve losses, and the mechanical efficiencies given below
are applied.
In general, the tendency is to make the liquid piston too small and the steam piston
too big with the result that the pump has no difficulty in meeting the required pres-
sure, but has to be overspeeded to meet its capacity. If the liquid piston is gener-
ously large, there is no incentive for the operator to overspeed the pump; and, if the
steam cylinder is not too large, it may be impossible to overspeed it.
A formula for estimating the required diameter of the steam-end cylinder is as
follows:
2
2 DL Pd – Ps
D s = ---------- × -----------------
E Pi – Pe
(Eq. 300-1)
where:
Ds = Steam piston diameter, inches
DL = Liquid piston diameter, inches
Ps = Pump suction pressure, psig
Pd = Pump discharge pressure, psig
Pi = Steam inlet pressure, psig
Pe = Steam exhaust pressure, psig
E = Mechanical efficiency
Stroke – Inches Approx. Eff.
Up to 6 .60
8 to 12 .70
Above 12 .75
Reduce above efficiencies by 1/10 if viscosity exceeds 4000 SSU or differential

pressure exceeds 300 psi.
Steam Consumption
The steam consumption of a steam-driven reciprocating pump may vary consider-
ably from one pump to another even though they are all identical in design and
under similar service conditions. The steam consumption will be affected by the
mechanical condition of the pump, the accuracy of the valve timing, the tightness of
the packing, etc. Figure 300-10 illustrates how to determine the approximate steam
rate of direct-acting duplex-steam pumps in pounds per hydraulic horsepower hour.
The steam rate of simplex pumps can be obtained by taking about 93% of that
obtained for a duplex pump. Figure 300-11 illustrates the formula with applicable
notes corresponding to Figure 300-10.
A simple direct-acting steam pump cannot take any advantage of expansion of the
steam. Therefore, the steam rate is not materially reduced if steam pressures higher
than about 150 psi are used. Thermodynamically, it is better to take advantage of
expansion above this pressure in other equipment. Exhaust back pressure always
increases the steam rate materially.
Inlet steam pressure is not mentioned in the formula or the graph. The formula
assumes that enough initial steam pressure is available to do the required amount of
work. This will be true if the steam cylinder is of the proper size. Steam pumps are
almost invariably operated with a hand or automatic valve, throttling the inlet steam
to provide the required pressure, and limit or regulate the speed.
Fig. 300-10 Steam Rate of Direct-Acting Duplex Steam Pumps
Fig. 300-11 Steam Rate of Direct-Acting Duplex Steam Pumps (1 of 2)
The following formula is very simple to use and will give a good approximation of the steam consumption of a simple
duplex steam pump in fair mechanical condition. Experience shows that under the most favorable circumstances,
some pumps do better, but that many with leaky valves or otherwise in poor condition do worse.
Q
S = ---------------------- [ P + r ( P b + P′ + 18.5 ) ]
57.5e s e v
(Eq. 300-2)
where:
S = Total steam consumed, pounds per hour.

Q = Gallons per minute of liquid pumped.
P = Difference between suction and discharge pressure, in pounds per square inch.
es = Steam efficiency – See table below.
Pb = Back pressure on exhaust pounds per square inch gage.
P´ = Friction m.e.p. referred to steam cylinder in lbs. per sq. in. – See table below.
ev = Volumetric efficiency; usually over .95, and usually taken as 1.0 for rough figures; may be as low
as .5 for pumps in bad condition.
r = cylinder ratio, or ratio of area of the steam cylinder to the liquid cylinder.
Approximate Steam Efficiency and Friction M.E.P.of Duplex Steam Pumps

Stroke of pump 3" 4" 5" 6" 8" 10" 12" 15" 18" 24"
Steam eff., es .35 .375 .39 .40 .425 .45 .475 .50 .525 .55
Friction m.e.p., P´ 31.8 28.8 25.8 23.4 20.4 17.4 15.6 14.4 13.2 11.4
outside packed pumps
Ditto, inside packed pumps 30 27 24 21 18 15 13.5 12 10.5 9
Notes: Simplex steam pumps will ordinarily have a steam efficiency from 7% to 10% higher than given for duplex pumps, largely
because they are built with smaller clearance and do not short-stroke when properly adjusted. The friction m.e.p. can be
markedly reduced by the use of high-class metallic packing.
Superheat. 100° of superheat will reduce steam consumption to 87%; 200° will reduce it to 78% of that shown.
Fig. 300-11 Steam Rate of Direct-Acting Duplex Steam Pumps (2 of 2)

Discussion
The formula given is sound theoretically, and the proper values of es and P´ will give true results. It is based
on the following assumptions:
1. Specific volume of steam in cylinder is
460
----------------------
P s + 18.5
(Eq. 300-3)
cu. ft per lb., which holds well between 25# and 125# (gage).
where: Ps is the available pressure in the steam line.
2. Gage pressure of steam in cylinder at end of stroke equals
Ps = 1
--- P + P′ + P b
r
(Eq. 300-4)
3. Steam efficiency is ratio of displacement of steam cylinder to steam actually used. The low efficiency invariably
found is mainly due to the clearance volume which traps steam from the steam pipe to the exhaust pipe without
doing any work, and secondarily, to cylinder condensation. Valve leakage also plays a part. Although taken as a
constant, this efficiency is apt to vary considerably with conditions.
This formula shows the steam consumed, but does not show whether the pump can actually perform the work
or not, either as regards capacity or pressure. The maximum pressure that the pump can put up is
theoretically
P = r ( P s – P b – P′ )
(Eq. 300-5)
However, at least a 25% additional margin of safety is desirable; the working pressure should be no more than
75% of that found above. The proper working capacity of a duplex pump in gallons per minute is
approximately
2
10D L
Q = -----------------
L + 10
(Eq. 300-6)
where:
D = Diameter of liquid end, inches.
L = Length of stroke, inches.
332 Power Pumps

Crank driven power pumps are typically used in high pressure, low-to-moderate
flowrate services on pipelines and in producing field applications such as water-
flood, mud pumps, and gathering systems.
Power pumps are divided into two common types; single-acting plunger and double-
or single-acting piston. Piston pumps are limited to approximately 1500 psig, but
plunger pumps typically go to 6000 psig and have been designed for discharges as
high as 30,000 psig.
Unlike gas-driven reciprocating pump speed, power pump speed is unaffected by

system pressure. Pumps are designed to operate at a constant speed, pressure, and
capacity. If system pressure increases, the power requirement increases proportion-
ately, but the pumping rate does not change.
There are many kinds of crank-driven pumps, but three types are commonly used.
These are the vertical single-acting triplex plunger; horizontal duplex with two
double-acting pistons or pairs of plungers; and the horizontal double-acting triplex
piston. Also available are vertical single-acting pumps with five, seven, or even nine
plungers. Horizontal pumps are also made in the multi-cylinder designs. Piston
pumps are usually less expensive than plunger pumps.
A triplex plunger pump with three-inch diameter plunger and five-inch stroke, oper-
ating at 200 RPM against 1000 psig, has a typical capacity of 250 gpm.
Injection of an external lubricant with an oil lubricator into the stuffing boxes of
both piston and plunger pumps is needed to prolong packing life. Internal packing
and piston rings are not lubricated directly by oil injection.
When suction pressure is lower than atmospheric pressure, a piston pump should
usually be selected, because the plunger’s outside-packed construction generally
permits greater leakage of air and the design of the liquid passages is not suited to
clearing air accumulations.
333 Sizing of Suction Lines

When sizing suction lines for either power or gas-driven pumps, the following
guidelines are recommended:
• Average fluid velocity should not exceed 3 ft/sec and should preferably equal
1 ft/sec.
• Suction pressure should exceed the maximum fluid vapor pressure by at least
3 psi – preferably 5 – to prevent cavitation.
Adequate net positive suction head (NPSH) is the single most important design
parameter for reciprocating pumps. Suction-line flow to a simplex pump changes
from no flow to 100% within approximately the first 40% of the suction stroke. This
means that the liquid in the suction line is accelerated to full velocity in about
0.2 seconds when a pump is operating at 60 strokes per minute. The size, length,
and fluid velocity in the suction line must ensure that NPSHA is sufficient consid-
ering the pulsating flow. Some cases require a suction stabilizer.
Power pumps operate at higher speeds than gas-driven pumps, and have higher
valve velocities (due to limited valve area), and higher valve-spring loads (for fast
action). Consequently, they require a higher NPSH and, with an increase in RPM,
increasing suction line acceleration head. Figure 300-12 gives allowances for accel-
eration head for double-acting duplex power pumps handling water and similar low
viscosity fluids, based on empirical data developed by reciprocating pump suppliers.
Fig. 300-12 Suction Line Liquid Acceleration Head (ft) for Double-Acting Duplex Power Pumps
Suction Line Acceleration Head for Suction Line Length (ft)
Average Suction Line
Velocity (fps) 25 50 75 100
0.5 1.7 3.3 5.0 6.5
1.0 3.3 6.0 59.8 13.0
2.0 6.5 13.0 19.5 26.0
Notes: Refer to Section 130 for a detailed discussion of acceleration head.
(1) For triplex pumps, use 57% of the values shown.
(2) For single-acting duplex and simplex pumps, use 174% of the values shown.
(3) Multiply values given above by the actual RPM divided by 60.
(4) Length of line is actual feet, not equivalent length. For pumps with suction stabilizers, length of line equals
10 pipe diameters.
(5) Acceleration head is added to the NPSH required by the pump.
(6) The NPSH requirement for a reciprocating pump, covering pressure loss from the inlet flange to the cylinder, is prima-
rily determined by the liquid velocity through the suction valve, the weight of the valve, spring loading on the valve,
and the liquid viscosity.
(7) 12-ft NPSH allowance for a reciprocating gas pump is desirable.
(8) 8-ft to 10-ft NPSH is sufficient with some slower speed pumps.
(9) Special close clearance simplex pumps are available when some vaporization on the suction side may be expected.
(10)For hydrocarbons, use 75% of the values shown.
334 Selecting a Reciprocating Pump

The following steps may be used to select a reciprocating pump. For additional
guidance, see Section 2100.
1. Determine process duty.
2. Calculate liquid properties, if necessary.
3. Determine pipe pressure losses.
4. Calculate the suction head (same as for centrifugal pump).
5. Calculate the discharge head (same as for centrifugal pump).
6. Calculate the total head (same as for centrifugal pump).
7. Convert total head to pressure rise.
8. Calculate the NPSHA.
The expression for calculating the NPSHA for a reciprocating pump is similar
to that for a centrifugal pump except that acceleration head is included. Accel-
eration head is the force required to accelerate the fluid in the suction line. The
NPSHA for a reciprocating pump may be obtained as follows:
NPSHA = h p + h s – h f – h vpa – h acc

(Eq. 300-7)
where:
hp = absolute pressure at suction source, e.g., a vessel (ft)
hs = static suction head (ft)
hf = friction head loss in suction piping including entrance losses (ft)
hvpa = vapor pressure of the fluid at pumping temperature (ft)
hacc = Acceleration head (ft)
(See Section 130 for calculating acceleration head.)
9. Calculate brake horsepower.
BHP = ( GPM × psi ) ⁄ ( 1715 × eff )

(Eq. 300-8)
where:
GPM = Flow Rate in U.S. gallons per minute
psi = total differential pressure (pounds per square inch)
eff = pump efficiency (non-dimensional)
Note Use the following equation to convert head in feet to psi:
psi = h t × SG ⁄ 2.31
(Eq. 300-9)
where:
ht = Total head (ft)
SG = Specific gravity of liquid
10. Select particular pump.
Using the pump manufacturers’ literature and catalogs, select the pump for the
conditions obtained in the calculation. If possible, avoid selecting the largest
piston or plunger size for the pump case. Also avoid pumps which would have
to operate continuously at maximum allowable speed.
11. Consult pump Vendor.
Discuss pump selection with the Vendor for further recommendations and as a
check of the selection procedure.
12. Prepare pump data sheet and specification. See specification for reciprocating
pumps and API Specification 674.
Pump Description Positive displacement-reciprocating-piston-duplex-direct acting-

gas driven (steam, air or process gas)
Typical Service Relief drum pumpout. Low pressure boiler feed. Water. Sludge.
Sump pump. Transfer.
Typical Pressure/Capacity Range 0–700 psig/0–500gpm
Max Allowable Temperature 350°F
Typical Speed Range 30 to 60 RPM (with piston speeds usually between 50 and
100 FPM)
Construction Features Normal duplex, double acting, simplex available. Normally C. I.
steam and liquid ends with steel or bronze rods & trim
Typical Control Method Speed control by throttling drive gas (steam, air, process gas),
usually manual
Advantages Self priming. Will operate at very low speeds. High efficiency.
Minimizes liquid emulsification. Handles viscous stocks. No elec-
trical power is required. Suitable for unattended remote installations
Disadvantages Pump speed is affected by system pressure. Subject to vapor lock
with low NPSH available. Will stall with too-high system back
pressure. Pulsating flow can affect sensitive instrumentation
downstream
Specification API 674. See also PMP-PC-1081 in this manual.
Data Sheet API 674, Appendix A. See also PMP-PC-1081 in this manual.
Fig. 300-13 Duplex, Direct-acting, Gas-driven, Piston Reciprocating Pump
Part Part Part

No. Name of Part No. Name of part No. Name of Part
1 Steam Cylinder 10 Valve Rod 19 Valve Cover

2 Steam Cylinder Head 11 Valve Rod Lever 20 Valve Stem
3 Steam Piston 12 Liquid End Piston Rod 21 Valve Spring
4 Steam Piston Rings 13 Gland 22 Valve
5 Steam End Piston Rod 14 Gland Bushing 23 Valve Seat
6 Condensate Drain 15 Stuffing Box
7 Steam Chest 16 Liquid Piston
8 Slide Valve 17 Liquid Piston Rings
9 Valve Rod Adjustment Nut 18 Cylinder Liner
Pump Description Positive displacement-reciprocating-plunger (power pump)

Typical Service High pressure/low flow. Gathering systems/pipeline. Waterflood.
Drawing rig. Mud pumps. Well workover.
Typical Pressure/Capacity Range 500–6000 psi/10–600 gpm
Typical Speed Range 0–450 RPM
Construction Features Vertical configurations available up to 200 HP. Available in duplex
through nonuplex, although triplex is most common. Crank driven
with motor, turbined with gearbox, or engine drivers. Steel liquid
end. Cast iron and steel power end. Self-contained lubrication
system.
Typical Control Method Variable speed or flow bypass
Advantages Higher pressures available than with piston pumps (up to 30,000 psi).
Self-priming. Constant delivery at high efficiency over wide pres-
sure range. Minimum fluid emulsification. Handles viscous stocks.
Can run dry for a limited time.
Disadvantages and Limitations Pulsing flow. Low capacity. High first cost and maintenance cost.
Low tolerance for abrasives. Subject to vapor lock at low suction
pressure with high vapor pressure stock.
Specification API 674. See also PMP-PC-1081 in this manual.
Data Sheet API 674, Appendix A. See also PMP-PC-1081 in this manual.
Fig. 300-14 Reciprocating Plunger Power Pump Copyright 1995 Ingersoll Dresser Pumps. Worthington is a
trademark of Ingersoll Dresser Pump Company
400 Rotary Pumps
Abstract
This section discusses engineering principles, pump types, mechanical features, and
system considerations for rotary pumps. A checklist for troubleshooting rotary
pump problems is included in Section 1100.
Contents Page
411 Slip
412 Volumetric Efficiency
413 Mechanical Efficiency
414 Suction Conditions
420 Pump Types and Mechanical Features 400-5
421 Screw Pumps
422 Gear Pumps
423 Eccentric-Rotor-Vane Pumps
440 Rotary Pump Descriptions 400-11
450 System Considerations 400-20
451 Relief Valves
452 Drivers
453 Instrumentation and Control
454 Strainers
455 Lubrication Before Start-Up
400 Rotary Pumps Pump Manual

Rotary pumps are positive displacement pumps, but unlike reciprocating pumps,
have relatively steady, non-pulsating flow. Rotation of the rotor(s) within the casing
traps pockets of liquid at suction conditions, elevates the fluid pressure, and then
pushes the fluid out the discharge.
Like other positive displacement pumps, the capacity of a rotary pump is directly
proportional to pump speed. Regardless of capacity or pump speed, the discharge
pressure of a rotary pump is that required to force the fluid through the discharge
system. The discharge pressure is limited only by the mechanical design of the
pump casing, or by the fluid viscosity and torque capability of the driver.
411 Slip
The actual capacity of a rotary pump is less than the calculated theoretical capacity.
The theoretical capacity is reduced by recirculation back through the clearances
between the rotor(s) and casing. The recirculated fluid is termed “slip”. Thus, the
delivered capacity (Q) is calculated:
Q = Qt - S
(Eq. 400-1)
where:
Qt = theoretical capacity
S = slip
Rotary pump capacity is reduced (i.e., slip is increased) by:
1. Decreasing viscosity
2. Increasing differential pressure
3. Increasing internal clearances between rotating and stationary parts
4. Decreasing pump speed
412 Volumetric Efficiency

The amount of slip is usually characterized by manufacturers as volumetric effi-
ciency. The volumetric efficiency of a rotary pump (Ev) is:
Ev = Q/Qt = (Qt - S)/Qt

(Eq. 400-2)
The volumetric efficiency of the pump, which is a function of pump speed,
viscosity, and differential pressure, impacts pump sizing and selection. The volu-
metric efficiency is one component of the pump’s overall efficiency. Other factors
include heating losses from fluid friction, mechanical losses in the bearings, and
drag due to viscosity.
Pump Manual 400 Rotary Pumps
413 Mechanical Efficiency

The overall mechanical efficiency (Em) is:
Em = (HHP/HP) × 100
(Eq. 400-3)
where:
HP = Input Horsepower
HHP = Hydraulic Horsepower = (Q)(DP)/1715
Q = Flow in gpm
DP = Differential Pressure in psi
The overall mechanical efficiency of rotary pumps generally ranges from 60% to
70%. It may be as high as 80% for a pump with low slip and very favorable condi-
tions, or as low as 50% under unfavorable conditions.
The rotary pump curve in Figure 400-1 shows the efficiencies and capacities
discussed above, with the resultant horsepowers. Pump speed and fluid viscosity are
assumed constant. Similar curves could be drawn as a function of speed or viscosity.
The “output horsepower” shown in the curve is equivalent to hydraulic horsepower
described above.
Fig. 400-1 Rotary Pump Curve From Pumps, Types Selection, Installation, Operation and
Maintenance by Kristal and Annett; 2nd Ed. (1953). Used with permission from
McGraw Hill.
414 Suction Conditions

Suction conditions have a great effect on performance. It is important to completely
fill the moving cavities of the pump to ensure quiet and efficient operation. The
more viscous the fluid, the greater the resistance to flow. Therefore, the rate of
filling the moving cavities is slower for higher viscosity fluids. Also, if the cavities
are moving too fast they will not fill completely. Pump speed must then be limited
to satisfy the effect of high viscosity and the NPSHA.
NPSHA for rotary pumps is calculated in the same way as centrifugal pumps (refer
to Section 220). NPSHR is sometimes reported by rotary pump manufacturers as
MSLA, maximum suction lift available. To convert to NPSHR:
NPSHR = (Inlet Vessel Pressure) - MSLA
The MSLA reported by a manufacturer should not be exceeded. This is the same as
saying NPSHA must be greater than NPSHR. A margin of at least 3 psi between
NPSHA and NPSHR is recommended.
Rotary pumps can handle fluids with dissolved or entrained gases. The gas expands
in proportion to the absolute pressure at the pump inlet and will displace an equal
volume of liquid. This must be taken into account when specifying liquid capacity.
If the fluid pumped is compressible, such as a mixture of oil and air, the volume
within each closure is reduced as it comes in contact with the discharge pressure.
This produces pressure pulsations, the intensity and frequency of which depend on
the discharge pressure, the number of closures formed per revolution, and the speed
of rotation. Under some conditions, the pressure pulsations are of high magnitude
and can cause damage to piping and fittings or even the pump, and will almost
certainly be accompanied by undesirable noise.
The effects of dissolved gas may be reduced by lowering the suction lift through
relocating the pump, increasing the suction line size, or changing the piping
arrangement.
Remember that there is a difference between entrained or dissolved gas and the
liquid vapor. Adequate NPSHA of the liquid must still be maintained for acceptable
performance, even though the pump may be specified to handle entrained or
dissolved gas.
An important characteristic of rotary pumps is their ability to handle wide viscosity
variations and very high viscosity fluids. They can handle a viscosity range from
35 to 1,000,000 SSU (2 to 225,000 centistokes), but are normally applied in the
100 to 250,000 SSU (20 to 50,000 centistokes) range.
Figure 400-2 illustrates the solubility of air in oil.
Fig. 400-2 Solubility of Air in Oil From Pumps, Types Selection, Installation, Operation and
Maintenance by Kristal and Annett; 2nd Ed. (1953). Used with permission from
McGraw Hill.
420 Pump Types and Mechanical Features

Rotary pumps are of three general types: screw pumps, gear pumps, and eccentric-
rotor pumps. There are several variations of each.
421 Screw Pumps

Screw pumps are the most common type of rotary pump found in the petroleum
industry. Figure 400-3 shows three sub-types of screw pumps: three-screw, two-
screw, and single-screw.
Fig. 400-3 Screw Pump Sub-Types
The three-screw pump, built by Delaval (IMO) and Alweiller, has three screws that
rotate within the stator or casing. The casing supports the rotors along their entire
length, and functions as a bearing journal. The rotors, stator, and mating teeth
between rotors are all lubricated by the process fluid. Because of this, the three-
screw is referred to as the “internal bearing” version of the screw-pump type. The
center screw, or power rotor, is coupled to the driver and drives the other two
screws, called idlers. A major advantage of the three-screw pump is that it requires
only one mechanical seal.
The two or twin-screw pump is similar to the three-screw, but the two screws are
designed with greater internal clearance. They are designed so that contact does not
occur between the rotors themselves, or between either rotor and the stator. Unlike
the three-screw pump, where the design provides contact between the three screws,
the rotors in the two or twin-screw pump are designed to avoid contact. The clear-
ances between the rotors are held by anti-friction bearings that maintain precise
alignment of the rotors. The bearings can be mounted internally or externally. In the
external-bearing version, the timing gears and bearings are mounted in external oil-
lubricated housings. The shafts protrude through the casing in four locations,
making four mechanical seals (or packing) necessary.
An advantage to the external-bearing two-screw pump over the three-screw pump
and the internal bearing two-screw pump, is that it is less susceptible to wear in
services with suspended solids; the trade-off being higher cost. The external-bearing
version also performs well in services with lower viscosities and lubricating capa-
bility. Neither the internal-bearing two-screw pump nor the internal-bearing three-
screw pump should be applied in services with suspended solids.
Specially designed two-screw pumps are used in multiphase service. Here the pump
transports a combination of liquid, gas, and solids. The most common application is
in oil fields where the crude has that composition. Helico-axial pumps, or mixed
flow, multistage centrifugal pumps are also used in this service as is the single screw
pump, discussed below. Multiphase pumps (MPPs) are discussed in more detail in
Section 770, Multiphase Pumps.
The single-screw—built by Moyno, Alweiller, and others—consists of a single
screw rotating in an elastomeric (rubber), teflon or ceramic stator. The pump is also
referred to as a progressive cavity pump (PCP). The single-screw is a slow-speed
pump and is physically large for the amount of fluid pumped. It can tolerate small
amounts of solids and dissolved gas in the pumped fluid and is often chosen for that
reason.
422 Gear Pumps

Gear pumps are commonly found in heavy-fuel, lube-oil, and crude-oil services.
They generally cost less than screw pumps, but have a shorter life. There are three
general sub-types, with variations: (1) external gear, (2) internal gear, and (3) eccen-
tric rotor pumps. Figure 400-4 shows external gear and internal gear pumps.
Fig. 400-4 Gear Pump Sub-Types
The external-gear pump consists of two counter-rotating gears, one driving and the
other an idler. The gear-forms used include spur, herringbone, helical (or spiral), and
Heliquad.
The rotors of external-gear pumps are supported by anti-friction bearings. The
bearing can be internally or externally mounted. Internal-bearing versions can only
be applied to clean lubricating fluids. External-bearing versions can be applied in
slightly less favorable services, but require four mechanical seals. The more
complex helical and heliquad gear-forms are provided with timing gears to prevent
tooth contact. Some herringbone versions also have timing gears.
The internal-gear pump is a reliable, inexpensive pump.
The outer gear of this pump rotates concentrically in the casing and is coupled to the
driver. The inner gear is eccentrically mounted and is an idler. The crescent is
stationary and seals the tooth cavities between suction and discharge. Internal gear
pumps are usually used only in clean services, but material options are available
which make them more suitable for services with some suspended solids than
internal-bearing, external-gear, and two-screw pumps, or three-screw pumps.
The lobe pump is used infrequently in petroleum industries. Lobe pumps come in
two- and three-lobe versions, the latter of which is shown in Figure 400-4. Lobe
pumps operate like gear pumps except the fluid is delivered to the discharge in a
smaller number of larger quantities.
423 Eccentric-Rotor-Vane Pumps

Eccentric-rotor pumps are available in several intriguing designs, few of which
have any practical value in the petroleum industries.
The sliding-vane pump is the more common version of the eccentric-rotor type. It
consists of an eccentrically rotating drum with radial slots which house sliding
vanes. The vanes form and seal the fluid cavities that shrink and elevate the fluid
pressure as the drum rotates. Sliding-vane pumps are inexpensive and unreliable due
to frictional wear and vane breakage.
The various other types of eccentric-rotor pumps are classified as rotary pumps, but
are not true rotaries because they convert rotational movement to reciprocating
movement, much like the cam-and-piston pump shown in Figure 400-5. Pumps of
this type are usually applied to hydraulic fluid (power fluid) services.
Fig. 400-5 Eccentric Rotor Pumps and Vane Pumps

The guidelines below, in addition to the applications chart shown in Figure 400-6,
can be used to help select the appropriate type of rotary pump. It is assumed that a
rotary pump has been selected rather than centrifugal or reciprocating pumps.
1. Screw Pumps Often the only type available for higher pressures or larger capacities. Usually
- General more expensive but will generally have a longer pump life than gear pumps.
- Three-Screw The most common screw-pump type and usually the first choice in screw
pumps. Considered the standard pump for lube-oil and seal-oil circulation for
major machinery trains, barge unloading and pipeline services. Also the stan-
dard pump for fuel-oil delivery to burners. Occasionally used in resid process
services. Discharge pressures to 3000 psig and capacities to 1000 gpm. Can be
directly coupled to the driver.
- Two-Screw Usually more expensive than the three-screw. Available in capacities
exceeding the range of three-screw pumps. Often applied in fuel oil, crude oil,
and asphalt transfer services, especially at marine and rail terminals. Also
occasionally used in resid process services. Discharge pressure to 1500 psig
and capacity to 5000 gpm. Can handle lower viscosities than three-screw
types.
- Single-Screw Seldom the appropriate choice for any application because of low reliability.
Occasionally applied to fluids with entrained solids, water treating, and oil
sludge handling services. In such a service the pump would be run at low
speed and would have a rubber stator to minimize erosion. May find one
applied to a non-newtonian or other shear sensitive fluid. Discharge pressure to
200 psig and capacity to 450 gpm.
2. Gear Pumps Generally less expensive than screw pumps, and used when an inexpensive
- General short-life pump can be tolerated. Also used in intermittent services.
- External-gear The “multi-purpose” rotary pump. Used in clean lubricating services. Some-
times used as an expendable pump in non-lubricating services because of low
cost. Discharge pressures to 700 psig and capacity to 600 gpm.
- Internal-gear Usually less expensive than external-gear pumps, and usually the first choice
in a gear pump. Occasionally applied in non-lubricating or slightly solid-
bearing fluid services. Discharge pressure to 200 psig and capacity to 150
gpm. A special design exists for pressures up to 1000 psig, for filtered liquids
and power fluid services. Smaller sizes can be directly coupled to driver, but
usually require a gear or belt drive.
- Lobe Not very common. May be found in services that might emulsify from liquid
shearing. In such a service the pump would run at a very slow speed to prevent
agitation.
3. Eccentric- Inexpensive, but wears excessively on all but the cleanest lubricating stocks.
Rotor-Vane Seldom used. May be applied in services where the pump may need to handle
entrained or dissolved gas. Also may be applied in services with moderate-
vacuum suction conditions. Discharge pressures to 200 psig.
- Cam and Piston Used in power fluids (hydraulics) services. Some special designs are applied to
higher vacuum services.
Fig. 400-6 Pump Application Chart for Rotary Pumps

Two-Screw Ext. Gear
w/Ext. Single- Ext. Gear w/Ext. Internal
Pump Description Three-Screw Bearing Screw w/Int. Brngs. Brngs. Gear
Diff. Pressure Range (PSI) 3000 1500 200 70 700 200
Flow Range (GPM) 1000 5000 450 600 600 150
Max. Temp. (F) 500 500 180 250 250 250
Horsepower Range (HP) 800 2000 800 400 400 40
Speed Range (RPM) 3600 1800 800 1800 1800 1800
Viscosity Range (SSU) 1,000,000 1,000,000 1,000,000 250,000 250,000 250,000
Variable Viscosity N Y M N M N
Relative Cost to (#) 3 5 4 2 3 1
Purchase
Relative Cost to Install (#) 3 4 2 2 3 1
Relative Cost to Maintain (#) 4 3 5 2 3 1
Self Priming Y Y M Y Y Y
Can Run Dry – Short Time N M N N M N
Will Emulsify M N N N N M
Field Alignment Required Y Y Y Y Y Y
Good for Some Entrained Gas Y M Y Y Y M
Good for Abrasives N N Y N N M
Parallel or Series Recommended P P P P P P
Bearings Lub. (Oil, Grease, Stock) S O S, G S O S
Coupling Rigid or Flexible F F F F F F
Legend: Y = Yes, N = No, M = Marginal, # = Relative Number (1-5, 1 is best), N/A = Not Applicable
(1) Pumps are commercially available outside the parameters shown. Pumps with values outside the parameters should be
avoided or special care should be taken to maintain their reliability.
(2) Rotary pumps are, by nature, self priming. However, they require liquid in the pump in order to lubricate internal parts.
The following can be further used to help select the appropriate rotary pump for the
service types listed:
Clean Services. All rotary pumps will perform well with clean stock that lubricates
(i.e., viscosity greater than 35 SSU). In general, screw pumps can be directly
coupled to the driver and are more efficient at high pressures, high flow rates, and
high viscosities than gear pumps. They are also more expensive. Gear pumps are
competitive in the lower flow, lower pressure applications and usually require gear
or belt drive arrangements.
Dirty Service. The single-screw should be the only pump considered for handling
stock with significant solids such as abrasive slurries, sewage, or sludge. An alterna-
tive for intermittent low pressure situations is use of a small, inexpensive internal-
gear pump which can be replaced as needed. The use of external timing gears in
two-screw pumps helps to maintain clearances between screws and case, so they can
tolerate fine, low-hardness contaminants. Another alternative in low flow services
might be an air driven diaphragm pump.
Lubrication and Hydraulic Systems. Normally, three-screw pumps are used in
this service. With clean, lubricating stock their continuous operation is reliable and
extended. Three-screw pumps perform best at high pressures with a clean, fixed
viscosity stock, where external bearings are not required.
High Pressures. Screw pumps are better suited than gear pumps in services
exceeding 400–500 psig. They have a higher mechanical efficiency at higher pres-
sures because of greater volumetric efficiency. However, at low pressures their
mechanical efficiency is lower than gear pumps because of the greater friction
losses due to larger internal surface areas. Higher pressures are possible with three-
screw pumps than two-screw.
High Viscosity . Two-screw pumps are more versatile than gear or three-screw in
handling stocks with variable viscosity. The most common two-screw pumps use
external timing gears and bearings, which make them more expensive than three-
screw pumps.
Self Priming. All rotary pumps can operate with considerable dissolved or
entrained gas in the stock pumped without losing suction. Continuous handling of
appreciable quantities of vapor along with liquid, especially under cavitating condi-
tions, can cause excessive noise and vibration and contribute to rapid wear.
Typically, screw pumps provide greater suction lift than gear pumps, and three-
screw pumps are better than two-screw.
Vacuum Service. Rotary pumps lubricated by special oils are often used in vacuum
services to pump air or other gases or vapors. Low-vapor-pressure oils are used to
lubricate the pumps and seal the clearance spaces. Liquid-ring vacuum pumps are
also available.
Non-Pulsating Flow. Typically, screw pumps provide smoother fluid discharge than
gear pumps. Three-screw pumps are better than two-screw.
Drive Arrangement. Typically, gear pumps and single screw pumps operate at
slow speeds which require belt or gear drives. Three-screw and two-screw pumps
almost always operate at motor speed, eliminating the need for a belt or gear drive.
Emulsifying Affect. All types are less likely than centrifugal to promote stock
emulsification. The lower the speed, the better.
Running Dry. These pumps are unforgiving if run dry; they fail quickly. Twin
screw and twin gear (with external gears/bearings) can tolerate running dry slightly
better than others.
440 Rotary Pump Descriptions

This section describes five rotary pumps.
Pump Description Positive Displacement – Three Screw – Rotary

Typical Service Heavy fuel oil; crude oil; asphalt. Lubricating stocks up to
1,000,000 SSU (225,000 cs). Lube oil, oil transfer services.
Typical Head/Capacity Range 50–3000 psig/0–1000 GPM
Speed Range 200–3600 RPM
Construction Features Driven rotor drives 2 idler rotors. Generally double-ended to
balance thrust in larger pumps. Steel case in hydrocarbon
service/C.I. or steel rotors. One bearing needed to center rotors.
One stuffing box or head seal. Close internal clearances
(0.005 inch – 0.015 inch).
Typical Control Method Variable speed or flow bypass.
Advantages Quieter, more efficient, smoother flow than gear pumps at high
pressures. Self priming. Non-pulsing flow. Will not emulsify. Will
pass vapors. Fewer seals and bearings than double screw pumps.
Constant capacity over wide range of pressures with high effi-
ciency. Can be mounted vertically and horizontally.
Disadvantages and Limitations More expensive than most gear pumps. Poor with non-lubricating
stock, abrasives or corrosive stock. Cannot run dry. Lower toler-
ance for abrasives than reciprocating pumps.
Specification API 676 (Special Purpose only). See also PMP-PC-458 in this
manual.
Data Sheet API 676, Appendix A (Special Purpose only).
Fig. 400-7 Three-Screw Rotary Pump Courtesy of IMO Industries: Pump Division
Pump Description Positive Displacement – Two-screw – External Bearings – Rotary

Typical Service Heavy fuel oil; crude oil; pipeline; unloading; asphalt. Lubricating
stocks to 1,000,000 SSU (225,000 cs).
Construction Features Driven rotor meshes with idler rotor and both are driven by timing
gears. Generally double-end to balance thrust in larger pumps. Four
stuffing boxes or mechanical seals. Steel case in hydrocarbon
service/C.I. or steel rotors. Externally lubricated timing gears and
bearings. Close internal clearances (0.005 inch - 0.015 inch).
Advantages Quieter, more efficient, smoother flow than gear pumps. Higher pres-
sures. Self priming. Non-pulsing flow. Will not emulsify. Will pass
vapors. Constant capacity over wide range of pressures with high
efficiency.
Disadvantages and Limitations More expensive than single screw, three-screw, or gear pumps. Poor
with non-lubricating or corrosive stock or abrasives. Cannot run dry.
Lower tolerance for abrasives than reciprocating pumps. Lower
mechanical efficiency than gear pumps at low pressures.
manual.
Data Sheet API 676, Appendix A (Special Purpose only)
Fig. 400-8 Two-Screw External Bearings Rotary Pump Courtesy of IMO Industries: Pump Division
Pump Description Positive Displacement – Single-screw (Progressive Cavity) –

Rotary
Typical Service Viscous crude or stock with suspended solids or abrasives; sludge;
services with wide variations in viscosity.
Construction Features Steel rotor in C.I. Soft teflon, Buna N, or cast iron stator; steel in
hydrocarbon service.
Advantages Self priming. Will pass solids and vapor. Usually less expensive
than two-screw pumps. Can remove rotor with pump in the line.
Uniform, non-pulsing discharge. Will not emulsify stock. Less
expensive than reciprocating pumps. Can handle non-newtonian or
shear sensitive fluids. Can be mounted horizontally or vertically.
Disadvantages and Limitations Less reliable than other positive displacement pumps. Cannot run
dry. Uses large amount of floor space for amount of fluid pumped.
Expensive to maintain and unreliable. Stator materials must be
inert to the fluid in order to prevent seizure.
manual.
Fig. 400-9 Single-Screw Rotary Pump (Progressive Cavity) Courtesy of Moyno Industrial Pumps, a Division of
Robbins & Myers
Pump Description Positive Displacement – External Gear – Rotary

Typical Service Low pressure, low capacity lubricating stocks. Pipelines - wet crude or
stocks. Lubrication and hydraulic oil systems. Air, gas, or vapor
vacuum service.
Speed Range 300/1800 RPM
Construction Features Can be external bearing or internal. Carbon steel case for
hydrocarbon service/steel, bronze, C.I. gears. Can be packed or with
mechanical seals. Close internal clearances (0.005 inch – 0.015 inch).
Advantages Self priming. Constant delivery at high efficiency over wide range of
pressures. Minimizes fluid emulsification.
Disadvantages and Limitations Cannot handle external forces from piping. Cannot run dry. Pumped
fluid must provide lubrication. Low tolerance for abrasives. Can have
high maintenance cost. Usually has four stuffing boxes.
Specification API 676 (Special Purpose only). See also PMP-PC-458 in this manual.
Fig. 400-10 Positive Displacement External Gear Rotary Pump Courtesy of Kinney Vacuum Company, Inc.
Pump Description Positive Displacement – Internal Gear – Rotary

Typical Service High pressure, low capacity lubricating stocks. Pipelines - wet crude or
stocks. Lubrication and hydraulic oil systems. Air, gas, or vapor
vacuum service.
Speed Range 300/1800 RPM
Construction Features Can be external bearing or internal. Carbon steel case for
hydrocarbon service/steel, bronze, CI gears. Can be packed or with
mechanical seals. Close internal clearances (0.005 inch – 0.015 inch).
Advantages Self priming. Constant delivery at high efficiency over wide range of
pressures. Minimizes fluid emulsification. Handles viscous stocks and
two phase flow. Material options available for suitability in service
with some suspended solids.
Disadvantages and Limitations Have no timing gears. More complex machining required than with
external gear pumps.
Specification API 676 (Special Purpose only). See also PMP-PC-458 in this manual.
Fig. 400-11 Positive Displacement Internal Gear Rotary Pump Courtesy of Viking Pump, Inc.
450 System Considerations
451 Relief Valves

Relief valves are required in the discharge of rotary pumps to prevent over-pres-
sure. Blocking in a rotary pump will cause failure of the weakest element in the
drive train, usually a cracked pump casing. Many rotary pumps can be obtained with
internal “built-in” relief valves. API 676 prohibits use of internal relief valves,
because they are usually unable to protect the pump from overheating in a relief
situation. Separate relief valves are preferred, piped back to the vessel in the inlet
system.
452 Drivers
Rotary pumps are almost always driven by electric motors, geared motors, or belt
driven with motors. Occasionally a rotary pump is driven by a steam turbine, and in
rare cases by an internal-combustion engine.
453 Instrumentation and Control

Capacity control is obtained either by varying the speed of the pump or by recircu-
lating a portion of the discharge. Recirculation is a common method of control, as
rotary pumps are most often operated at constant speed with electric motor drivers.
454 Strainers
Permanent suction strainers should usually be installed to protect rotary pumps from
damage by foreign material. This requirement applies to most rotary pump installa-
tions, except for inexpensive pumps in relatively clean systems where the initial cost
does not justify a strainer to protect the pump. A dirty or plugged strainer may also
cause pump damage by increasing the friction loss in the suction piping and
reducing the NPSHA. A low pressure alarm downstream of the strainer may alle-
viate this problem.
Even though strainers are provided, the suction lines to rotary pumps should be
carefully flushed out before final connection of piping to the pumps.
455 Lubrication Before Start-Up

Because of possible or necessary internal contact, rotary pumps must be lubricated
before being started. When they are installed to handle appreciable quantities of
vapor even for a short time, as in priming, special precautions may be required to
insure proper lubrication before start-up. In self-priming installations for low-
viscosity stocks, rotary pumps should ordinarily be installed at the bottom of a “U”
in the piping which can be filled with liquid to provide lubrication during startup or
if the pump loses suction. In intermittent services, foot valves are also recom-
mended for low-viscosity stocks.
500 Metering Pumps
Abstract
This section covers metering pump principles and application criteria. See
Section 1100 for troubleshooting information.
Contents Page
511 General
512 Hydraulic End
513 Proportioning Control
520 Pump Applications 500-6
500 Metering Pumps Pump Manual
511 General
Metering pumps deliver accurate quantities of liquid into a process or system. They
usually handle a small discharge volume (typically between 0.5 gph and 10 gph)
and high discharge pressure (up to 30,000 psig). The volume delivered is infinitely
controllable between limits and virtually independent of discharge pressure.
Most metering pumps are positive displacement type. They usually are driven by an
electric motor through an integral reduction gear. Air-driven units are common in
producing locations. Proportioning mechanisms are usually integrated with the drive
mechanism.
Since metering and proportioning duties commonly call for different volumes to be
delivered at different times, most metering pumps are the variable-capacity type,
usually variable stroke. Advantages and disadvantages of different types of metering
pumps are shown in Figure 500-7. Metering duties are normally accomplished with
special designs of piston-diaphragm pump heads where high accuracy is required.
Plunger type diaphragms may be used where higher pressures are required. Other
positive displacement pumps are not excluded from metering service. Rotary
pumps, for example, may be used for metering viscous fluids but cannot reproduce
the accuracy of a true metering pump, which is normally on the order of ±1% or
better.
512 Hydraulic End

Two types of hydraulic ends are most commonly used:
• Packed Plunger Pump
• Diaphragm Pump
Packed Plunger Pump

The packed plunger pump is the basic design from which all hydraulic-end types
originate. Pumping is achieved by reciprocating motion of a plunger, which draws
in and discharges liquid. Check valves in the inlet and discharge lines direct the
flow in the desired direction (double-ball check valves are the most common and are
recommended). Figure 500-1 shows a typical packed plunger pump.
Pump Manual 500 Metering Pumps
Fig. 500-1 Packed Plunger Pump Courtesy of Marcel Dekker, Inc.
A basic packed plunger pump can deliver up to 30,000 psig with around 1% volu-
metric accuracy. Packed plunger pumps are more suited for high pressures in low-
flow situations than diaphragm pumps. These pumps are also inexpensive and
operate simply. Their disadvantages include the continuous leakage of process fluid
across the plunger packing, which is necessary to cool and lubricate the packing.
They also operate poorly with abrasive fluids.
Diaphragm Pump
Diaphragm pumps provide an effective solution to leakage problems and, to a lesser
extent, abrasive problems. There are two kinds of diaphragm pumps:
• Mechanical Diaphragm Drive
• Hydraulic Diaphragm Drive
Figures 500-2 and 500-3 illustrate the basic principles of each pump.
Fig. 500-2 Mechanical Diaphragm Driver Courtesy of Marcel Dekker, Inc.
Fig. 500-3 Hydraulic Diaphragm Driver Courtesy of Marcel Dekker, Inc.
Although mechanical drives are both simple and cheap, they have a short diaphragm
life and are only suitable for very light duty (e.g., automobile fuel pumps). They
will not be discussed further in this manual.
Hydraulic drives may have a single or double diaphragm. A single diaphragm is the
most common and suitable for most services. However, a double diaphragm may be
necessary for extremely toxic services.
Diaphragms are usually made of teflon or stainless steel, but elastomers or elas-
tomer-coated steel diaphragms are also available. The double diaphragm provides
positive isolation between the process fluid and the drive fluid (hydraulic oil). The
diaphragm interspace may be designed with alarms to alert personnel to contamina-
tion by process fluid (e.g., conductivity) which may indicate outer diaphragm
failure.
To prevent diaphragm overstress, relief valves are incorporated into the drive
system. Figure 500-4 illustrates relief-valve principles.
Fig. 500-4 Relief Valve Principles From The Metering Pump Handbook by McCabe, Lanckton
and Dwyer. Industrial Press, Inc., Copyright 1984. Reprinted with permission
The hydraulic drive system looks similar to a packed plunger pump. However, it has
a number of advantages over the simple plunger pump:
• The plunger works in an ideal fluid (i.e., good lubricity, clean, etc.).
• The hydraulic drive uses relief valves to avoid diaphragm overstress. This
feature is the equivalent of a discharge pressure cut-off.
• It pumps corrosive and abrasive materials with much lower wear rates and
better reliability than packed plunger pumps.
• Field repairs can be made quickly.
513 Proportioning Control

Most pumps are driven by fixed-speed electric motors. Capacity control is achieved
by adjusting the pump stroke. Pump stroke can be adjusted by using a kinematic
mechanism to vary the stroke or a lost motion device to limit the stroke.
Lost-motion devices are suitable only in small power applications (less than
0.5 HP) because they have a tendency to shock-load the mechanism. Lost motion
devices are either mechanical or hydraulic.
Kinematic mechanisms are preferred for stroke adjustment. Figure 500-5 illustrates
kinematic device operating principles.
520 Pump Applications

Actual pumps may have a combination of features and may be arranged with
hydraulic ends in parallel and/or series. Series pumping is rare but required for high
pressures (greater than 1000 psi, approximately). Parallel operation helps achieve
desired volumes.
One major advantage of parallel operation is that the pulsating flow characteristic of
a single unit can be effectively “smoothed” by carefully phasing the drive for each
liquid end.
Figure 500-6 and Figure 500-7 summarize metering pump application criteria.
Fig. 500-5 Kinematic Operating Devices for Proportioning Control Courtesy of Marcel Dekker, Inc.and From The
Metering Pump Handbook by McCabe, Lanckton and Dwyer. Industrial Press, Inc., Copyright 1984.
Reprinted with permission.
Fig. 500-6 Metering Pump Application Criteria

Liquid End Type Recommended for: Not Recommended for:
Packed Plunger • Very High Discharge Pressure • Corrosive Fluids
• Temperature Over 250°F • Abrasive Fluids
• Low Vapor Pressure Fluids • Applications whose trace contamination
of pumpage with packing lubricants is
not permitted
Single Diaphragm • Corrosive Fluids • High discharge pressure (>1500 psi)
(hydraulic drive) • Applications requiring high • Temperatures over 175°F
reliability (elastomer diaphragm)
• Over 250°F (teflon diaphragm)
Double Diaphragm • Very Corrosive or hazardous fluids • High discharge pressure (>1000 psi)
• Abrasive fluids
• Duties requiring guaranteed isolation
from drive oil
• Applications requiring early warning of
diaphragm failure
Fig. 500-7 Comparison of Metering Pumps

Pump Type Advantage(s) Disadvantage(s)
Mechanical Diaphragm • Least expensive. • Accuracy not as good as plunger and
• Can handle most fluids. piston-diaphragm pumps.
• Glandless. • Diaphragm subject to fatigue failure.
• Limited to low pressure deliveries.
Piston Diaphragm • High accuracy. • Diaphragm may be subject to wear if
• Good diaphragm life. pumped fluids include particulate
matter.
• Can handle most fluids.
• Glandless.
• Readily rendered in double diaphragm
form for fail-safe characteristics.
Electromagnetic-Driven • Particularly suitable for micro-metering • Limited capacity.
Diaphragm with precise pulse • More complex control system
operation. (needs digital signal input).
• Glandless.
Abstract
This section covers the four most common types of multiphase pumps that are
commercially available, an overview of manufacturers, upstream applications of
each type, typical application opportunities, the qualification process, economics to
be considered, review of selected installations (with lessons learned), and new
developments as of 2008.
Contents Page
611 Definition of a Multiphase Fluid
612 Most Common Types and Applications of MPPs
613 Purpose of the Manual
614 Upstream Oil Field Applications
615 Refinery Applications
616 What Kinds of Pumps are MPPs?
617 Conclusion
620 Commercially Available MPP Types 600-13
621 General
622 Twin Screw MPP
623 Helico-Axial Pump
624 Progressing Cavity Pump (PCP) or Single Screw Pump
625 Electric Submersible Pump (ESP)
626 General MPP Selection Guidelines
630 Design Recommendations 600-39
631 General
632 Mechanical Seals
633 Screw Sealant for Twin Screw Pumps
634 Rotor Design for Twin Screw Pumps
March 2009  2009 Chevron U.S.A. Inc. All rights reserved. 600-1
600 Multiphase Pumps Pump Manual
635 Effect of Slugs—Gas and Liquid

636 Motor Selection and Use of Adjustable Speed Drives (ASD)
637 Methods to Reduce Sand Erosion
638 Recycle Stream Flashing (Surface Pumps)
639 Suction and Discharge Piping (Surface Pumps)
640 MPP Materials 600-56
641 General
642 Twin Screw Pumps
643 Helico-Axial, PCP, and ESP Materials
644 Mechanical Seal Materials (all types of MPPs)
650 Manufacturers—Twin Screw MPPs 600-59
651 General
652 Bornemann MPPs
653 Leistritz MPPs
654 Flowserve MPPs
655 Recommendations and Comments
660 Manufacturers—Helico-Axial MPPs 600-67
661 General
662 Framo
663 Sulzer
670 Manufacturers—Progressing Cavity MPPs 600-71
671 General
672 Moyno
673 seepex
680 Manufacturers—Electric Submersible (ESP) MPPs 600-75
681 General
682 Centrilift
683 Schlumberger (Reda)
600-2  2009 Chevron U.S.A. Inc. All rights reserved. March 2009
Pump Manual 600 Multiphase Pumps
690 Sizing of the MPP, Its Driver, and Associated Facilities

(Upstream Applications) 600-80
691 Twin Screw MPP Design Guide for Upstream Applications
692 Sizing
693 Pump Sizing Calculation
694 Gas Volume Fraction (GVF)
695 The Recycle, Slip, Flashing Factor for Twin Screw MPPs
696 Pump Efficiency for Twin Screw MPPs
697 Overall Mechanical Efficiency Calculation
698 Volumetric Efficiency Calculation
699 Pump Sizing Spreadsheet for Twin Screw MPPs
6910 Piping and Instrumentation Diagram (P&ID) for a Twin Screw MPP
6911 Separator Sizing for Twin Screw MPPs
6100 Typical MPP Application Opportunities 600-93
6101 General
6102 Upstream Application Opportunities—Listing
6103 Upstream Application Opportunities—Details
6104 Downstream Applications
6110 Technology Qualification Process (TQP) 600-101
6111 Introduction
6112 Chevron TQP
6113 Definition of Qualification
6114 Machinery TQP
6115 Technology Development Stages (TDSs) for MPPs
6116 TQP Summary
6120 Economics 600-108
6121 General
6122 Existing Facilities
6123 New Facilities
6124 Factors to Consider
6125 Examples
6126 Cost Comparison Study for an exceedingly high GVF of 98 Percent—MPPs
versus a Conventional Separation System
6130 Selected MPP Installations (including Lessons Learned) 600-117

6131 General
6132 Humble’s Fluid Flow Test Facility MPP
6133 Trinidad
6134 Mitsue Field, Slave Lake, Canada (later moved to Princess Field in Canada)
6135 Mitsue Pump Moved to the Princess Field
6136 Main Pass 313, Offshore Platform in the Gulf of Mexico
6137 Humble, Texas, Production Field
6138 Duri Trial, Indonesia
6139 COB Facility, El Tigre Field, Venezuela; Boscan Field, Venezuela; Kome,
Miandoum; and Belobo Fields, Chad
61310 Main Pass 59 A
6140 New Developments (as of 2008) 600-141
6141 General
6142 Twin Screw Pump Developments
6143 Helico-Axial Pump Developments
6144 PCP Developments
6145 ESP Developments
6146 Mechanical Seal Developments
6150 Definitions and Acronyms 600-146
6151 Definitions
6152 Acronyms
6160 References 600-150
6161 Company Specifications
6162 American Petroleum Institute (API)
6163 National Association of Corrosion Engineers (NACE)
610 Introduction
611 Definition of a Multiphase Fluid

A multiphase fluid is defined as a fluid that consists of two or more phases in which,
in the most basic case, one phase is a gas and one phase is a liquid. The multiphase
fluid often involves three or more substances, such as crude oil, gas, and water.
Additionally, multiphase fluids can also contain wax, natural gas hydrates, and sand
or other particulate.
A multiphase pump (MPP) handles multiphase fluids, but it can also pump single
phase fluids, such as liquids or gas (for a designated period of time).
Additional definitions used in multiphase pumping are provided in Section 615.
612 Most Common Types and Applications of MPPs

This manual describes the four most common types of MPPs. Other types exist, but
are less common and are not discussed.
The most common types and applications of MPPs are:
• Twin Screw Pumps
• Helico-Axial Pumps
• Progressing Cavity Pumps (PCPs)
• Electric Submersible Pumps (ESPs)
The most common MPP in the petroleum industry is the twin screw type, and it is
usually the type recommended for Chevron applications.
Twin screw and helico-axial pumps have been installed onshore at grade, on
offshore platforms, and on the seabed floor. PCPs and ESPs have been used mostly
in downhole onshore applications and in offshore dry-tree applications supported
from platforms.
Each type has its own particular application considerations. For example, a helico-
axial MPP usually includes a buffer tank ahead of the pump to minimize the effect
of slugs. The other three types do not deploy such a tank.
Most of the MPP applications are upstream, pumping from an oil well or oil field.
Multiphase pumping applications exist in refineries, as well, such as pumping
gaseous liquid from a flare gas knockout drum.
As of the writing of this manual, the installation list for the 94 MPPs that Chevron
(including legacy-Texaco) has installed is shown in Figure 600-1.
Fig. 600-1 Chevron Multiphase Pump (MPP) Applications (1 of 2)
600-6

Loc No. No. GVF Dif. P Max Max
No. Location Country Pur Oper Opertg Years Service Manufacturer Pump Size Flow (bpde) (%) (psi) RPM BHP Drive H2S (ppm)
1 Trinidad Trinidad 1 1 1990 to Gassy Crude Leistritz L4MK 126 to 11,000 82 160 1750 55 Motor ??
Present? 70
Trinidad Trinidad 1 1 1992 to Gassy Crude Leistritz L4MK 164 to 19,500 65 100 1750 76 Motor ??
Present? 72
2 GOMBU Main USA 1 0 1995 to? Gassy Crude Leistritz L4MK 149 to 21,000 95 130 1750 65 Motor ??
Pass 313 (refurbishing) 96
GOMBU Main USA 1 1 2007 to Present Gassy Crude Leistritz L4HK 365 to 151,975 95 440 to 910 to 1134 to Natural Gas ??
Pass 59A 175 640 1200 1707 Engine
3 Mitsue (moved Canada 1 0 1995 to 2002 Sandy, Gassy Leistritz L4HK 200 to 40,000 95 900 3000 680 Motor ??
to Princess) Crude 54
4 Humble USA 1 0 1998 to 2002 Sandy, Gassy Leistritz L4MK 82 to 120 225 Motor ??
 2009 Chevron U.S.A. Inc. All rights reserved.
Test Pump (shut down by Crude 40

operations)
Humble USA 1 0 1997 to 2004 Gassy Crude Leistritz L4NG 13,000 92 75 1750 44 Motor ??
Production (shut down by 126/98
operations)
5 Captain A Great Britain 1 1 1998 to Present Gassy, API 18 Centrilift ESP 10,000 63 660 6600 317 Motor ??
Crude
6 Captain B Great Britain 5 5 2000 to Present Gassy, API 19 Weir HSP 10,000 75 660 6600 317 Hydraulic ??
Crude Turbine
Duri Test Pump Indonesia 1 0 3 months Sandy, Gassy Sulzer MPP7 125,000 75 to 150 3600 700 Motor
Crude 100
7 Duri Indonesia 1 6/98 to 12/98 Sandy, Gassy Bornemann MW 9.5zk to 124,981 75 to 231 1900 349 Motor ??
(moved to (moved) Crude 67 100
Minas)
8 Minas Indonesia 1 0 2000 to 2003 Sandy, Gassy Bornemann MW 9.5zk to 75 to Motor 500
(moved from (shut down by Crude 67 100
Duri) operations)
Minas Indonesia 2 0 6/98 to 12/98 Sandy, Gassy Bornemann MW 8.7zk to 60,075 30 to 110 1800 219 Motor 200
(shut down by Crude 46 99
operations)
9 Midway Sunset, USA 1 0 2000 Diatomacious Bornemann MPC 208 59 to 290 100 Motor 10,000 to
CA (replaced due to Crude 90 20,000
operating
conditions)
Midway Sunset, USA 1 2000 to Present Diatomacious Bornemann MW 7.3zk to 7,381 59 290 1385 63 Motor 10,000 to
CA (moved to (relocated) Crude 33 20,000
Station 222)
Midway Sunset, USA 1 1 2001 to Present Diatomacious Bornemann MW 7.3zk to 7,381 59 to 290 1385 63 Motor 10,000 to
CA Crude 33 90 20,000
Midway Sunset, USA 1 1 2002 to Present Diatomacious Bornemann MW 8.5zk to 70,943 81 270 1760 402 Motor 10,000 to
CA Crude 67 20,000
Pump Manual
March 2009
CA Crude 85 20,000
CA Crude 67DS 20,000
10 Station 222, CA USA 1 1 2001 to Present Gassy Oil/Water Leistritz LMK 186 to 31,000 20 to 100 1150 100 Motor None
Emulsion 110 100
Station 222, CA USA 1 1 2004 to Present Gassy Oil/Water Bornemann MW 7.3zk to 21,000 20 to 100 120 Motor None
(moved from (relocated) Emulsion 33 100
Midway
Sunset)
Fig. 600-1 Chevron Multiphase Pump (MPP) Applications (2 of 2)
March 2009
Pump Manual
Loc No. No. GVF Dif. P Max Max
No. Location Country Pur Oper Opertg Years Service Manufacturer Pump Size Flow (bpde) (%) (psi) RPM BHP Drive H2S (ppm)
11 Indian Colonial, USA 1 1 2001 to Present Gassy Oil/Water Bornemann MW 7.3 100 Motor None
CA Emulsion
12 El Tigre Venezuela 2 2 2001 to Present Gassy Crude Bornemann MW 9.3zk to 60,377 92 220 1522 618 Motor None
(Hamaca) 53
Production
El Tigre Venezuela 7 7 2001 to Present Gassy Crude Bornemann MW 9.3zk to 135,698 70 to 268 1760 937 Motor None
(Hamaca) 90 90
Production
El Tigre Venezuela 6 6 2002 to Present Gassy Crude Bornemann MW 9.3zk to 135,698 92 268 1760 937 Motor None
(Hamaca) 90
Production
El Tigre Venezuela 5 5 2003 to Present Gassy Crude Bornemann MW 9.3zk to 135,698 92 268 1760 937 Motor None
(Hamaca) 90
Production
El Tigre Venezuela 4 4 1998 to Present Gassy Watery Flowserve NLXSHP 30 to 180 1800 600 Motor None
(Hamaca) Crude 5JSR 40
Gathering
Station
El Tigre Venezuela 4 4 2004 to Present Gassy Watery Flowserve NLXSHP 30 to 180 1200 600 Motor None
(Hamaca) Crude 5JSR 40
Gathering
Station
El Tigre Venezuela 4 0 2002 to 2005 Gassy Crude Flowserve NPS14HP 75,017 10 to 800 1787 1600 Motor None
(Hamaca) (Failed)
Transfer
El Tigre Venezuela 5 5 2005 to Present Gassy Crude Bornemann MW 10.6zk 75,017 10 to 800 1787 1600 Motor None
(Hamaca) 20
Transfer
13 Boscan Venezuela 11 11 1998 to Present Gassy, Sandy Warren Colfax GTS 208 20,000 20 to 200 to Motor None
(0.75%) Crude 30 350
Boscan Venezuela 4 4 2003 to Present Gassy, Sandy Warren Colfax GTS 268 40,000 40 500 Motor None
(0.75%) Crude
14 Lake Maricaibo Venezuela 1 1 2005 to Present Gassy, Sandy Weatherford 3,000 30 to 55 Motor None
(<1.0%) Crude 70
15 Kome Chad 3 3 2003 to Present Gassy Crude Leistritz L4HK 330 to 68,454 2 1150 215 Motor None
126
Kome Chad 4 4 2003 to Present Gassy Crude Leistritz L4HK 330 to 156,812 56 175 1150 684 Motor None
214

Miandoum Chad 3 3 2003 to Present Gassy Crude Leistritz L4HK 330 to 59,558 2 815 1150 960 Motor None
100
Belobo Chad 4 4 2003 to Present Gassy Crude Leistritz L4HK 330 to 133,240 38 365 1150 615 Motor None
189
16 Gulf of Thailand Thailand 1 0 2004 to 2008 Gassy Crude Bornemann MPC 268 ?? 99 410 ?? ?? Motor ??
Totals 94 79
600-7
613 Purpose of the Manual

Previously, some Chevron personnel have been reluctant to install MPPs because
they are more comfortable with conventional systems. Among the goals of this
manual are to convey the facts about MPPs, along with their applications and
benefits, and to improve the comfort level of Chevron personnel.
To amplify a bit further, the purpose of this manual is to provide:
• Assistance to the following in understanding MPP technology and how to apply
this technology to drive down the cost of production and the handling of mixed
oil, water, sand, and gas streams:
– Upstream, midstream, and downstream facility engineers
– Reservoir engineers
– Production engineers
– Project engineers
– OPCO discipline engineers
– Project managers
– Business unit managers
– Asset managers
• A text to assist in specifying MPPs;
• An understanding of the economics of MPPs;
• Examples of existing MPP applications;
• Scenarios that point to opportunities for MPPs where they provide an economic
advantage, e.g., onshore, offshore platforms, and subsea;
• A reference tool for people to better learn the technology and economics of
multiphase pumping;
• A teaching tool to be used in conjunction with a “power point” presentation to
various organizations to help realize the economic importance of MPPs.
614 Upstream Oil Field Applications

In an oil field, MPPs are usually located near an individual well or at a manifold,
where production flow lines combine several wells. These pumps may be located
onshore, on offshore platforms, on the seabed floor (subsea), or downhole in an oil
well.
In a traditional surface oil field application, an MPP can replace a conventional
system consisting of:
• Separation vessel
• Gas compressor
• Pump
• Gas production line
• Liquid production line
• Sand or particulate handling system
By comparison, a multiphase pumping system only consists of:

• The MPP itself, with or without a buffer tank, which replaces the separator, the
pump, and the compressor;
• A single combined gas and liquid production line that replaces the individual
gas and oil lines, the gas line being from the compressor discharge and the oil
line being from the pump discharge (refer to Figure 600-2).
Fig. 600-2 Comparison of Conventional versus MPP Design Showing MPP Installation
Advantages
Conventional Separation
Multiphase
In particular, a twin screw multiphase pumping system has the following

advantages:
• One piece of equipment, the MPP, instead of several;
• A single pipeline to the processing facility, instead of two;
• Significantly reduced capital cost;
• Less weight and smaller footprint than a conventional system (important for an
offshore platform, where weight and space are extremely costly);
• For subsea, a simplified seabed support structure.
If gas concentrations and power costs are high, an MPP may not be the most
economical choice compared to the conventional system. Each application must be
evaluated based on its own economical situation.
615 Refinery Applications

MPPs are used in refining to pump multiphase fluids having gas content above
5 percent. The bottoms fluid from a flare knockout drum is one example. MPPs can
also replace centrifugal pumps that are not delivering the desired flowrate due to
entrained or free gas.
616 What Kinds of Pumps are MPPs?

In the Pump Manual, pumps are generally divided into five categories as shown in
Figure 600-3.
Fig. 600-3 Pump Categories (Courtesy of the Hydraulic Institute)
PUMPS
Reciprocating Rotary Positive

Centrifugal Metering Miscellaneous
Positive Displacement
Displacement
In Figure 600-3, the MPPs described in this manual fit into the centrifugal (helico-
axial pumps and ESPs) and rotary positive displacement (twin screw and PCPs)
boxes of the above chart.
Kinetic centrifugal pumps can also be classified as in Figure 600-4.
In Figure 600-4, ESPs and helico-axial pumps fall under KINETIC, Centrifugal,
Turbine Type, VERTICAL TYPE Single and Multi Stage, Deep Well Turbine
(Including Submersibles).
Positive displacement pumps are classified in Figure 600-5.
In Figure 600-5, twin screw MPPs are designated as positive displacement, rotary
pumps, screw, and multiple. MPP PCPs are designated as positive displacement,
rotary pumps, screw, and single.
Fig. 600-4 Kinetic Pump Classifications (Courtesy of the Hydraulic Institute)
Fig. 600-5 Positive Displacement Pump Classifications (Courtesy of the Hydraulic Institute)
3,6721 6,03/(;
+25,=217$/

'28%/(
67($0
$&7,1*
9(57,&$/
3/81*(5 '83/(;
5(&,352&$7,1* +25,=217$/ 6,1*/( 6,03/(;

3,6721
38036 $&7,1*
32:(5
'83/(;

9(57,&$/ '28%/( 3/81*(5 08/7,3/(;
$&7,1*
+25,=217$/ 3/81*(5 6,03/(;

&21752//(' 3,6721 '83/(;
92/80(
0(&+$1,&$//<
9(57,&$/ ',$3+5$*0 08/7,3/(;
&283/('
326,7,9(
',63/$&(0(17 %/$'(%8&.(7 +<'5$8/,&$//<
&283/('
9$1(
5ROOHURU6OLSSHU
%/2:&$6(

$;,$/
3,6721
5$',$/

)/(;,%/(78%(
)/(;,%/( )/(;,%/(9$1(
0(0%(5 )/(;,%/(/,1(5

6,1*/(
527$5<38036 /2%(
08/7,3/(

(;7(51$/
*($5
,17(51$/

&,5&80)(5(17,$/ 6,1*/(
3,6721
08/7,3/(

6,1*/(
6&5(:

08/7,3/(
617 Conclusion
MPP technology is mature. These machines, especially the twin screw type, are
reliable. Many have been operating for years with little operator attention. There are
currently hundreds of installations throughout the world, with many more in
progress and anticipated. Chevron alone has had 94 applications in operation since
1990. Section 6120 lists and describes several of these installations.
Questions or help with a potential MPP application should be addressed to
Bob Heyl at: [email protected].
Facilities Engineering Department (FED)
1400 Smith St.
Houston, TX
Office phone: 713-372-7272
Alternately, access contacts at ETC MEPS team at:
http://etc.chevron.com/fe-mee/machinery/default.asp
620 Commercially Available MPP Types
621 General
This section describes the four most commonly used types of commercially
available MPPs. An MPP is usually installed in an upstream oil field, but they can
also be applied in a refinery wherever liquid and gas are combined in a mixed
stream and transported.
The four types that are described in this manual are summarized in Figure 600-6.
Fig. 600-6 MPP Summary

Max
Max Installed
MPP Type Manufacturers Installed Pressure Max GVF
Capacity Differential (%)
(bpd) (psi)
Twin Screw Bornemann 900,000 1015 100(1)
Leistritz 330,000 1450 100(1)
Flowserve 280,000 1000 100(1)
Helico-Axial Framo 100,000 900 100(1)

Sulzer 650,000 1200 100(1)
Progressive Cavity (PCP) Moyno 60,000 900 40

seepex 50,000 600 40
Electric Submersible (ESP) Centrilift 14,000 5000 60

Schlumberger (Reda) 9,000 4000 60
1 Note: 100% is possible when the pump is supplied with a relatively small amount of externally supplied liquid either as
screw sealant for the twin screw pumps or directly into the flow stream for helico-axial pumps.
Figure 600-7 includes several additional characteristics.
Comparing the Four Types of MPPs

Figure 600-7 is a general overall comparison of the four types of MPPs. These four
types of pumps will be discussed in much more detail in Section 622, Section 623,
Section 624, and Section 625.
Fig. 600-7 Comparison of Four Types of Multiphase Pumps (MPPs)

Type of MPP
Progressing Cavity Electric
Twin Screw Helico-Axial (PCP) Submersible
General Design
Location Surface/subsea Surface/subsea Surface/downhole Surface/downhole
seabed seabed (land or sea) (land or sea)
Mounting orientation Horizontal Vertical/horizontal Vertical/horizontal Vertical
Rotors and Max Temperature
Rotors (C=contacting, N N C N
N=noncontacting)
Max temperature (°F) 600 500 300 with elastomeric 300 to 400
stator
Speed (rpm) 600 to 3,600 2,000 to 7,000 300 to 550 1,800 to 5,400
Case
Removable liners Yes No No No
Bearings Antifriction Antifriction/tilting Antifriction Sleeve
pad
Drivers ASD elect motors, ASD elect motors, ASD elect motors ASD elect motors
nat. gas engines nat. gas engines,
hydraulic turbines
Max GVF (%) 100% with external 100% with special 40% (downhole) 60% with gas
screw sealant design features and separators and
external liquid handlers
supply
Fluid Flow
Slug handling 1 2 5 5
(1=unaffected, 5=greatly (with buffer tank)
affected)
Erosion Resistance 1 2 4 4
(1= very resistant, 5=not
resistant)
Efficiency
Max viscosity (SSU) No limit for 300 SSU No limit for production 300 SSU
production or or refinery services
refinery services
Mechanical Seals
Number 4 sets (single, 2 sets (one single, 1 (single or dual 2 sets (single or
or dual one dual pressurized) dual pressurized)
unpressurized, or pressurized)
dual pressurized)
Maximum capacity (bpd)/ 900,000/1,450 650,000/1,200 60,000/900 14,000/5,000
differential pressure (psi)
Lubrication Self-contained or Externally supplied Self-contained lube oil Pumped fluid
externally supplied lube oil lubricated
lube oil
Reliability
Maintenance Easy (1) or 3 4 2 5
Difficult (5)
Field (F) or Shop (S) repair S S F or S S
Mean time between failure 5+ 3+ 2 to 3 2 to 3
(Years)
Figure 600-8 is a comparison of the advantages and disadvantages of the four types
of MPPs.
Fig. 600-8 Advantages/Disadvantages of Four Types of Multiphase Pumps (MPPs)

Type of MPP
Electric Submers-
Twin Screw Helico-Axial Progressing Cavity (PCP) ible (ESP)
Advantages Good flow and Good flow and One set of mechanical 2 sets of
pressure ranges pressure ranges seals mechanical seals
100% GVF with 100% GVF with Fewer bearings than twin
external liquid screw external liquid, 95% screw or helico-axial
sealant, 95% without
without
Reliable Fairly reliable Can be repaired in the
field
Erosion resistant Erosion resistant Slow speed
Self-contained lube Self-contained lube oil
oil system system
Several wells can Several wells can
feed one pump feed one pump
Pumps heavy Pumps heavy viscosity
viscosity oil easily oil easily
Does not tend to Does not tend to form
form emulsions emulsions
Disadvantages 4 sets of mechanical 2 sets of mechanical Not as reliable as the Not as reliable as
seals seals twin screw or helico-axial the twin screw or
helico-axial
Usually needs a Usually needs a Usually deployed Usually deployed
shop repair shop repair downhole, one pump per downhole, one
well pump per well
Pressurized lube oil Lubricated by the Lubricated by the
system pumped stream pumped stream
Not as reliable as Highly susceptible to Highly susceptible
the twin screw sand erosion to sand erosion
Limited flow range Limited GVF Limited GVF
Viscosity limited to Viscosity limited to
300 SSU light oil
Flow and pressure range Flow and pressure
limited by the well’s range limited by
casing size the well’s casing
size
Tends to form Tends to form
emulsions emulsions
Motor and electric
cable are
downhole causing
many problems
Twin screw pumps and progressing cavity pumps are both classified as rotary
positive displacement pumps, each designed to API 676. However, each supplier
will have his own set of exceptions to this specification. Helico-axial pumps are a
type of centrifugal pump that combines aspects of axial and centrifugal flow. They
are loosely designed in accordance with API 610, but again, suppliers have many
exceptions. ESPs are centrifugal or mixed flow types. Various API Recommended
Practices provide guidelines about sizing, operating, testing, and maintaining. They
are listed in Section 6160.
Twin screws and PCPs will pump a virtually constant capacity, regardless of the
magnitude of the backpressure, as long as the driver has adequate power. If the
pump lacks adequate horsepower to overcome the backpressure, the pump will stall.
A pump with adequate power has the potential to overpressure its case or the
discharge piping. To prevent this, a relief or pressure limiting valve is required
between the pump’s discharge flange and the discharge block valve. Pressure relief
or pressure limiting valves are always installed if these pumps are used.
Unlike rotary positive displacement pumps, helico-axial and ESP pumps use
centrifugal force and high velocities to increase pressure at the pump discharge
flange. At zero flow or “shutoff”, these pumps provide their greatest discharge
pressure. If this pressure is greater than the design pressure of the discharge pipe, a
pressure relief valve must also be installed with these pumps.
For additional information, refer to each type of pump in its specific subsection of
this Section 600 and also in other sections of the Pump Manual. An electronic
version of the entire Pump Manual can be found at the Chevron Engineering
Standards website.
Both the twin screw pump and the helico-axial pumps are surface machines located
onshore, on offshore platforms, or subsea. They are not deployed downhole.
Though the helico-axial pump can handle a higher pressure boost than the twin
screw pump, the twin screw pump is superior in almost every other category. Based
on the complicated nature of the helico-axial pump and Chevron’s experience at
Duri, the twin screw pump has proven itself as the better choice for most
applications. See Section 6138, Lessons Learned for additional reasons. By a wide
margin, twin screw pumps are installed in more surface multiphase applications
than any other type.
The main advantages of a PCP over the twin screw pump and the helico-axial pump
is that it involves fewer mechanical seals, has fewer bearings, and can be repaired
more easily in the field. A PCP’s maximum flowrate and pressure boost is lower
than the twin screw pump or the helico-axial pump. Because the PCP’s rotor
continuously contacts the stator, the PCP will wear, and reliability suffers.
ESPs and PCPs are usually installed downhole in one well. No matter what
installation orientation (vertical or horizontal), they are not considered reliable.
Their capacity is designed to match the production of the well, which is definitely
lower than the capacity of a twin screw pump or a helico-axial pump. An ESP’s
maximum pressure boost may be higher than any of the other three pumps.
Most all of our industry’s surface MPP experience has been with twin screw pumps.
Indeed, as of 2008, there exist throughout the world 600 twin screw pumps and only
50 helico-axial pumps. Chevron (and legacy-Texaco) experience has also been
almost entirely with twin screw MPPs, having installed 88. The initial twin screw
pump installations were carefully monitored by Chevron to learn as much as
possible from each. They are listed in Figure 600-9, and these installations and
others are referred to throughout this manual as examples.
Fig. 600-9 Monitored Twin Screw Pump Installations

Location or Year
Oil Field Installed Manufacturer Type Comments
1. Humble’s Flow 1992 Leistritz Twin Screw Tested heavy and light oil with
Test Facility, Texas varying amounts of water, sand,
and GVFs.
2. Trinidad 1992 Leistritz Twin Screw Pumped a multiphase fluid with an
appreciable amount of sand.
3. Mitsue Field, 1993 Leistritz Twin Screw Experienced severe slugging from
Canada crude oil wells a mile away. Moved
to the Princess Field.
4. Princess Field, 1995 Leistritz Twin Screw Moved from Mitsue. Pumped
Canada multiphase fluids 20 miles to
processing facility.
5. Main Pass 313 1993 Leistritz Twin Screw Decreased wellhead pressure
Platform, Gulf of increasing production.
Mexico
6. Humble Oil Field, 1997 Leistritz Twin Screw Decreased wellhead pressure
Texas increasing production. Damaged
beyond repair in 2006 when over
pressured from well rework.
7. Duri Field, 1998 Bornemann Twin Screw Tested with a Sulzer helico-axial
Indonesia pump to determine which type to
buy for the Minas Light Oil Steam
Flood project.
8. Duri Field, 1998 Sulzer Helico-axial Tested with a Bornemann twin
Indonesia screw pump to determine which
type to buy for the Minas Light Oil
Steam Flood (LOSF) project.
9. El Tigre Field, 2005 Bornemann Twin Screw Five Bornemann pumps replaced 4
Venezuela unreliable Flowserve twin screw
pumps.
10. Main Pass 59A 2007 Leistritz Twin Screw Boosts wellhead pressure to match
Platform, Gulf of system pressure. The 1,700 HP
Mexico natural gas engine is the largest
currently driving an MPP.
622 Twin Screw MPP

General
Twin screw pumps are by far the most common type of surface-installed MPP and
the design usually recommended for most petroleum industry applications,
including Chevron’s. These pumps are covered in the third edition of API 676,
expected to be published in 2009.
Twin screw pumps can be deployed in a refinery or in an upstream oil field whether
onshore, on a platform, or subsea on the seabed floor. They are mounted
horizontally, though attempts at vertical installation downhole in an oil well have
been attempted. Unfortunately, none of these downhole installations have proven to
operate reliably. In oil fields, they pump multiphase fluid from a single well or from
a discharge manifold supplied by several wells.
An external view of an MPP twin screw pump is shown in Figure 600-10, and an
internal view showing the major components is shown in Figure 600-11.
Fig. 600-10 External View of Twin Screw MPP in Minas, Indonesia
Rotors and Stators

As Figure 600-11 shows, twin screw pumps consist of two rotors side by side, held
apart by bearings. The screws are either an integral part of the shaft or manufactured
separately and heat shrunk or pressed onto the shaft.
Fig. 600-11 Internal View of Twin Screw MPP (Courtesy of Leistritz Pumps)
The MPP rotor is stiffer than the rotors in a pure liquid twin screw pump. Unlike a
liquid pump, there is no contact between the MPP screws and the stator or case. The
possibility of contact is further reduced by increasing the clearances to compensate
for either gas slugs or particulate. (Refer to Section 634 for design recommendations.)
Speed
The speed of the pump is usually 600 rpm to 1,800 rpm, but they have run reliably
at 3,600 rpm to achieve higher capacities. Design speed is a function of the service
and pumped fluid characteristics. For example, the Chevron Princess Pump ran
reliably at 3,600 rpm for years. (Refer to Section 6135.)
Case
The twin screw case is robust. If particulate is present, the case is usually bored to
accept a replaceable liner.
Bearings
The twin screw uses radial antifriction bearings. Thrust bearings are theoretically
not required. However, one of the radial antifriction bearings is designed to handle a
minor thrust load that can be present during heavy slugging.
Drivers
The most common twin screw driver is an electric motor with an adjustable speed
drive (ASD) or often, more specifically, a variable frequency drive (VFD). Other
drivers used are natural gas or diesel engines. Twin screw MPPs are not usually
driven by steam or gas turbines. Steam is not usually available at the right
conditions in the upstream environment, and the speed, expense, or horsepower of
gas turbines is usually too high for twin screw operation. The driver is coupled to
one rotor, and the second rotor is driven by the first through a set of gears at one end
of the rotor set.
Gas Volume Fraction (GVF)

Gas volume fraction is the volume of gas as a percent of the total volume of all the
fluids (gas and liquid) at pump suction conditions. A properly designed and applied
twin screw pump can pump from 0 to 100 percent GVF indefinitely, with an
adequate and reliable supply of a screw sealant. (Refer to the Screw Sealant item at
end of this section.)
Fluid Flow
As shown in Figure 600-11, multiphase fluid enters the twin screw pump at its
center. The fluid is split into two equal streams, each stream channeled to opposite
ends of the case where the fluid enters the screws at suction pressure. The fluid is
then pushed to the discharge at the pump’s center. Because the flow is split, with
virtually equal portions entering the opposite ends of the screws at the same time,
thrust on the pump is theoretically zero. Slugs of liquid following slugs of pure gas
are also hydraulically balanced and will not damage the pump.
Erosion
Sand and other particulates can cause wear in twin screw pumps by wedging into
the clearances between rotors and between the screw edges of each rotor and the
case or stator liner. As discussed, these clearances can be adjusted to reduce erosion.
Sand in the crude oil may not necessarily cause erosion if the crude oil’s viscosity is
high enough and/or the gravity is heavy enough (roughly an API gravity of
30 degrees or lower). For example, a test at Duri, Indonesia, in 1998 showed that
sand flowed through the pump without appreciable wear. (Refer to Section 637 and
Section 638, especially Lessons Learned, item A.)
Efficiency
There are two definitions of efficiency: mechanical efficiency and volumetric
efficiency. Mechanical efficiency, expressed as a percent, is simply the theoretical
power needed to pump a specified flow and pressure, divided by the actual power
delivered by the driver to the pump’s shaft. Volumetric efficiency is the amount of
volume delivered by the pump, divided by the theoretical total amount of volume
transferable, again expressed as a percent.
Both efficiencies suffer if the internal clearances of a twin screw pump are
increased. Larger clearances increase “slip”, defined as the amount of fluid that
leaks backward from the high pressure side of each screw flight to the low pressure
side. Larger clearances increase slip and reduce efficiency.
Pump efficiency can decrease from:

• Increased clearances to compensate for particulate;
• Increased clearances to compensate for the expected high operating tempera-
ture, because the MPP handles gas (high GVFs increase the pumping tempera-
ture more than low GVFs, requiring greater clearances);
• Low viscosity of gas that slips more easily than liquid;
• A required high pump differential pressure that increases the “slip” within the
pump;
• Pumped liquid when its viscosity is low.
Pump efficiency can increase from:
• High viscosity liquid (above API 30 gravity) that seals the clearances better
than a low viscosity liquid;
• Increasing the pump speed, which decreases “slip” and improves efficiency.
Mechanical Seals
The process fluid in the pump is separated from the outside atmosphere by
mechanical seals that, in most designs, seal against pump suction pressure. Except
for startup conditions, this pressure is usually low, a feature that directionally
improves the seal’s reliability.
Each twin screw pump has four sets of mechanical seals, one set at the end of each
screw. For a dual seal arrangement, this amounts to eight single seals in one pump.
Mechanical seals are further discussed in Section 632.
Seals using the following API 682 seal flush plans have been found to be very
reliable:
• API Seal Flush Plan 11, or 31, or 32 for single seals;
• API seal plan 52 for unpressurized dual seals with a buffer fluid;
• API Seal Flush Plan 53A, or 53B, or 53C, or 54 for pressurized dual seals with
a barrier fluid.
The type of seal and its required seal flush plan are selected in accordance with the
specific field conditions.
Maximum Capacity and Differential Pressure

The maximum designed twin screw MPP capacity currently available is
approximately 900,000 bpd. The maximum differential pressure is 1,450 psi. These
maximums are not likely to be achieved simultaneously or in the same pump, since
they are affected by:
• GVF percentage (the higher the GVF, the lower the rates and the differential
pressure achievable);
• Viscosity of the liquid (the higher the viscosity, the higher the capacity and
differential pressure achievable);
• Amount and size of the particulate (the larger the particulate size, the larger the
clearances and the amount of slip, reducing the pump’s capacity and differen-
tial pressure achievable).
Lubrication
Twin screw pumps use antifriction radial bearings, one of which is capable of
resisting some thrust. They are enclosed in their own bearing housing. Usually each
bearing housing contains its own separate lubricating oil reservoir, although a
centralized forced feed lubrication system is also used on occasion.
Chevron requires the lubricating oil to stay below 180°F (82°C) under the worst
pumping and atmospheric conditions. Mineral oil tends to oxidize, and for every 10°
above 180°F (82°C), the life of mineral is cut in half. There are three ways to
alleviate the problem if the atmospheric temperature is hot:
• For a self-contained bearing housing, this can be accomplished with an air
cooled bearing housing by attaching a fan to the pump shaft. For example, the
Duri pump operated under hot conditions, and its oil temperature stayed below
180°F (82°C) by using an air cooled bearing housing fan.
• A centralized forced lubrication system can also be designed to keep the oil
below 180°F (82°C). For example, the MPP in the Mitsue field used a force
feed circulating system.
• Finally, synthetic oil that allows for a higher operating temperature will work
satisfactorily. However, this could lead to mistakes if operators were to add
mineral oil as makeup oil instead of the synthetic oil.
Installation of an RTD or thermocouple in the lube oil of each bearing housing or
touching the outer race of each bearing is recommended. If installed at the outer
race, the bearing temperature will operate approximately 20° hotter than the lube
oil, and the limit should be increased to 200°F (93°C).
Reliability
If properly designed and applied, twin screw pumps have proven to be reliable,
normally operating longer than 5 years.
Screw Sealant
A twin screw MPP always needs liquid to seal the clearances between the screws
and the case. Without this liquid, and if not shut down by its high temperature
shutdown instrumentation, the MPP will stop pumping, heat up, and seize. This
liquid is called screw sealant, and the amount needed is approximately 4 to
5 percent of the pump’s capacity. This liquid must be contained in the suction flow
stream or injected either into the suction line or at the ends of each screw at the
mechanical seals, such as when supplied by the seal flush liquid.
623 Helico-Axial Pump

General
Similar to twin screw pumps, helico-axial pumps are found onshore at grade, on
offshore platforms, and subsea. They pump multiphase fluid from a single well or
from a discharge manifold fed by several wells. These pumps can be mounted in the
horizontal or vertical positions.
In this pump design, the pump increases pressure using several stages of open
impellers that resemble augers or screws. Multiphase fluid moves from one stage to
another, helically and axially (parallel to the pump shaft).
The external view of a helico-axial pump is shown in Figure 600-12, and an internal
view showing the major components is shown in Figure 600-13.
Fig. 600-12 External View of Helico-axial Pump Installed in Legacy-Texaco’s Humble Flow Facility, Before Installation in
Duri, Indonesia (1998)
MPP
7-stage helico-axial
125,000 bpd
700 hp motor w/ VFD (3600 rpm)
There are approximately 42 helico-axial pumps throughout the world. At the time of
this writing, Framo has 18 installed subsea pumps on the seabed floor, mostly in the
Norwegian sector of the North Sea, plus 2 onshore and 2 on offshore platforms.
Sulzer has built 20 pumps, of which 17 are installed onshore, 2 on offshore
platforms, and 1 subsea.
Rotors and Stators

As can be seen in Figure 600-13, the rotating element of a helico-axial pump
consists of several compression cells, each of which is composed of a rotor attached
to the pump’s shaft and a stator attached to the case. The stator redirects the flow
into the inlet of the next rotor or compression cell. The number of cells depends
upon the differential pressure required.
Fig. 600-13 Internal View of Helico-axial Pump (Courtesy of Sulzer Pumps)
Speed
The speed of these pumps varies from 2,000 rpm to 7,000 rpm. Design speed is a
function of the service and the pumped fluid characteristics.
Case
The case is a pressure vessel like any multistage centrifugal pump. Unlike the twin
screw pump, the case does not have a replaceable liner.
Bearings
Radial loads are supported by hydrodynamic or sleeve bearings if the design speeds
are greater than 3,600 rpm. At or less than this speed, the bearings are of the
antifriction type. Thrust bearings are usually tilting pad or “Kingsbury” type for any
speed.
Drivers
For most common applications, the pumps are driven by an electric motor.
However, a natural gas engine or a diesel engine can also be used for surface
applications. For subsea, a hydraulic drive or an electric motor is commonly used.
To reach the higher operating speeds, a gearbox, variable speed coupling, or, for the
electric motor, an ASD is required.
As with the twin screw pump, changing the speed allows the pump to operate at
different capacities. The difference with this pump versus the twin screw pump is
that changing the speed also changes the differential head that the pump can
produce.
Gas Volume Fraction (GVF)

A properly designed and applied helico-axial pump can pump from 0 to 100 percent
GVF indefinitely with an adequate and reliable supply of a liquid to the inlet. Note
that a pump designed for a low GVF may not pump at all at a high GVF. (Refer to
Section 6138, especially Helico-axial MPP Test Results.)
Fluid Flow
Multiphase fluid moves from one stage to another, helically and axially (parallel to
the pump shaft). Between each impeller is a stator (diffuser) or set of vanes attached
to the pump case that directs the fluid from the discharge of one impeller to the
suction of the next. The impeller openings become progressively smaller with each
successive stage toward the discharge to compensate for the compression of the gas.
Erosion
Sand and other particulates cause wear in a helico-axial pump due to the angle of
impact with the rotor and stator. This angle of impact changes depending upon the
speed and the flowrate. To reduce wear, the pump is run as slowly as possible and at
its best efficiency point. Also, helico-axial pump manufacturers coat their rotors and
stators with a hard material, such as tungsten carbide, or they gas harden them with
nitride or boride gas to minimize erosion.
Efficiency
The mechanical efficiency of a helico-axial pump is usually lower than that of the
twin screw pump. This is especially true at higher viscosities (greater than
300 SSU), where centrifugal pump efficiencies fall off dramatically. For example, at
Duri, the mechanical efficiency of the helico-axial pump was 22 percent, while that
of the twin screw pump was 45 percent. (Refer to Section 6138.)
Mechanical Seals
In Figure 600-13, the mechanical seal system consists of a mechanical seal at both
the motor and outboard ends.

The maximum pump capacity is approximately 650,000 bpd with multiphase fluid
measured at inlet conditions. Maximum pump differential pressure is approximately
1,200 psi. As with the twin screw pumps, these maximums are independent of each
other and are not likely to be achievable simultaneously or even with the same pump.
These maximums are affected by:

• Viscosity of the liquid (the lower the viscosity, the higher the rates and differen-
tial pressures achievable);
• Percent gas in the fluid (GVF). (The higher the GVF, the lower the rates and
differential pressures achievable. For example, a helico-axial pump produces
half the differential pressure if the GVF increases from 80 to 90 percent and
half of that if it increases to 95 percent.)
Lubrication
A circulating lubrication system is required to cool and lubricate all the bearings. It
usually consists of a reservoir, lube oil pump, cooler, filters, and associated alarms
and other instrument devices.
Reliability
The helico-axial pump is complicated by its circulating lubrication system and high
speed. These complications have been known to affect pump reliability. For
example, the side by side twin screw and helico-axial pump tests at Duri
demonstrated how this complexity can cause problems. (Refer to Section 6138.)
Buffer Tank
Slugs of liquid or slugs of gas are common for any MPP stream, but unlike the twin
screw pump, in which the flow is split to balance slugging, the helico-axial pump
often uses a “buffer” tank (actually a vessel) installed ahead of the pump. The
“buffer” tank dampens the effect of any slug and allows the pump to operate with a
lower thrust bearing capacity. The buffer tank also causes the gas and liquid to
become more homogeneous before entering the pump and provides some residence
time enabling particulate to settle out. Figure 600-14 shows such a buffer tank. Note
the stand pipe in the center of the tank, which has holes for gas to enter into the
liquid stream going through the stand pipe.
Fig. 600-14 “Buffer” Tank (Courtesy of Sulzer Pumps)
624 Progressing Cavity Pump (PCP) or Single Screw Pump

General
Multiphase PCPs are usually installed horizontally on the surface. They can also be
installed vertically, downhole in a single well, driven by a long shaft with an electric
motor from the surface. Like the twin screw pump, the PCP is a positive
displacement (PD) pump. The detailed comments made about PD pumps in the twin
screw section apply to these pumps, as well, including overpressuring potential and
the need for a pressure limiting valve in the discharge piping before the first
discharge block valve.
External views of the progressing cavity pump are shown in Figure 600-15
(horizontal orientation) and Figure 600-16 (vertical orientation). Internal views of
the major components are shown in Figure 600-17 and Figure 600-18.
Fig. 600-15 External View of PCP (Courtesy of Tarby Pumps) Fig. 600-16 External View of PCP in Downhole
Orientation (Courtesy of seepex)
Fig. 600-17 Internal View of PCP (Courtesy of Tarby Pumps)
Fig. 600-18 Internal View of PCP in Downhole Orientation (Courtesy of Robbins and Myers, Inc.” makers of
Moyno pumps)
Rotor and Stators (Including Elastomeric Stators)

The progressing cavity pump is essentially a single, serpentine screw rotor that
revolves inside a stationary liner called the stator. With rare exceptions, the stator is
an elastomer, while the rotor is a hardened steel alloy. Unlike the screws in a twin
screw pump, the rotor is in intimate contact with the stator. Continuous lubrication
between the two is absolutely necessary to prevent excessive heat buildup between
the rotor and the stator and corresponding wear.
Tests for elastomeric decompression and swelling must be run to determine the
compatibility of the elastomeric stator with the pumped fluid and to avoid excessive
heat buildup.
Elastomeric Decompression. In some cases, the elastomeric stator becomes
infused with the pumped gas during normal operation. If this occurs, the stators can
be destroyed from the explosive decompression of the infused gas as the pump is
de-pressured. To prevent this and to decide if a proposed stator is acceptable, a
compatibility test is required before finalizing the selection of the elastomer. The
test involves placing an actual sample (usually crude oil and gas) in an enclosed
pressurized chamber with the proposed elastomer. After an elapsed period of time,
the chamber pressure is lowered to determine if any of the infused gas destroys the
elastomer.
Elastomeric Swelling. The elastomeric stator can also swell from the multiphase
fluid’s chemistry. When purchasing a PCP, the amount of elastomeric swelling must
be tested for fluid compatibility. This swelling, if any, increases the contact pressure
between the rotor and the stator and must be considered in sizing the bore of the
stator. If the allowance for swelling is not enough, the contacting force will be
excessive, overheating the stator and causing the pump to fail. If the design
allowance for swelling is too large, the pump will not reach its design capacity.
Elastomer selection is therefore a major concern. Choosing an elastomer that is
compatible with the pumped fluid, preferably with no gas infusion and no swelling,
is vital to the life and performance of the pump.
Speed
PCPs run at slow speeds, usually from 300 rpm to a maximum 550 rpm. This slow
speed is required due the intimate contact between the rotor and the stator.
Case
The case material for these pumps is usually carbon steel, but other materials can be
provided, if required, and 316 SS is often recommended. With one rare design
exception, the case is lined with the elastomer. (Refer to General in Section 624 and
Metal to Metal PCP in Section 6144.)
Bearings
One set of antifriction bearings is provided on the motor end for radial and thrust
located between the mechanical seal and the coupling.
Drivers
These pumps are usually electric motor driven with either an ASD or a gear set to
reduce the speed and to adjust the capacity. Like the twin screw pump, the ASD
allows the PCP to operate at different speeds and different capacities, enabling it to
match the well’s production from its initial flowrate throughout its lifetime.
GVF
A PCP cannot operate without liquid. If this occurs, the rotor will grab the
elastomeric stator, causing intense friction and heat, which will immediately destroy
the stator. For this reason, and based on Chevron experience, a PCP should be
limited to a 30 percent GVF (unless an adequate and reliable external or recyled
liquid stream is introduced into the suction, in which case, the inlet stream’s GVF
may be raised to 40 percent). (Note that some supplier’s marketing documents
promote using the pumps at GVFs as high as 99 percent in research tests with
specialized elastomers for the stator and coatings for the rotor to reduce friction and
heat. Due to Chevron’s experience with stator failures from transient conditions in
the field which caused the pump to run dry, it is strongly recommended that the
GVF be limited to a maximum of 40 percent.)
Fluid Flow
Fluid is displaced from inlet to outlet as the gas/liquid mixture is trapped in cavities
that are formed between the rotor and the stator. Unlike the helico-axial pump, one
of the advantages of the PCP is that the flow is uniform, with low shear, meaning
that tight emulsions will not be formed by the PCP.
Erosion
Particulate in the pumped stream will become partially embedded in the elastomeric
stator. The particulate will contact the rotor with each revolution, causing wear, and
increasing the slip, with the resulting loss of the pump’s capacity and differential
pressure.
Efficiency
As with all PD pumps, the volumetric efficiency is high until wear and slip occur.
Mechanical Seal
A surface PCP can use a single seal or a pressurized dual seal. In some of the earlier
applications, packing was used by companies other than Chevron. Packing is not
recommended due to the required leakage of the pumped fluid as lubricant for the
packing. This leakage causes safety and environmental concerns. Also, if the
packing fails, gas and liquid will be released, causing an even more severe concern.

The maximum capacity of a surface PCP MPP can be as high as 60,000 bpd, and the
maximum differential pressure can be 900 psi, again not with the same pump. The
reasons are the same as previously described for the twin screw and helico-axial
pumps.
Lubrication
PCPs usually have one set of bearings contained in a bearing housing and lubricated
by oil or grease. In a vertical downhole application, the shaft is centered in the
tubing by bushings which are product lubricated.
Reliability
The mean time between failures averages 2 years to 3 years. The most common
failure mode is the elastomeric stator, followed by mechanical seals. Unlike a twin
screw pump or a helico-axial pump, a PCP pump is often repaired in the field.
625 Electric Submersible Pump (ESP)

General
An ESP is a multistage, centrifugal or axial flow pump that is almost always
installed downhole in an oil well. The ESP pump itself typically consists of many
small diameter impellers on a shaft, with product lubricated sleeve bearings between
each impeller. The entire unit can be 100 or more feet long.
ESPs have been installed on the surface, in a horizontal position, but this is NOT
recommended as their reliability in this has been unsatisfactory. An ESP fits in an
oil well caisson, and therefore, its diameter and capacity are restricted. The suction
is from the well through a screened inlet, and the discharge flows directly into the
production tubing.
An external view of a vertical downhole electric submersible pump is shown in
Figure 600-19, and an internal view is shown in Figure 600-20. Not all multiphase
ESP pumps will necessarily have all these components.
Fig. 600-19 External View of Vertical Electrical Submersible Pump (Courtesy of Schlumberger)
Fig. 600-20 Cross-Sectional View of Vertical Electrical Submersible Pump (Courtesy of Schlumberger)
Rotors and Stators (including Gas Separators and Gas Handlers)

An ESP assembly for a multiphase application, from bottom to top, consists of an
electric motor, a protector (seal), a gas separator, often a gas handler, and the ESP
pump itself. Like helico-axial pumps, the number of ESP impeller stages is chosen
to meet the design differential pressure. To do this, the pumps generally require
numerous stages. All this makes the overall length of the pump and motor long,
over 100 feet in some cases.
ESPs can be outfitted with a variety of devices to separate the gas from the
multiphase fluid stream. A gas separator is usually a centrifugal or rotary device
installed between the motor and the pump. It vents the separated gas into the
annulus between the caisson and pump above the liquid level.
A gas handler can be connected between the gas separator and the ESP. It increases
the suction pressure to the ESP, re-liquefying some of the gas and lowering the GVF
to the ESP. It also decreases the size of the remaining gas bubbles and homogenizes
the mixture.
The protector is described in this section in Bearings and in Protector or Seal
System.
Speed
Speeds range from 1,800 rpm to 5,400 rpm using an ASD.
Case
Case diameter and therefore the impeller diameter is determined by the size of the
caisson. Case material is normally carbon steel which can be coated or replaced
with proprietary materials. (Refer to Section 680 for details.)
Bearings
ESPs have sleeve bearings that are lubricated by the fluid being pumped. In the
preferred design, a bearing exists between each impeller, and therefore, the number
of bearings is equal to the number of impellers, as many as 300 in some cases. This
bearing arrangement is only in the preferred design and must be specified. Unlike
the ESP itself, the electric motor is oil lubricated, made possible by a “seal bag” in
the protector or seal system section. Included below the “seal bag” is the pump’s
thrust bearing, also lubricated by oil.
Driver
An ESP is driven by an electric motor, also installed downhole. The ESP motor is
installed below the pump, with the pumped fluid flowing around the motor before it
enters the pump’s suction. This design is intended to provide needed cooling for the
motor. However, this can cause problems, as the motor will not be adequately
cooled by streams containing a high GVF or streams containing a high percentage
of particulate. With high GVF streams, the gas provides insufficient heat transfer,
while for high particulate streams, the particulate coats or packs around the motor,
severely restricting heat transfer. For this reason, most ESP MPP manufacturers
recommend that the pump be controlled or shut down based on high motor
temperatures instead of traditional high motor amperes.
As with the other pumps described earlier, ESPs are often controlled by ASDs. The
ASD allows the pump to operate at different speeds and different capacities. As with
any centrifugal pump, controlling the pump speed means the pump will operate on a
different head/capacity curve, which can present problems for operations.
GVF
Based on Chevron experience, the GVF for an ESP (without special gas separation
devices) should be limited to 40 percent, even though manufacturers’ marketing
brochures indicate that they can handle GVFs as high as 90 percent with specialized
gas separators and gas handlers. Based on Chevron experience, if these devices are
included for actual field operations use, the GVF should be limited to 60 percent.
Fluid Flow
Fluid flows vertically upward, past the motor into the gas separator, then into the
gas handler (when supplied), and finally into the ESP itself. The fluid gains pressure
through each impeller and diffuser stage. After leaving the last stage, the fluid
discharges into the production tubing.
Erosion
Particulate causes erosion to the impellers and, to a larger extent, to the bearings,
causing them to wear rapidly. Mean time between failures can be as low as a few
months, depending on the amount of sand being pumped.
Efficiency
The mechanical efficiency of an ESP is usually lower than the twin screw pump.
This is especially true at higher viscosities (above 300 SSU), at which centrifugal
pump efficiencies fall off dramatically.
Protector or Seal System

As mentioned, a protector exists between the motor and gas separator. Located in
the protector is a “seal bag” that protects and segregates the motor’s lubricating oil
from the produced fluid, and it also equalizes the motor’s pressure to that of the
pump. Also, the protector houses an oil lubricated thrust bearing to handle the thrust
from the ESP.

ESP stages are added or subtracted such that the pump’s pressure matches
production pressure requirements. Increasing the impeller diameter increases the
pump’s capacity. However, the impeller diameter is limited by the caisson size.
Pumps can be designed for increased speed to raise their capacity. Maximum rates
are 14,000 bpd, and one of the manufacturers claims that its ESP can develop a
differential pressure up to 5,000 psi.
Reliability
Typically, ESPs are pulled from the well every 2 years to 3 years for required
maintenance. Despite attempts to improve reliability, ESPs are not reliable for a
variety of reasons, some of which are as follows:
• ESPs experience electrical cable shorting problems, motor problems, bearing
problems, seal bag problems, and other mechanical problems all made worse by
higher downhole temperatures, corrosion, erosion, and gas.
• They are often installed in a deviated or slanted well. Installing them in nonver-
tical orientations causes the bearings to wear unevenly and more quickly.
• As previously discussed, radial sleeve bearings (as many as 300 in some cases)
are lubricated by the fluid being pumped. If the fluid contains particulate, the
particulate enters the bearing with resultant wear.
• If the fluid has a high GVF, the bearings will not be adequately lubricated,
causing premature failure.
Horizontal installation of these pumps negatively affects the reliability, as well. The
horizontal position, the long pump shaft length, and the increased probability of
misalignment cause the bearings to wear unevenly and more rapidly than if located
in the vertical position.
626 General MPP Selection Guidelines

As stated earlier, an MPP is a single piece of equipment used to pump oil, water,
gas, and sand. It is most often the economical choice for liquid with a GVF greater
than 5 percent.
The flowcharts shown in Figure 600-21, Figure 600-22, Figure 600-23, and
Figure 600-24 provide a general guideline for selecting an MPP for a particular
generic application. They are meant to provide some guidance but are not to be
taken as the final determining factor in the selection of a particular type of MPP. If
an MPP installation is considered, an ETC specialist or a local expert should be
consulted.
Fig. 600-21 Pump Selection Flowchart
I need an MPP
Y=Yes N=No
Y Onshore N
or
Platform
Y N
Subsea
Y
Seabed
Y N
Downhole
N
Downhole
Y Depth
< 5500 ft
Above
Y Ground
N
Wet Tree
Y
Y
See SME
N
ESP Go to Sheet C
Go to Sheet C Go to Sheet A
Fig. 600-22 Pump Selection Flowchart Sheet A
Sheet A
Viscosity at
Y All Design
Conditions
< 300 SSU
Y Viscosity at N
Any Design
Condition
> 300 SSU
Y Particulate N
> 0.01 wt%
N
Y DP < 5,000
Go to psi
Sheet B
Y N
GVF < 95%
Y Flowing T N
< 300 F
Y Total Flow N
< 500,000
bpd
Helico-axial Specially Designed

Helico-axial System
See SME
Fig. 600-23 Pump Selection Flowchart Sheet B
Sheet B
Y
DP < 1,400 psi
Y
GVF < 95%
Y FT < 600F
Total Flow
< 300,000
bpd N
Y Specially Designed
Twin Screw System
See SME
Twin Screw
Fig. 600-24 Pump Selection Flowchart Sheet C
Sheet C
Y Total Flow
< 17,000 bpd
N
Y
GVF < 40%
N
Y
T < 300F
N Y N
GVF < 60%
DP < 900 psi

N
Y
ESP See SME
PCP or ESP
630 Design Recommendations
631 General
Most of Chevron MPP experience has been with twin screw pumps because they are
the most versatile, reliable, and appropriate pumps for most Company applications.
While most of the recommendations in this section pertain to all types of MPP
designs, some apply only to twin screw pumps and are so noted.

The subject of mechanical seals is complex. The following is a brief discussion of
mechanical seals as they apply to MPPs.
Mechanical seals used in MPPs should be purchased in accordance with Chevron
specification PMP-SC-4662, which modifies API 682. API 682, Annex A, provides
a good tutorial on seals, their usage, and selection procedures. Both the Chevron
standard and API 682 discuss aspects of mechanical seals, including materials, seal
part codes, flush plans, auxiliary hardware, and more.
Chevron has a PSA with both John Crane and Flowserve. These agreements include
a seal selection guide. A copy of the guide exists in the Chevron Pump Manual—
Engineering Guidelines (Gray Book), Section 850. Bergmann also makes a quality
mechanical seal and has extensive experience applying them to MPPs.
The following types of mechanical seals and API seal plans are commonly used in
MPPs (refer to API 682 for details):
• Single Seals (using an API Seal Flush Plan 11 as shown in Figure 600-25, Seal
Flush Plan 31 as shown in Figure 600-26, or Seal Flush Plan 32 as shown in
Figure 600-27).
• Unpressurized Dual Seals (with a buffer fluid using API Seal Flush Plan 52 as
shown in Figure 600-28).
• Pressurized Dual Seals (with a barrier fluid using API Seal Flush Plan 53A as
shown in Figure 600-29 or Seal Flush Plan 53B as shown in Figure 600-30 or
Seal Flush Plan 53C as shown in Figure 600-31 or Seal Flush Plan 54 as shown
in Figure 600-32).
The following paragraphs illustrate examples of the various seal types. They are
termed examples, since many different seal arrangements of each type are possible.
The specific arrangement depends on such conditions as the amount of particulate,
the temperature, and the hazardous composition (if it is hazardous) of the pumped
stream.
For example, in a single seal, the normal seal with several small springs (pusher
type) may be replaced by a bellows seal designed for high temperatures. In another
case, due to the amount of particulate in the stream, the standard Type C stationary
bellows assembly may need to be switched to one that rotates so the bellows does
not get clogged with particulate.
Single Mechanical Seals

All mechanical seals rely on liquid to lubricate and cool the seal faces. Without a
thin film of liquid, the seal faces make contact, heat up, and fail. In an MPP, the
pumped fluid usually contains a significant amount of gas, including periods of
100 percent gas slugging. Gas does not adequately cool and lubricate the seal faces.
Thus, for a single seal without special design considerations, seal failure is likely.
Seal failure will usually result in leakage of the pumped fluid (including its gas
phase) to the atmosphere.
In MPPs, a single seal can be used if it includes an internal close clearance throat
bushing inside the seal chamber and other special design techniques. The extent of
these other design techniques employed depends upon the percentage of GVF and
the amount of particulate in the pumped stream. This single seal design is simpler
and less expensive than a dual seal, and therefore, it is usually the type
recommended for most MPP applications.
With the use of the internal close clearance throat bushing inside the seal chamber,
seal flush flows through the close clearance into the MPP. The bushing’s clearance
is designed to keep the pressure in the seal chamber 20 psi or so above the pump’s
internal pressure. The purpose of the bushing is to prevent any gas from being
released to the atmosphere if the single seal fails. If it fails, only the liquid flush
would be released. This same flush can also act as screw sealant for a twin screw
pump.
The following are four API seal flush plans that can be used in conjunction with the
close clearance throat bushing:
• API Flush Plan 11: If the multiphase stream contains no particulate and is
below 200°F (93.3°C), API Seal Flush Plan 11 is acceptable. This plan takes a
liquid flush from the pump discharge and returns it across the seal faces. The
discharge line is enlarged as required to trap liquid that is recycled from the
bottom of the line, where the liquid has settled, back to the seals. Enough liquid
needs to be trapped to enable the appropriate seal flush liquid to be available
during normal GVF conditions and during design periods of 100 percent gas
slugs. (In the Bornemann twin screw pumps, the fluid is trapped inside the
pump case itself.) (Refer to Figure 600-25 in this section.)
Fig. 600-25 API Seal Flush Plan 11 (Courtesy of John Crane)
• API Flush Plan 21: If flashing is expected in API Seal Flush Plan 11, a cooler
will be required. API Seal Flush Plan 21 is the same as Seal Flush Plan 11 only
with the cooler added.
• API Flush Plan 31: If a small amount of particulate (less than 0.01 wt. %) is
present and the discharge fluid is below 200°F (93.3°C), API Seal Flush Plan
31 is often used. This plan also takes a liquid flush from the enlarged discharge
line through a centrifuge or cyclone separator back to the seals. The centrifuge
separates the particulate from the pumped liquid and rejects the particulate back
to the pump’s suction. Note that for the centrifuge to work properly there must
be a large difference between the gravities of the particulate and the fluid being
pumped. Carbon, for example, will not centrifuge out, but sand will. (Refer to
Figure 600-26 in this section.)
• API Flush Plan 32: If particulate is present in amounts above 0.01 wt. %, an
external liquid flush (API Seal Flush Plan 32) is recommended. This approach
involves a continuous, clean, 100 percent liquid stream from an external source
that enters the seal cavity to cool and lubricate. Water is often used for the
flush. (Refer to Figure 600-27 in this section.)
Dual Mechanical Seals

Dual seals are more complicated and more expensive than single seals, and
therefore, single seals are more common and preferred. However, except in
extremely specialized circumstances, dual seals are still considered necessary for
safety reasons when H2S or other hazardous components are present or if the
pumped stream will auto-ignite upon seal leakage.
Unpressurized Dual Mechanical Seals. An unpressurized dual seal consists of two
separate mechanical seals, with a buffer fluid circulated between the seals. It is
shown in API Standard 682, Seal Flush Plan 52. (Also refer to Figure 600-28 in this
section.)
The buffer fluid supplied is intended to be more environmentally benign and safer
for personnel than the pumped fluid. The buffer fluid, by definition, has a pressure
below that of the seal chamber and provides a flush fluid which lubricates and cools
the outer seal faces. The inner seal flush is provided by any of the seal plans covered
under the single seal section above. If the inner seal leaks, the pumped fluid (if it
does not flash as its pressure is reduced) will flow from the seal chamber into the
buffer fluid reservoir, raising its level, and setting off a high level alarm. If the
pumped fluid flashes as it enters the reservoir, the pressure in the buffer fluid
reservoir increases, and a high pressure alarm is initiated. If the outer seal leaks, the
more benign buffer fluid will leak to the environment, the reservoir level will drop,
a low level switch will alarm, and the pump will shut down. The pumped fluid will
not leak to the atmosphere unless both the inner and outer seals leak at the same
time.
The circulating buffer fluid system consists of a reservoir with a cooler inside, level
switches, a pressure gage, an orifice, and assorted valves. The buffer fluid circulates
from the reservoir into the chamber between the two seals and then back to the
reservoir.
Pressurized Dual Mechanical Seals. An example of a pressurized dual seal

consists of two separate mechanical seals, with a relatively benign barrier fluid
circulated between the seals to lubricate and cool both the inner and outer seal faces.
The barrier fluid pressure is kept above the seal chamber pressure, which is usually
about suction pressure. Thus, if the inner seal leaks, the barrier fluid flows into the
pumped fluid. If an outboard seal leaks, barrier fluid flows to the atmosphere. In
both cases, the barrier fluid reservoir level will drop, and instrumentation in it will
shut down the pump. The pumped fluid will not leak to the atmosphere unless both
seals leak simultaneously.
For an illustration in the form of a figure, refer in this section to API Seal Flush
Plan 53A (see Figure 600-29), Plan 53B (see Figure 600-30), Plan 53C (see
Figure 600-31), and Plan 54 (see Figure 600-32).
Fig. 600-29 API Seal Flush Plan 53A (Courtesy of John Crane)
Fig. 600-30 API Seal Flush Plan 53B (Courtesy of John Crane)
Fig. 600-31 API Seal Flush Plan 53C (Courtesy of John Crane)
The circulating barrier fluid system consists of a pumping ring, a reservoir with an
internal cooler, several level switches, a pressure gage, a pressure switch, and
assorted valves. The barrier fluid circulates from the reservoir into the seal chamber
between the two seals and then back to the reservoir.
The pressure switch keeps the barrier fluid pressure above the pump’s seal chamber
pressure, which is slightly above pump suction pressure. If the pressure switch
malfunctions, the barrier fluid pressure can become lower than the suction pressure.
This is called pressure reversal. Pressure reversal will likely overcome the closing
pressure of the springs and open the faces, creating a leak. If purchasing dual seals,
this situation should be considered, especially for startup conditions. The amount of
pressure reversal that the seal can handle before leaking should be recorded by the
supplier on the seal data sheets.
Suction pressure is highest immediately before and during the startup of an MPP.
After startup, the MPP lowers the suction pressure. The barrier fluid must be
designed to operate above the highest possible suction and seal chamber pressure
expected.
There are a number of other components in this barrier fluid system, not shown in
the API figure, including automatic shutdowns to prevent leakage to the
atmosphere.
Comparison of the Seal Types. As mentioned earlier, the single seal with the close
clearance bushing is the most common, because it is simpler and less expensive than
the dual seals. The next most common is the pressurized dual seal, used where H2S
or other toxic components are included in the MPP stream. The least common is the
unpressurized dual seal. The only known application that comes close to the API
unpressurized dual seal is the Bornemann “poor man’s” mechanical seal. The inner
seal is the same as one would expect in an unpressurized dual seal. The outboard
seal is a contacting lip seal, which, not being a mechanical seal, differentiates it
from an API unpressurized dual seal. The area between the seals is flushed with an
unpressurized buffer fluid at atmospheric pressure. (For more details, refer to
Mechanical Seals in Section 652.)
Materials for Mechanical Seals

The standard mechanical seal faces are silicon carbide for the seal ring (flexible
spring mounted seal face) and carbon for the mating ring. Silicon carbide is
generally selected, since it has high hardness, excellent corrosion resistance, high
thermal conductivity, and a low coefficient of friction against carbon. It comes in
two types, reaction bonded silicon carbide and self-sintered silicon carbide.
Reaction bonded silicon carbide is preferred, since it has a lower coefficient of
friction against carbon, is less brittle, and is not as hard. It is more suitable for fluids
that have a pH from 4 to 11. Silicon carbide has a maximum temperature of 800°F
(426.7°C).
Tungsten carbide is also used as a seal ring material instead of the silicon carbide to
run against the carbon mating ring. It has a maximum temperature of 750°F
(398.9°C).
Seals with silicon carbide or tungsten carbide running against carbon are standard
seal offerings and are usually satisfactory if the pumped fluid temperatures are
below 750°F (398.9°C) and little or no particulate is present.
Two hard faces, such as tungsten carbide versus silicon carbide, are required if sand
or other abrasives are contained in the pumped stream. Unless it is certain that no
particulate is present, tungsten carbide versus silicon carbide should be used with
API Seal Flush Plans 11, 31, 32, 52, 53A, 53B, 53C, and 54. This face to face
combination runs well, even in abrasive water streams.
Of course, the toughest application is a problem when the pump and its seal must
exist in a remote location pumping a hot, high GVF stream (90 percent and above)
containing particulate (quartz) and without access to a clean, cool, external seal
flush. In this case, the pumped fluid must be used as the seal flush. The most
experienced seal design to date would run hard face against hard face, usually
tungsten carbide vs. silicon carbide. Diamond seals (running diamond face against
diamond face) have been developed recently, and experience is being accumulated
which preliminarily indicates that these seals will likely be the seal face material of
choice for this application in the future. Today, however, there is little experience
with these seals, and they are not yet recommended. With or without the diamond
seal, the seal flush would come from a downstream separator vessel or from the
bottom of the enlarged discharge line acting as a separator. Again, this enlarged
discharge line segment needs to be designed so that it has enough residence time to
knock out the required amount of seal flush liquid to sustain the seals during normal
multiphase operating conditions and during design periods of 100 percent gas
slugging.
Another new seal developed by Chevron for this application is a single mechanical
seal, cooled and lubricated by grease fed from a canister. This type of seal was
tested for several years at the Chevron Humble facility and found to be reliable.
However, the seal is not patented and has not yet been deployed commercially.
Testing was only performed on suction pressures of approximately 6 psi.
Both the grease canister seal and the diamond seal are discussed in Section 6146,
specifically in Grease Canister Seal Flush System.
Refer to the Chevron Pump Manual—Engineering Guidelines (Gray Book),

Section 800, and API 682, and Chevron API 682 exception specification,
PMP-SC-4662, for an in depth discussion on all types of mechanical seals and the
various seal flush plans.
633 Screw Sealant for Twin Screw Pumps

In a twin screw pump, the screws are held apart by bearings with only 0.008 inch to
0.010 inch clearance between the screws and with the same clearance between the
screws and the pump case or liner. Liquid must fill these clearances. This liquid is
often called screw sealant. The design flowrate of the screw sealant must reliably
deliver 4 to 5 percent of the pump’s design capacity or the pump will vapor lock,
stop pumping, heat up, have internal rubbing between the rotors and stators, and
cause severe damage.
Screw sealant can be supplied by the liquid portion of the multiphase fluid entering
the pump if the GVF is always less than 95 percent (absolutely no slugs of gas).
Otherwise, a source of screw sealant must be supplied from a downstream separator
(using the pumped liquid) or from an external supply that is unrelated to the fluid
being pumped, e.g., water.
A flush to the mechanical seal can serve two purposes, acting as both the seal flush
and the screw sealant. It can come from an enlarged section of the discharge pipe or
a downstream in-line separator (e.g., a gas liquid cylindrical cyclone) or a large
downstream separator vessel (API Flush Plans 11 or 31) or an external supply
(API Flush Plan 32) to the single or dual seal. The flush then enters the seal
chamber and the pump where it becomes the screw sealant. API Seal Flush Plan 32
is used for services with a GVF greater than 95 percent.
If the fluid being pumped is below a 95 percent GVF, an external supply of screw
sealant is not theoretically required. However, caution is advised in these situations
because multiphase flow is almost never homogeneous. For instance, with a
90 percent GVF one might think that no screw sealant is required. However,
90 percent GVF is an average. Even at 90 percent GVF, slugs of pure liquid
(0 percent GVF) and pure gas (100 percent GVF) can occur. These slugs must be
designed for and communicated to the MPP supplier on the API 676 data sheets. As
with the mechanical seal flush, a clean, cool, particulate free, external liquid steam
should be supplied. For example, a screw sealant was needed in the Mitsue pump
with an average GVF of 75 percent, yet it experienced severe liquid and gas
slugging. (Refer to Section 6134 Lessons Learned, item C.)
Slugging is described in detail in Section 635.
For a dual mechanical seal, the screw sealant must be fed directly into the seal flush
port immediately inside the inner seal, at the end of each screw.
As explained in Section 633, if the screw sealant comes from a discharge separator,
the level in the separator will drop as the gas slug is moved by the pump. If the gas
slug lasts for a longer duration than designed for, the separator will run dry, starv-
ing the pump of screw sealant. With the external supply of screw sealant, a high
temperature sensor will pick up the increased heat inside the pump and shut down
the pump before damage occurs. If a downstream separator vessel is used, a level
alarm and shutdown will protect the pump. Refer to Figure 600-33 for a skidded
MPP with a separator built into the discharge line.
Fig. 600-33 Skidded MPP with a Built-In Separator in the Pump’s Enlarged Discharge Line (Courtesy of Leistritz
Corporation)
In conclusion, the downstream separator must be sufficiently sized, taking into

account the duration of a 100 percent gas slug. One method to determine the size
and duration of slugging is to model the suction piping system by using OLGA, a
computer simulation program.
634 Rotor Design for Twin Screw Pumps

A liquid-only twin screw pump is designed to pump liquid (no gas), and the rotors
are held apart by using the pumped liquid as a lubricating cushion and preventing
wear between the parts. For an MPP, the rotors must never touch, and the rotor shaft
must be stiffer and, therefore, larger in diameter. For this reason, a liquid-only twin
screw pump should never be used for multiphase service. The twin screw rotors
should be designed such that the rotors will deflect no more than half the designed
internal clearance under the most severe operating conditions.
A twin screw MPP will also operate hotter than one that pumps pure liquid.
Compressing the gas in the pumped fluid causes the pump to run hotter, expanding
the internal parts. Internal clearances must be increased. Otherwise, the parts will
rub and seize. To avoid this, the rotor clearances should be designed to allow the
pump to operate without the rotors contacting each other. To prevent wear, the pump
should be designed for a temperature of 300°F (148.89°C) above the maximum
expected suction temperature.
For example, the 300°F (148.89°C) recommendation is based in part on the Mitsue
pump that experienced a temperature rise of 280°F (137.8°C). (Refer to
Section 6134 Lessons Learned, items A and B.)
Unless specifically designed to pump 100 percent gas, MPPs are not designed to
pump pure gas efficiently. The screw flight spacing (screw locks) are designed for
100 percent liquid, and the increased clearances, required for gas, make the pump
relatively inefficient, typically 30 to 50 percent. Note, however, that in comparing
the twin screw pump to the helico-axial pump, PCP, or ESP, at any GVF, the twin
screw pump’s efficiency is considerably superior.
Bornemann offers a “regressive screw” twin screw pump design in which the screw
locks become smaller towards the discharge. This design is marketed as improving
the pump’s volumetric efficiency by as much as 15 percent. (Refer to Section 634.)
If particulate is present, the clearance may also need to be increased further to
compensate for the size of expected particulate. When developing quotes, a
particulate size distribution and an analysis of the particulate showing the percent
quartz, silica, clay, etc., should be provided on the API 676 data sheets given to the
MPP suppliers to determine the correct clearances. With properly designed
clearances, the screws can expand when handling pure gas, and particulate can pass
without rubbing or causing erosion. If only one (gas slugs or particulate) is present,
the clearances are sized for only that criteria.
635 Effect of Slugs—Gas and Liquid

Slugging can be defined as the alternating of large pockets of liquids and gas to the
pump. A twin screw type pump will not be damaged by severe slugging. This is
because twin screw MPPs, besides being of a very robust design, are designed to
split the incoming flow into two equal parts, each part entering the screws from the
opposite ends of the pump at exactly the same time. The resulting forces are
opposed and cancel each other. This design is one of the primary reasons that the
twin screw pump is used for most applications. Other types of MPPs do not have
this feature and are, therefore, more susceptible to damage from slugging.
Slugging can be very severe if the MPP is located a good distance from the well(s).
Before starting the pump, the low points of the suction piping are filled with liquid.
During startup, the liquid pockets are flushed into the pump. After this is done,
liquid again starts to fill the low points, allowing gas to flow in the upper radius of
the pipe, feeding the MPP mostly pure gas. Startup slugging is referred to as terrain
slugging, and the effects can be severe. If a liquid slug enters the pump after a gas
slug, the suction pressure drops suddenly, and the liquid level in the downstream
separator rises significantly. The cooler liquid causes the temperature in the pump to
drop significantly, as well.
The reverse happens if gas slugs occur after liquid slugs. The pump’s suction
pressure rises suddenly, and the discharge separator level drops, while the pump
temperature rises significantly.
After the pump has been operating, terrain slugs are often followed by
hydrodynamic slugs. Hydrodynamic slugs are caused by the well’s production
stream, its GVF, and the terrain-caused ups and downs of the pump’s suction line.
As described above, the bottom of the suction line begins to fill with liquid,
especially at the low points, while the gas flows over the liquid in the upper portion
of the line. Liquid slowly fills the line sufficiently such that, eventually, the gas
pushes it through and into the MPP. Hydrodynamic slugs are usually more frequent
but less severe than terrain slugs.
Liquid slugging also affects the design size of the downstream facilities. During a
liquid slug, the MPP pumps at its theoretical liquid flowrate (not at the liquid
percentage of the multiphase flowrate). The downstream separator, in combination
with the liquid flow line capacity, should be sized for this 100 percent liquid rate for
the expected duration of the liquid slug. A Pipephase or OLGA computer simulation
is required to determine the slug duration and to assist in the proper sizing of the
separator and the discharge flow lines.
If practical, one way to avoid slugging or at least minimize its severity is to locate
the MPP close to the well(s) such that the number of terrain ups and downs in the
suction line is kept to a minimum. For example, the Chevron Mitsue pump
experienced both types of the slugging discussed above, but other companies have
had similar experiences. (Refer to Section 6134 Lessons Learned, item D.)
636 Motor Selection and Use of Adjustable Speed Drives (ASD)

One of the characteristics of a twin screw pump and the progressive cavity pump is
that they are positive displacement pumps (capacity is proportional to speed) and
constant torque machines (under the same fluid conditions, the required torque is
dependent only on differential pressure). They operate independently of speed and
gas volume fraction. If the pump is started against the full system backpressure, the
motor must develop a large torque. Most motors do not develop full torque until
they reach full speed. Therefore, if the pump starts under a load, a larger HP motor
will be necessary than will be required for normal operation. A larger HP motor
increases the cost of the pump/motor skid substantially.
There are two commonly used methods to avoid this:
1. Installation of a startup recirculation line (piping from pump discharge or a
downstream separator back to a suction header or suction tank) that includes a
pressure control valve;
2. Installation of an ASD on the motor, a motor specifically designed for an ASD.
Solution 1 allows the MPP to start at a low discharge pressure, where the MPP
requires a low torque from the motor. The pressure control valve in the recycle line
is opened on pump startup allowing the motor to reach full speed and full torque
before it closes on a planned basis, thereby building the discharge pressure on pump
slowly to keep the required torque below that developed by the motor. To do this
properly, the operator needs to be certain that the suction, discharge, and
recirculation lines are full of liquid before the pump is started. This approach could
be used after startup during normal operation to reduce the net production by
recycling flow through the bypass line. However, this approach wastes energy. Also,
if the bypass line is left open too long, the fluid will heat up, possibly flash, and
increase the GVF.
Solution 2 provides the advantage that the motor provides the required flow control
under operating conditions. An ASD allows the operator to vary the speed and,
therefore, the flowrate of the pump, to better match the field’s production. Finally,
by increasing the pump speed, an ASD system can be used to maintain production
as the pump wears.
Solutions 1 and 2 can be applied independently. However, both should be
implemented.
Other drives, such as natural gas or diesel engine drives, can also be used. These
drives are not constant torque drives and must be sized such that the pump’s
required torque never exceeds the torque available from the drive. Again, the drive
may be oversized to accomplish this, or a recycle line may be used.
For the helico-axial pumps, which are not constant torque, standard electric motors
and typical discharge flow control valves are used, often with a recycle line to keep
the pump above the pump’s minimum flowrate (thermal or stable), or an ASD drive
and its properly designed motor is used. The flow control valve and recycle line will
likely waste energy, so the ASD option is the preferred solution.
637 Methods to Reduce Sand Erosion

Twin Screw Pumps
Based on the Duri test, sand rates equivalent to 90 bbl/day will not cause erosion in
a twin screw pump as long as the sand is flowing in crude oil at or below 22 degrees
API with a viscosity above 330 cp. In this test, it is believed that no erosion
occurred because the sand stayed in suspension in the heavy, viscous crude oil. A
computer simulation also indicated that a pump handling crude oil between
22 degrees API and 30 degrees API is not likely to erode with sand present. The
amount of erosion from sand is expected to increase if the API gravity is above
30 degrees. If the sand is pumped with only gas, steam, or water, the pump will
definitely erode quickly. This was shown at Duri during the steam injection test.
(Refer to Twin Screw MPP Results in Section 6138.)
Erosion with water and particulate was also demonstrated at the University of
Erlangen in Nuremberg, Germany. During the 1990s, a sand erosion test was run on
a conventional twin screw pump. During the test, a stream of pure water containing
a 0.9 vol. % of sand was pumped while the pump’s differential pressure was
maintained at approximately 220 psi. The sand was pure quartz with a particle
distribution that extended above and below the pump’s internal clearances. In
1-1/2 hours, the pump wore out completely.
If a twin screw pump is installed with sand in its feed, several design improvements
should be implemented to directionally reduce the effect:
1. Installing a removable liner with Satellite or tungsten carbide coating.
Suggestions for coatings are described in Section 641;
2. Gas hardening the rotors (Again, refer to Section 641);
3. Coating the screw edge;
4. Operating the pump at a slower speed (balancing the cost of the larger sized,
slower pump against the smaller sized, higher speed pump that will wear more
quickly);
5. Using an ASD with a properly designed motor or a variable speed engine,
which allows the pump to run at speeds above its normal operating speed;
6. Using the maximum number of screw turns or locks, more than that necessary
to develop the required discharge pressure;
7. Prohibiting balance holes on the periphery of the screws. (They create a place
for rapid erosion.)
By using hard coated liners and hardened items 1, 2, and 3, the amount of erosion is
reduced as long as the particulate in the stream is softer than the coatings. Allowing
the pump to operate at slower speeds (item 4) reduces the erosion, since the erosion
rate, “e”, equals the velocity, “v”, raised to the 3rd or 4th power (e = v3). If the pump
is designed to provide the required flowrate at a speed less than the motor’s
maximum rpm and an ASD is used to drive the motor, the speed of the pump and
motor can be increased to compensate for any erosion suffered during normal
operation (item 5). Confirm with the manufacturer that the speed increase is
acceptable. Using item 6 decreases the pressure boost across each stage, decreasing
slip, therefore decreasing erosion. Implementing item 7 eliminates a location where
erosion can get a foothold, destroying the coating around it. (It is common for
manufacturers to balance their rotors by drilling holes in the tip of various screws.)
The most commonly used design modifications to prevent erosion are items 1, 2, 3,
4, and 5. If sand is present, an MPP specialist should be consulted before purchasing
the pump.
In a twin screw pump, rotor clearances are small, on the order of 0.008 inch, almost
always smaller than a grain of sand. Sand tends to wedge between the screw edges
and the bore of the pump, causing erosion. Once erosion begins, the slip or the flow
of fluid backwards across the screw edges will increase causing additional erosion
which continues an ever more rapid rate of erosion.
Helico-Axial, PCP, and ESP Pumps

Since a certain speed is required to develop the required pressure differential and
flowrate, these types of pumps typically rely on proprietary coatings and materials
to minimize erosion. (Refer to Section 660 for Helico-axial pumps, Section 670 for
PCPs, and Section 680 for ESPs.)
638 Recycle Stream Flashing (Surface Pumps)

Twin Screw Pumps
Liquid near its boiling point will flash into gas if a pressure drop occurs or if its
temperature is raised, such as, if a liquid near its boiling point is mixed with a hotter
fluid. The result is that a small volume of liquid will flash into a very large volume
of gas.
As mentioned before, if the GVF of a pumped stream is very high, a recycle screw
sealant stream will be required. If the screw sealant stream is the pumped fluid
recycled from a discharge separator, it is likely to be near its flash point, and
flashing needs to be evaluated.
Usually a flow control valve is installed on the screw sealant line from the
downstream separator to the suction side of the pump. As the fluid passes through
the control valve, the fluid pressure drops to that of the suction. As this pressure
reduction occurs, the hot screw sealant fluid may flash. Also, if hot enough, the
sealant fluid may cause some of the incoming pumped stream to vaporize. Of
course, for a given sealant fluid temperature, more of the incoming stream’s volume
will flash if it is a lighter gravity fluid rather than a heavier gravity fluid. The net
effect of this flashing is that the evolved gas occupies pump capacity. This gas
volume replaces liquid volume in the pump, decreasing the pump’s volumetric
efficiency and causing the pump to be built larger to accommodate the increased gas
volume.
Once a flow diagram is identified, a process simulation should be done to identify
areas where flashing can occur and to quantify the amount of flashing. HYSIM or
PRO II are two simulation programs that can be used for this purpose.
Potential solutions to minimize the effects of flashing (other than oversizing the
pump) are shown in the following list (the recommended solutions are items 3 or 4):
1. Mixing the recycled stream into the suction stream as far upstream of the
pump’s suction flange as is possible. This allows for some atmospheric cooling
of the recycled stream and for re-condensing some of the gas before it reaches
the pump.
2. Mixing the recycled stream into a suction vessel.
3. Installing a cooler or condenser in the recycle stream after the pressure
reduction valve, injecting the recycle stream into the suction or into the
mechanical seal flush port. The HYSIM process simulator should be run again
to size the cooler. Usually the cooler is a small air cooled exchanger.
4. Supplying the screw sealant from a cool, external source.
For example, Mitsue used these solutions. (Refer to Section 6134 Lessons Learned,
item J.)
Recycle, Slip, Flashing Factor. The pump’s size needs to be larger than the size
that is first calculated. This is necessary to compensate for flashing and the
increased clearances (with increased slip) needed to compensate for heat and
particulate. To do this, one simply multiplies the initial design capacity by a recycle,
slip, flashing factor in accordance with the following:
• Use a factor of 1.10 percent for screw sealant that is the same temperature as
pump suction (solution 3 or solution 4);
• For uncooled recycled streams:

– Use a factor of 1.15 percent for all GVFs and a heavy crude (API less than
25 degrees);
– Use a factor of 1.25 percent for GVFs less than 50 percent and a light
crude (API greater than 25 degrees);
– Use a factor of 1.35 percent for GVFs greater than 50 percent and a light
crude (API greater than 25 degrees).
Refer to Section 693 for an example of how this is applied. The need for the recycle,
slip, flashing factor was determined from the Mitsue pump. (Refer to Section 6134
Lessons Learned, item J.)
Helico-Axial, PCPs, and ESP

The helico-axial pump and surface PCPs or ESPs may also require a recycle stream
for low flowrates. If so, flashing must be considered. Solutions 1, 2, and 3 in
Section 638 (twin screw pumps) can be applied.
639 Suction and Discharge Piping (Surface Pumps)

Model the Piping
During the early design stages, the multiphase simulator, Pipephase, should be run
on the MPP’s inlet pipe from the production wells through the discharge piping to
the downstream separation. If Pipephase indicates slugging, a transient simulator,
such as OLGA, should be used to predict the size, frequency, and duration of the
slugs. This information is necessary to properly size the pump and associated
equipment. For example, the OLGA simulation of the Mitsue installation indicated
significant slugging. (Refer to Section 6134 Lessons Learned, item G.)
Suction Strainer
A strainer should always be installed in the MPP’s suction line to catch debris. It
should be a permanent, basket strainer, not just a startup “witch’s hat”, conical
screen. If sized properly, this basket strainer will prevent small pieces of debris from
entering and damaging the pump. The openings or mesh size of the strainer is often
1/8 inch, unless wax is present. If wax is present, a filter is used instead of a strainer.
The sizing of the strainer mesh or filter should be determined for each application,
and the mesh size should be jointly developed by the purchaser and the pump
supplier.
Note The combination of wax and sand in a stream is particularly challenging
since sand will plug the filter quickly. Consult an SME on multiphase pumping.
The strainer or filter should be a duplex stainless steel unit that allows switching
from one unit to the other while the MPP runs. The open area of the screen should
be 150 percent of the pipe flow area, unless wax is present. For wax, 200 percent is
recommended. The strainer should be designed to withstand as large a differential
pressure as possible. This depends upon the size and pressure rating of the strainer
housing and the mesh itself. When purchasing the strainer, determine the strainer’s
collapse pressure. Finally, a high differential pressure across the strainer should
trigger an alarm, followed by a shutdown that is lower than the collapse pressure.
Refer to the Princess MPP and the Main Pass 313 installations. (Refer to
Section 6135 Lessons Learned, item B and Section 6136 Lessons Learned, item F.)
Inlet Pipe Orientation

The pump should be located as close as possible to the source of the pumped fluid.
The suction piping shall be kept as short as possible, with as few elbows as possible.
High spots where gas can get trapped or low spots where liquid can collect should
be minimized, since they will create slugging. Any of these vertical legs that are
absolutely necessary should be located as far from the pump as possible.
Inlet piping should enter a twin screw pump from a plane perpendicular (either
horizontal or vertical depending on the pump’s suction flange) to the axis of the
screws. Upstream of the pump inlet flange the pipe should be a straight run for at
least 10 pipe diameters or 10 feet, whichever is longer. An elbow should not be
installed immediately at the pump’s inlet flange.
Pipe that runs parallel to the axis of the twin screw pump’s rotors and that have an
elbow bolted directly to the pump flange must be avoided. This will channel more
flow to one side of the pump than the other. In a slugging situation, this will
increase the pump’s axial thrust forces beyond design and cause a failure.
Pressure Restriction Location

A pressure restriction (a control valve or orifice, etc.) shall not be installed near the
pump’s discharge flange. If such a device is installed near the pump, the pressure
will surge to unacceptable levels, instantaneously, when a slug of liquid follows gas.
A pressure restriction, if necessary, should be installed on the downstream side of
the liquid separator in the discharge line. If a liquid separator is not being used, the
restriction should be installed as far down the discharge line as possible. Locating
restrictions at the proper points will tend to dampen any pressure surge due to
slugging. For example, in the Mitsue installation, slugs caused a significant pressure
spike due to a pressure restriction near the discharge flange. (Refer to Section 6134
Lessons Learned, item E.)
640 MPP Materials
641 General
This section describes materials for MPPs that are designed to operate with streams
containing sand or other hard particulate.
Materials, coatings, and material hardening processes are improving rapidly, and
selecting the most erosion resistant material, coatings, or hardening process can be
difficult. An ETC or local expert should be consulted if sand or hard particulate is
involved.
642 Twin Screw Pumps

If twin screw MPPs handle sand, the rotors and case liner or case (if supplied
without a liner) must be fabricated using hard materials, material hardened by
various methods, or materials coated with overlays suited to withstand abrasion and
erosion. For twin screw pumps, the following methods have been used:
Pump Case
The pump case or the pump casing under the liner should be carbon steel or better.
Cast or malleable iron is not acceptable. A pump case liner is strongly
recommended.
Pump Case Liner

Alternative 1—Stellite 12 Weld Overlay. Stellite can be applied to the case or case
liner using the following general weld procedure:
• Preheat the base material to 482°F (250°C) with a maximum interpass tempera-
ture of 350°F to 400°F (176.7°C to 204.4°C);
• Apply a weld overlay, called a buffer or butter layer, consisting of 316 L SS
weld rod. The finished thickness of this layer, after machining, should be
0.080 in. (2.0 mm);
• Apply two layers of Stellite 12, with a final total finished thickness after
machining of 0.12 in. (3.0 mm);
• Post heat treat at 1,094°F (590°C), plus or minus 50°F (10°C). This tempera-
ture should be held for 1 hour. Heating and cooling rates shall be 104°F (40°C)
per hour.
Stellite 12 weld overlay has a hardness of 45 HRC.
Alternative 2—Titanium. Manufacture the entire liner out of titanium.
Alternative 3—Chrome Overlay the Liner. Applying chrome overlay to the pump
case or the pump case liner is not recommended. The process of adding chrome over
base metal is often performed poorly, and as a result, the chrome peels off the base
material. Chrome overlay has a hardness of 68 HRC.
Recommendations. Alternative 3 is not acceptable. Alternative 1 and Alternative 2
are recommended.
For all of the hardening or coating processes, the supplier should submit with his bid
detailed procedures, including weld procedures, and pre- and post-weld heat
treatments.
Rotors (Screws)
Gas Hardening. Gas hardening is usually used to harden twin screw rotors, since
their profile is difficult to apply a welded material to and then remachine to the
proper profile.
There are two commonly used gas hardening techniques:

Alternative 1—Nitride Soak. The entire twin screw rotors can be hardened with
nitride gas for 48 hours. The depth of the hardness is approximately 0.003 in.
(0.076 mm). Nitriding has a hardness of 58 HRC.
Alternative 2—Boride Gas Diffusion Deposition Coating. The twin screws can be
hardened with boride gas diffusion over the entire surface of the screws (edge, both
flanks, and the root).
• The parent material should be 400 series SS, capable of resisting hot carbonic
acid corrosion.
• Before applying the boride diffusion coating, the parent material should be heat
treated to 1,650°F (898.9°C), minimum, and slow cooled.
• A boride diffusion coating should be applied with a finished hardness of
1,600 DPH or 1,600 HV, minimum.
• Finished thickness of the coating should be 0.004 in. (0.102 mm), minimum.
• After coating, no check cracks should be visible by the naked eye.
Suggested suppliers are the pump manufacturers, since the coating is purchased
when the pump is bought. Boriding has a hardness of 2,000 HV (V = Vickers).
If particulate is present in the pumped stream, using Alternative 2 for the screws
lasts longer. However, it is more expensive and difficult to apply. The supplier’s
procedure needs to be reviewed and the results inspected.
Alternative 3—Stelliting the Screw Edges. In addition to Alternative 1 or
Alternative 2, Alternative 3 may be used. In Alternative 3, only the tips or edges of
the screws are stellited. Alternative should not be used by itself, but only in
combination with Alternative 1 or Alternative 2.
Alternative 4—Other Options. The pump supplier should be encouraged to offer
hardening alternatives, including price differences that are harder than those
mentioned above and/or that could coat more of the screws than just the edges or
tips. The pump supplier should submit such details with his bid.
643 Helico-Axial, PCP, and ESP Materials

These types of pumps use proprietary materials. Refer to Section 660 for Helico-
axial pump materials, Section 670 for PCP materials, and Section 680 for ESP
materials.
644 Mechanical Seal Materials (all types of MPPs)

For most helico-axial and twin screw pump applications running below 300°F
(148.9°C) or with no particulate or hydrogen sulfide or highly acidic or basic fluids,
the typical materials for the seal parts are:
• Seal ring: silicon carbide
• Mating ring: carbon
• Hardware (screws, springs, retaining rings, etc.): stainless steel

• Secondary seal elastomeric “O-Ring”: Viton
• Close clearance bushing: Bronze
For other upstream applications, refer to Materials for Mechanical Seals in
Section 632 of this manual, and, if necessary for other applications, consult Chevron
Pump Manual Seal Selection Guides PMP-DG-4662 CRN and PMP-DG-4662 FS.
650 Manufacturers—Twin Screw MPPs

651 General
The three most common manufacturers of twin screw MPPs are:
• Bornemann
• Leistritz
• Flowserve
The most common twin screw MPPs are Bornemann and Leistritz. They supply
approximately 90 percent of these pumps that are sold. They are well constructed
and highly reliable. Both manufacturers supply packaged units on skids.
As of this writing, Flowserve has few installations, and they have had problems with
their construction and reliability. We strongly recommend against purchasing their
pumps at this time. For example, Chevron has had problems with Flowserve pumps
in Venezuela’s Boscan and El Tigre fields, as well as in Chad. (Refer to
Section 6139.)
All three pumps use identical flow patterns. The pump’s inlet or suction flange is at
axial center of the pump case. The inlet fluid is split in two streams, with equal
portions directed to each end of the pump where the fluid enters the screws. The
screws move the fluid to the center of the screws where the discharge is located. All
three manufacturers incorporate four sets of mechanical seals, and all have bearings
outboard of the seals.
Note Bornemann also has a design in which the discharge is at the end of the
screws by the mechanical seals. This design should be avoided, if possible, since in
this design, the mechanical seals are required to seal against discharge pressure.
These seals are more difficult to design and their reliability is inferior.
In summary, at the moment, it is strongly recommended that only Bornemann or
Leistritz be considered for Chevron applications. Their pumps are well constructed,
highly reliable, and can be supplied as packaged units on skids.
Note Colfax owns Warren, Houttuin, Allweiler, and IMO brands. Warren has
manufactured a liquid screw pump for many years. We have had success with the
Warren pumps, which are pumping 40 percent GVF multiphase flow for
approximately 15 years in Venezuela. While not yet recommended for purchase by
Chevron, Colfax is currently attempting to enter the MPP market in a substantial
way.
Refer to Section 6130 for a discussion of several Chevron and legacy Texaco MPP
installations, including several “lessons learned” from each. Several
recommendations included in this manual are derived from these “lessons learned”.
A complete list of Chevron and legacy Texaco MPP installations is included in
Figure 600-1.
652 Bornemann MPPs

Applications
Bornemann has made more twin screw MPPs than any other supplier, with over 400
installations worldwide. Bornemann claims its largest pump, MPC series, can pump
900,000 bpd at a differential pressure of 1,015 psi. It also makes a smaller version,
the MW series. Figure 600-34 lists 15 significant installations.
As shown in Figure 600-34, Bornemann pumps are installed onshore, on offshore
platforms, and, recently, subsea. The pumps are installed in a wide range of
applications, such as increasing production by reducing wellhead pressure,
transporting multiphase fluids several miles through one pipeline, and taking suction
from the annulus of several oil wells where the multiphase fluid averages over
98 percent gas. Note that, in Figure 600-34, one of the Bornemann MPPs was
purchased by the Chevron Midway Sunset field in Bakersfield, California. It
performed so well that eight of the pumps are now in operation at that site.
Fig. 600-34 Significant Bornemann MPP Installations (Information Courtesy of Bornemann Corporation) (1 of 2)
Onshore
Location or Offshore Capacity Delta P GVF
Oil Field Customer Subsea Date No. Model (bpd) (psi) (%) Driver HP
Tunisia Poseidon Offshore 1989 1 MPC355-45F 60,000 928 96 Electric -
Duri, Caltex Onshore 1997 1 MW9.5zk-67 125,000 231 75 Electric 350
Indonesia
Bakersfield, Chevron Onshore 2000 1 MW7.3zk-33 7,400 290 60 Electric 63
CA
Venezuela Sincor Onshore 2000 1 MW8.5zk-67 74,000 510 82 Electric 774
Minas, Caltex Onshore 2000 1 MW8.7zk-46 60,000 110 65 Electric 161
Indonesia
CA
CA
Russia Tatoilgaz Onshore 2002 2 MW7.3xk-43 17,000 218 75 Electric 80
CA
CA
Canada CNRI Onshore 2003 1 MW9.5xk-90 159,000 171 95 Electric 646
US Gulf of BP Subsea 2006 3 SMPC 335- - 725 70 Electric 1,475
Mexico, King 50
Field
Fig. 600-34 Significant Bornemann MPP Installations (Information Courtesy of Bornemann Corporation) (2 of 2)
Onshore
Location or Offshore Capacity Delta P GVF
Oil Field Customer Subsea Date No. Model (bpd) (psi) (%) Driver HP
India Cairn Onshore 2006 2 MW106-132 364,000 180 95 Electric 1,530
Energy
Algeria AGIP Onshore 2007 2 MPC-400-79 136,000 798 90 Electric -
Canada Imperial Oil Onshore 2008 2 MW8.5zk-85 86,000 189 95 Electric 370
Recently, three Bornemann MPPs were installed subsea for the British Petroleum
King Field in the Gulf of Mexico. BP claims the pumps will increase production by
20 percent and overall recovery by 7 percent, extending the life of the field. The
pumps are on the ocean floor under 5,500 feet of water. Each of the pumps is driven
by a subsea electric motor controlled by an ASD on a host platform. The motor
supplies 6,600 volts to the subsea pump, 15 miles away. These are the first twin
screw MPPs to be located subsea on the seabed floor.
Orientation and Drivers

Bornemann pumps are designed to be installed horizontally. Usually, they are driven
by electric motor drives with ASDs, although they may be driven by natural gas
engines or other drivers. This enables the pump’s capacity to be controlled to match
well production. Flow is usually controlled based on a set suction pressure or
wellhead manifold pressure.
Packaging
For new installations, Bornemann pumps can be purchased directly from
Bornemann, or they can be packaged by Aker Kvaerner as skid mounted units or
units ready for grouting. The packages include strainers, valves, piping, pressure
and temperature gages, and discharge separators.
Testing Facilities
Bornemann does an inhouse test of its equipment with water and air. After the MPP
pumps water for a while, Bornemann injects air into the water until the temperature
rises to a pre-established level. Bornemann does not test with particulate. This
testing is not considered to be a good indicator of how a pump will react while
pumping an actual multiphase fluid.
Particulate Handling and Speed

These pumps are usually run at 1,800 rpm but may be run at 3,600 rpm or 1,200 rpm,
based on the amount of particulate entrained in the stream, expected well production
rates, and the economics of the project.
The Bornemann pump demonstrated during the Duri test that it could pump and not
retain a substantial quantity of sand without internal damage. (Refer to Section 6138
for details.)
Materials and Coatings

Bornemann pumps are usually fabricated from carbon steel. A fully replaceable
liner is offered. The liner can be Stellite weld overlaid, and the screws can be
borided to provide better wear resistance to sand. The other coating materials
mentioned in Section 640 can also be provided. For example, these features were
included in the Caltex Duri twin screw MPP, manufactured in 1997. (Refer to listing
in Figure 600-34 and Section 6138.)
Mechanical Seals
Bornemann will supply its MPPs with a single seal using the close clearance throat
bushing or a pressurized dual seal or an unpressurized dual seal. The Bornemann
unpressurized dual seal is unique, not supplied by any other twin screw MPP
manufacturer. It is called the “poor man’s” seal. It consists of an inner seal, identical
to an unpressurized dual seal with a lip seal on the outboard end, which
differentiates it from the unpressurized dual seal described in API 682. The inner
seal is cooled and lubricated from a flush that is routed to the seal from two
connections in the case. Like all unpressurized dual seals, the “poor man’s” seal
injects the area between the inner or primary mechanical seal and the outer
secondary lip seal with a buffer fluid at atmospheric pressure supplied from two
tanks built into the MPP case. Bornemann claims it has supplied this seal to
95 percent of its customers, amounting to over 450 units. Bornemann also claims
that the seal lasts approximately 2 years, with the inner seal, as well as the lip seal,
failing at just about the same time. Chevron has a few of these seals in the
Bakersfield area.
Distinctive Features
Bornemann MPPs include several features not normally found in any other MPPs,
such as:
• A large pump case with an internal chamber, designed to separate and retain a
required portion of the liquid from the feed stream;
• The poor man’s seal described above;
• An adjustable internal circulation valve that circulates the trapped liquid back
to the pump suction. If fully open, the pump’s entire capacity can be circulated.
Bornemann claims that this provision unloads the pump and motor during
startup. The valve’s plug contains a groove, such that, if the valve is fully
closed, it still circulates 3 to 4 percent of the pump’s capacity back to suction.
This is usually enough to supply the screw sealant needed if the pump is fed a
fluid with a GVF above 95 percent.
Note that these features are not always considered positive. There is no
instrumentation to show if and when the internal chamber runs dry. The potential for
this occurring is high if pumping streams exceed a GVF of 98 percent.
Bornemann also includes a high temperature shutdown in the case that shuts the
pump down if the screw sealant liquid in the case is depleted. If liquid is depleted,
the pump heats up and shuts down due to high temperature. If a stream with a GVF
greater than 98 percent is anticipated, an external source of liquid screw
sealant/flush should be used. (Refer to Section 638.)
653 Leistritz MPPs

Applications
Leistritz model L4HK, its largest twin screw pump, can pump 330,000 bpd at
1,200 psi differential pressure. Leistritz is working to increase its differential
pressure capability to 2,100 psi. Leistritz makes another, smaller model, L4MK. Its
maximum flowrate is 220,000 bpd at a maximum pressure boost of 300 psi. Both
models can handle streams with GVFs of up to 100 percent. At GVFs above
90 percent, Leistritz requires a separate system for screw sealant and/or the
mechanical seal flush, from a wide section of the discharge line or from a discharge
vessel with a level gage, piping, and valves. The level gage allows monitoring of the
screw sealant/flush flowrate and assurance that it is flowing all the time. Above
98 percent, a clean, cool, reliable external sourced screw sealant/flush is
recommended. (Refer to Section 638.)
The first MPPs purchased by Chevron and legacy-Texaco (Trinidad, Humble,
Mitsue, and Main Pass 313) were all manufactured by Leistritz. They have proven
to be reliable, and Leistritz has supplied over 70 MPPs. Figure 600-35 lists 15 of
Leistritz’s more significant installations.
Leistritz’s current installations are onshore and on offshore platforms. One of them,
shown in Figure 600-35, was installed by Chevron in 2007 on an unmanned
offshore platform, Main Pass 59A in the Gulf of Mexico, driven by a large
(1,700 HP) natural gas engine. Leistritz calls this installation the “largest of its
kind”. The pump is designed to increase production by increasing the wellhead
pressure of 18 wells to match that of the additional adjacent wells such that a
common production line can be used. The GVF was 95 percent. This is a highly
profitable installation, making $80,000/day based on oil at $80/bbl, resulting in a
payout period of only 0.17 years. (Refer to Section 6131.)
Fig. 600-35 Significant Leistritz MPP Installations (Information Courtesy of Leistritz Corporation) (1 of 2)
Onshore
Location or Oil Offshore Capacity Delta P GVF
Field Customer Subsea Date No. Model (bpd) (psi) (%) Driver HP
Humble, TX Texaco Onshore 1997 1 L4NG 21,000 130 90 Electric 65
Mitsue/Princess Chevron Onshore 1995 1 L4HK 40,000 400 75 Electric 700
Main Pass 313 Chevron Offshore 1995 1 L4NK 26,000 125 98 Electric 75
Moreia Field Petrobras Offshore 1998 1 L4HK 61,000 580 88 Electric 800
Peace River Shell Onshore 2007 4 L4MK 87,100 300 95 Electric 600
Canada
Omen Shell Onshore 1998 1 L4HK 61,000 940 86 Electric 710
Canada
Chad Esso Onshore 2003 5 L4HK 150,800 840 56 Electric 960
Neiva, Colombia Colombia Onshore 2002 1 L4MK 45,000 265 58 Electric 300
Neiva, Colombia Colombia Onshore 2003 1 L4MK 56,000 430 58 Electric 700
Matarie, LA Cent. Expl. Offshore 2007 1 L4HK 132,100 250 95 Electric 820
Fig. 600-35 Significant Leistritz MPP Installations (Information Courtesy of Leistritz Corporation) (2 of 2)
Onshore
Main Pass 59A Chevron Offshore 2007 1 L4HK 152,100 640 95 Gas 1,700
Engine
Alberta, Canada Devon Onshore 2003 1 L4MK 22,000 231 95 Electric 125
Bakersfield, CA Area Onshore 2005 1 L4MK 130,000 190 35 Electric 800
Energy
China Bohai Oil Offshore 1996 1 L4HK 37,000 525 92 Electric 650
Corp
Brazil Petrobras Subsea Soon 1 ? 75,000 870 90 Electric -
It is expected that shortly a Leistritz pump will be installed on the ocean floor off
the coast of Brazil. That pump, SBMS-500, will pump from a water depth of
2,100 feet and produce a pressure boost of 870 psi. This pump will experience an
average GVF of 87 percent but is designed for 100 percent with an external supply
of screw sealant. Slugs are expected because the pump is located 1-1/2 km from the
well. Lubricating oil is supplied to the pump and motor from a tank located on the
host platform. This system is unique and patented.

Leistritz pumps are installed horizontally. Usually they are driven by an electric
motor controlled by an ASD, although they may be driven by natural gas engines or
other drivers. This enables the flow to be controlled to match production
requirements. Flow is usually controlled from suction pressure or wellhead manifold
pressure.
Packaging
For new installations, Leistritz pumps are usually packaged as skid mounted units or
units ready for grouting. Units include strainers, valves, piping, pressure and
temperature gages, and discharge separators. Leistritz surface pumps are commonly
packaged by Fluid Power Systems in Houston, Texas, while its subsea pumps are
packaged by Cameron. Of course, replacement pumps or new units can be
purchased as individual items.
Testing Facilities
Fluid Power Systems will test the Leistritz pumps using water and natural gas.

These pumps are usually run at 1,800 rpm but may be run at 3,600 rpm or
1,200 rpm, based on the amount of particulate entrained in the stream and the
economics of the project. For example, the Princess pump ran reliably for several
years at 3,600 rpm. (Refer to Section 6134.)

Leistritz pumps are usually fabricated from carbon steel. A fully replaceable liner is
offered. The liner can be Stellite weld overlaid, and the screws can be nitrided to
provide better wear resistance to sand. The other coating materials mentioned in
Section 640 can be provided.
Mechanical Seals
Leistritz has sold MPPs with the single seal using a close clearance throat bushing,
as well as pressurized dual seals. To date, Leistritz has not sold any MPPs with the
unpressurized dual seal. If requested, Leistritz will use cartridge seals designed in
accordance with API 682.
Leistritz pumps are more basic than Bornemann pumps, in that the liquid separation
chamber and internal recycle valve are not supplied. On the other hand, Leistritz is
easy to work with and will modify its equipment, backed by sound technical
reasoning. For example, the Mitsue pump was the first MPP in the industry to use
the API flush plan #32 with a throat bushing, where the flush also acted as a screw
sealant. This was accomplished after discussions between Leistritz and Chevron.
Like Bornemann, the Leistritz pump includes a high temperature shutdown if the
screw sealant/flush is lost. If fluid is depleted, the pump heats up and shuts down
due to high temperature. If a stream with a GVF greater than 98 percent is
anticipated, an external source of liquid screw sealant/flush should be used. (Refer
to Section 638.)
Leistritz also makes a skid mounted MPP that takes suction from a well’s casing if
rod pumps are used. It can also take suction from several well casings. The MPP
lowers the casing gas pressure that, in turn, increases the downhole liquid level in
the annulus. All of this allows the operator to speed up the rod pump to increase
crude oil production. The Leistritz design involves adding a small stream of liquid
from the discharge of the rod pump to the MPP’s suction to act as the seal flush and
screw sealant.
654 Flowserve MPPs

Note Currently not recommended due to numerous failures in Chevron and non-
Chevron applications.
Applications
Flowserve (formerly Ingersoll Rand) claims to have supplied 48 MPPs. Its largest,
MP1, pumps up to 280,000 bpd at a differential pressure of 1,000 psi. A list of
Flowserve installations is currently not available.
Flowserve’s largest pump is installed in Venezuela at Sincor. Flowserve claims that
the installation is successful, but this claim has not been verified. Most of the other
Flowserve MPPs are onshore, but one is installed on an offshore platform in the
Middle East. Flowserve is also looking for a user with whom it can develop a
downhole twin screw pump.

Flowserve pumps are designed to be installed horizontally. Usually, they are driven
by electric motor drives controlled by ASDs, although they may be driven by
natural gas engines or other drivers. This enables the flow to be controlled to match
production requirements. Flow is usually controlled from suction pressure or
wellhead manifold pressure.
Packaging
For new installations, Flowserve pumps are usually packaged as skid mounted units
or units ready for grouting. Units include strainers, valves, piping, pressure and
temperature gages, and discharge separators. Flowserve surface pumps are
commonly packaged by outside contractors, working from Flowserve engineering
standards. Of course, replacement pumps or new units can be purchased as
individual items.
Testing Facilities
The Flowserve testing facilities are in Brantford, Ontario, Canada. The facilities can
now test with water, oil, and injected air. Particulate cannot be added to the test
flowstream.

These pumps are usually run at 1,800 rpm but are also run slower, based on the
amount of particulate entrained in the stream and the economics of the project.

Flowserve’s standard material for the pump case is carbon steel, but the pumps can
also be made of stainless steel or duplex stainless steel. Flowserve is testing various
coatings for the screw tips, such as deloran, Stellite, and tungsten carbide. For the
case bore, chrome plating and Stellite are offered.
Mechanical Seals
Flowserve has sold MPPs with the single seal using a close clearance throat
bushing, as well as pressurized dual seals.
Like Bornemann, the Flowserve MP1 includes a liquid separation chamber with
injection back to suction to act as a screw sealant.

As of this writing, it is strongly recommended that only Bornemann or Leistritz twin
screw pumps be considered for Chevron applications. Currently, because of
Flowserve’s proven poor reliability in Chevron and non-Chevron installations, the
purchase of Flowserve MPPs is not recommended. Flowserve’s ability to provide a
reliable MPP is lacking. They are considerably behind Bornemann and Leistritz.
Currently, Flowserve is in the process of a complete pump redesign that also
includes its production and quality control processes. Hopefully, Flowserve will
improve.
660 Manufacturers—Helico-Axial MPPs
661 General
Framo, a Norwegian company, and Sulzer, a French company, are licensed to
manufacture helico-axial pumps based on the Poseidon technology that was
developed jointly by Total, Statoil, and the French Petroleum Institute.
The helico-axial type of MPP is discussed briefly in Section 623. It is an axial
flow/centrifugal pump in which pressure is boosted by multiple impellers in series.
Between each impeller, a diffuser directs the flow from the discharge of one
impeller to the suction of the next. An external view of this pump is shown in
Figure 600-12 and an internal view in Figure 600-13.
662 Framo
Applications
Framo built more than 60 helico-axial pumps, 23 of which are still operating. The
maximum capacity available is 100,000 bpd. The Framo MPP design can deliver a
pressure boost as high as 900 psi. Table 8 shows some of their installations.
Figure 600-36 shows that the Framo installations are predominately subsea, with 18
subsea MPPs. In addition, Framo has 2 installations onshore and 3 on offshore
platform decks.
Framo MPPs are used in a wide variety of applications, including decreasing
wellhead backpressure to increase production and to extend the life of a field. Its
subsea applications provide the pressure needed to get the production fluid to the
ocean’s surface, thereby increasing production. Framo also has units that pump fluid
through a single pipeline to a processing facility several miles away.

Framo can supply pumps that are horizontally or vertically mounted. Most Framo
pump drivers are electric. One is driven by a hydraulic turbine. Framo has also used
diesel and gas engine drivers.
Packaging
Framo offers a complete subsea package, including all relevant subsea tooling,
topside power, a subsea electrical distribution system, umbilical chords, and control
systems. One Framo subsea unit in the Lufeng (Statoil) field has been operating
since 1997. During that time, the pump lasted 7 years with no pump intervention.
This pump is not included in Figure 600-36, because it is not an MPP. It is a
conventional centrifugal pump, deployed subsea. This pump simply illustrates that
Framo has the ability to deploy reliable subsea pumps.
Fig. 600-36 Significant Framo MPP Installations (Information Courtesy of Framo)

Onshore Press.
Location or Offshore Capacity Boost GVF
Field Customer Subsea Date No. Each (bpd) (psi) (%) Driver HP
Poseidon Statoil Offshore 1994 1 20,000 600 40 electric 1,000
Gulfax A
SMUBS Norske Shell Subsea 1994 1 29,000 800 42 hydraulic 1,000
Zafiro Mobil Subsea 1998 3 35,000 500 75 electric 1,150
ETAP BP Subsea 1999 3 42,000 320 64 electric 1,200
Gulfax A&B Statoil Offshore 1999 3 40,000 360 50 electric 1,500
Lennox BHP UK Onshore 2003 1 75,000 430 65 electric 1,150
Celba Amerada Subsea 2003 5 11,000 600 75 electric 1,100
Hess
Mutineer/ Santos Subsea 2004 3 45,000 430 40 electric 1,500
Exeter
Huwaila ADCO Onshore 2006 2 23,000 1,100 65 electric 1,300
Schiehalion BP Subsea 2006 4 51,000 400 74 electric 2,400
Tordis Statoil Subsea 2006 4 53,000 500 68 electric 3,000
Shaybah Saudi Arabia Onshore 2006 1 100,000 450 87 electric 1,300
Vincent Woodside E. Subsea 2007 4 100,000 600 70 electric 2,400
Brenda #2 Oilexco N.S. Subsea 2008 1 100,000 500 65 electric 1,500
Testing Facilities
Framo has a complete full-scale test facility in Norway. It can test subsea pumps,
fully submerged and dry mounted, under actual field conditions. The Framo test
facility includes a multiphase test loop equipped with a separator, pumps, and
compressor.

Note that these pumps run fast, up to 5,100 rpm in Framo’s case. With this high
speed, the pump would be expected to suffer erosion if sand is present. Framo tested
one of its pumps with a high concentration of sand, using different coatings on the
impellers and diffusers. Framo claims that “the impeller/diffuser erosion was
minimal”. It is not clear what is meant by “minimal”, but apparently, some wear did
occur, and the duration of the test is unknown. Coatings that were tested include
Stellite and tungsten carbide. Chevron has no experience with sand erosion in a
helico-axial pump.

Framo can make its pump’s cases out of carbon steel for onshore applications or
duplex or super duplex for subsea and offshore platform applications. For
particulate handling and reduced wear, Framo will coat their impellers and diffusers
with different coatings, including Stellite and tungsten carbide.
663 Sulzer
Applications
Sulzer has built 21 helico-axial pumps. The maximum capacity available is Model
MPP12 at approximately 650,000 bpd. Sulzer also has MPPs deployed that can
deliver a maximum pressure boost of 1200 psi. Figure 600-37 shows Sulzer
installations.
From Figure 600-37, it can be seen that most of Sulzer’s installations are either
onshore or on an offshore platform. It includes only one MPP for subsea, the
Nautilus, but it was never deployed. It was a vertical pump, while the remaining
pumps listed in Figure 600-37 are horizontal.
Fig. 600-37 Significant Sulzer MPP Installations (Information Courtesy of Sulzer Pumps)
Onshore Press.
Location or Offshore Capacity Boost GVF
Field Customer Subsea Date No. Model Each (bpd) (psi) (%) Driver HP
France Elf Onshore 1994 1 MPP3 50,000 600 100 Electric 600
Siberia JSC Onshore 1997 2 MPP7 42,000 200 86 Electric 540
Chernogornefi
Nautilus(1) Total Subsea 1998 1 MPP4 50,000 600 71 Electric 1,750
Duri, Caltex Pacific Onshore 1998 1 MPP7 125,000 150 100 Electric 700
Indonesia
North Sea Dunbar Dev. Offshore 1999 2 MPP5 90,000 1,100 90 Electric 6,000
Saudi Aramco Onshore 2000 1 MPP7 47,000 300 59 Electric 750
Arabia
Siberia Yukos Onshore 2001 2 MPP11 280,000 540 91 Electric 8,800
Siberia TNK Onshore 2001 2 MPP 7 75,000 210 90 Electric 2,650
Nizhnevartovsk
Siberia Yukos Onshore 2002 2 MPP11 280,000 540 91 Electric 8,800
Siberia TNK Onshore 2003 1 MPP7 150,000 210 90 Electric 3,210
Nizhnevartovsk
Siberia TNK BP Onshore 2004 1 MPP7 260,000 110 90 Electric 3,210
Algeria Agip Onshore 2006 2 MPP7 60,000 1,100 90 Electric 3,350
North Sea BP Offshore 2007 1 MPP8 180,000 1,200 95 Electric 2,650
Korea Lotte Dassan Onshore 2007 2 MPP6 14,000 340 20 Electric 70
1 Never deployed.
One of the pumps listed in Figure 600-37 is the Chevron Duri helico-axial pump,
installed for testing in 1998. It was run for only a few months. (Refer to
Section 6138.)
Like all the other MPPs, the Sulzer installations were predominately used to
transport multiphase fluid long distances. Sulzer also has some that increased
production by reducing wellhead backpressure, where it claims the economic
payouts are usually less than 2 years. Finally, Sulzer claims it has re-activated dead
wells using their MPPs.

Sulzer can supply pumps that are horizontally or vertically mounted. All pumps
supplied to date are electric motor driven.
Packaging
To date, Sulzer has not packaged its pumps on skids.
Testing Facilities
Sulzer has an existing test loop in France and another loop under construction in the
UK. Both will be able test an MPP with a mixture of nitrogen and water. Each one
has the same maximum capacity of 4,000 m3/hr (600,000 bpd), but the UK unit is
considered larger because it will be able to pump more water at 700 m3/hr
(11,000 bpd) and nitrogen at 3,500 m3/hr (530,000 bpd). Sulzer can test at just about
any suction pressure up to 430 psi.
Because Sulzer tests with a water/nitrogen mixture, it needs to make adjustments to
match the pump’s characteristics to that of a proposed crude oil/natural gas mixture.
Sulzer does this by adjusting the pump suction’s water/nitrogen density such that it
is the same as that expected under crude oil/natural gas.

If sand is present, Sulzer coats its impellers and diffusers with a proprietary coating,
called Sulzer Metco, “SUME”. Sulzer claims the coating is 5 to 10 times more wear
resistant to sand than stainless steel and/or carbide coatings. Sulzer has a couple of
installations in Siberia where the sand content is 300 ppm.
Like Framo helico-axial pumps, Sulzer MPPs run fast, up to 6,800 rpm. With this
speed, if sand is present, erosion would be expected, even with a SUME coating.

Sulzer makes its pump cases out of carbon steel for onshore applications and either
duplex or super duplex for offshore platform installations. For applications that
involve particulate in the incoming flowstream, Sulzer will coat the impellers and
diffusers with SUME. (Refer to Particulate Handling and Speed in Section 663.)

On several occasions, Chevron has requested that Framo demonstrate or present
data to show that its pump’s wear from particulate is indeed minimal, but Chevron
has not received much supporting evidence from Framo, other than statements that
the wear is minimal. Framo claims that the angle at which the flowstream
particulate hits the impeller vane has been optimized, minimizing the wear. (This
angle is usually found to be roughly 30 degrees.) The problem is that the angle of
the incoming flowstream particulate (as it leaves the case and contacts the impeller)
will change, depending on its speed and density. Thus, in reality, a variable angle is
required, such as perhaps, one provided by a variable inlet guide vane.
Framo has tested its MPP with quartz sand in water using a tungsten carbide coating
on the leading edge of the impellers, the wear rings, and other potentially wearing
parts. The pump eroded somewhat with a significant reduction in efficiency.
Chevron has no operational experience with sand erosion in a helico-axial pump
and, to date, does not recommend these pumps be used in particulate containing
streams. Anyone who considers a helico-axial pump for use where sand is present
should investigate the erosion issue more thoroughly before proceeding with any
purchase.
The reliability of helico-axial pumps is suspect. Some applications have MTBFs on
the order of 5 years, but many have MTBFs of only 2 years. At a presentation,
Framo listed a number of installations, but when questioned about their reliability,
Framo often found fault with the installation, the application, or the operation of the
equipment for the failures which often occurred within 2 years.
670 Manufacturers—Progressing Cavity MPPs
671 General
There are two major manufacturers of progressing cavity pumps (PCPs) that can be
designed for multiphase applications. They are Moyno and seepex. Bornemann,
Netzsch, Mono, and Tarby also manufacture PCPs, but they have not actively tried
to gain MPP business and are not discussed in this section.
As discussed in Section 620, a PCP is a single, hardened steel alloy serpentine rotor
that revolves inside and seals against a stationary stator, usually manufactured of an
elastomeric material. Lubrication between the two is absolutely necessary. The
pumped fluid enters the pump near the coupling and is discharged axially at the end
of the screw.
Most of the PCPs used in multiphase service are installed on the surface, lying
horizontally. Some are also installed vertically, downhole, run by a long shaft from
an electric motor on the surface. Also, their speed is often controlled by an ASD.
PCP manufacturers claim they can pump multiphase fluids up to 100 percent gas.
Despite this, and because a PCP must never run dry, surface PCPs should be limited
to a GVF of 30 percent, unless an adequate and reliable external or recycled liquid
stream is introduced into suction that keeps the GVF at or below 40 percent. The
maximum GVF for PCPs deployed down hole should be 40 percent. Finally, PCPs
should not be used if slugging will occur, because it is likely that the GVF will
exceed 40 percent under these conditions. As mentioned earlier, slugging always
leads to a GVF of 100 percent. Chevron’s experience to date indicates that, even if
pumped fluids are restricted to GVFs in the 40 percent range, the MTBFs for PCPs
will typically run in the 2 year to 3 year range.
672 Moyno
Applications
The Moyno MPP is called the “Moyno Tri-Phase System”. Moyno claims to have
manufactured 150. Most of the Moyno installations are in Canada and the United
States. A lesser number are installed in Venezuela, Trinidad, Mexico, and Colombia.
All Moyno machines are installed on the surface. Most are onshore, with a few on
offshore platforms. Moyno has not deployed any pumps subsea.
Moyno claims it can pump up to 60,000 bpd of multiphase fluid to 900 psi with a
GVF up to 95 percent. Moyno claims that its PCP MPPs are reliable, running up to
10 years with minimal maintenance. Figure 600-38 is a list of some of the Moyno
installations.
As can be seen from Figure 600-38, most Moyno pumps are small, with only one
having a flowrate as high as 60,000 bpd with a 125 HP electric motor. What is
unusual is that many of Moyno’s earlier pumps used packing. The later Moyno
installations use a modified single or dual pressurized mechanical seal. Packing or
even a conventional single mechanical seal, without a close clearance bushing, is
not recommended for environmental and safety reasons. The modified single
mechanical seal typically uses an API flush plan 11, 21, or 32 to the seal chamber,
entering the pump through a close clearance throat bushing. This type of single seal
is described in Single Mechanical Seals in Section 632.
Fig. 600-38 Significant Moyno MPP Installations (Information Courtesy of Moyno)(1)
Location or Capacity Seal or
Oil Field Customer Date No. Model (bpd) Packing Driver HP
Canada Renaissance E Jan-97 1 BFD-2 29,000 Packing Electric 100
Canada Probe Explor Jan-98 1 BPD-4 7,000 Packing Electric 30
Venezuela Maxus Jul-98 1 BPD-4 12,000 Packing Electric 75
Canada TriLink Apr-99 1 BPD-5 7,000 Packing Electric 20
USA Mobil Exp Aug-99 1 BPD-2 29,000 Dual S. Electric 125
Canada BP Amoco Nov-99 1 BPD-4 7,000 Packing Electric 30
USA U of Tulsa Sep-00 1 BPD-6 11,000 Single S. Electric 75
Canada Pancanadian Dec-00 1 BPD-4 12,000 Packing Electric 50
Canada AEC Oil/Gas Mar-08 1 BPD-4 7,000 Single S. Electric 60
Canada Triumph E. Apr-08 1 BPD-2 3,000 Packing Electric 10
USA Exxon/Mobil Aug-08 1 BPD-2 29,000 Dual S. Electric 125
Colombia BP Exp Apr-00 1 BPD-2 29,000 Dual S. Electric 125
USA Control CTR Sep-08 1 BPD-1 60,000 Single S. Electric 125
Trinidad Onyx Res Jan-08 1 BPD-4 5,000 Dual S. Electric 50
Mexico Pemex - 1 BPD-4 15,000 Dual S. Electric 100
1 Note: Information for Location (onshore, offshore or subsea), Delta P, and GVF is not available
Most of the Moyno pumps reduce wellhead backpressure to increase production and
extend the life of oil wells. Moyno also points out that its pumps have, in some
cases, brought back to life “dead” wells.

Moyno can supply pumps for use in the horizontal position. The drivers to date have
all been electric motors, most driven through ASDs.
Packaging
Moyno will provide packaged pump skids with valving and instrumentation if
requested.
Testing Facilities
Moyno does not own a test loop testing facility. It has used others in the U.S. and
internationally.

Moyno MPPs run with a maximum speed of 550 rpm. Moyno claims it has
experience pumping sand at concentrations up to 5 percent. The Moyno MPPs have
also pumped salt, which can be just as abrasive.
Materials and Corrosion

The standard material for Moyno pumps is 316 SS for the rotor and a nitrile based
elastomer for the stator. Moyno coats its rotors with tungsten carbide if particulate is
in the fluid stream. Other materials are available upon request.
Moyno pumps are conventional, with an elastomer stator and metal rotor. They have
no unique features.
673 seepex
Applications
seepex has factories in Ohio and Germany. Before 2004, seepex supplied
approximately 50 MPPs through another company to extract water and methane gas
from coal bed seams. Since 2004, seepex has made 16 PCP MPPs with 12 still
installed—all onshore. Currently, seepex has an order for two additional pumps to
be deployed on offshore platforms. seepex does not have any subsea pumps yet, but
one is scheduled for development by 2010.
seepex installations include reducing wellhead backpressure to gain more
production. seepex also claims to have successfully replaced other suppliers’ pumps
that were gas locking. A complete list of seepex installations is not currently
available.
The maximum flowrate available for a single seepex MPP is currently 50,000 bpd.
The pump’s maximum pressure boost is 600 psi. However, seepex claims a higher
pressure boost can be obtained by connecting pumps in series. A system that uses
PD pumps in series is very difficult to control and often leads to failure. PCP
pumps, as has been mentioned several times earlier, cannot run dry, thus the control
system for series operation is very complicated. Unless a pressure vessel is installed
between the pumps to provide a cushion of fluid for the second pump, the second
pump can easily run dry if the flow from the first pump is not adequate. Series
operation of PCPs should be avoided, unless there is just no other pump or pumping
system that can meet the application need, and the pressure vessel between the two
pumps is included in the system design.

They are all motor driven, usually using ASDs.
Packaging
seepex does not provide packaged pump skids with valving and instrumentation.
Testing Facilities
seepex claims it has tested its pumps in its flow loop in Ohio at a continuous GVF
of 99 percent and up to 100 percent for a short duration. seepex uses nitrogen for
testing. seepex claims that 1 percent nitrogen or less is enough to keep the pump
lubricated and running without damage. seepex further claims that nitrogen and also
natural gas, directly from an oil well, provide better lubrication than air. To date,
Chevron has no experience with these tests or the testing facilities. Until more
information is available, these pumps should be operated at a GVF of 40 percent or
less.
seepex further claims that its pumps exhibit good reliability because it uses:
• A complicated computer program to optimize the exact interference fit between
the rotor and stator, which minimizes the contact surface;
• A proprietary elastomer, hydrogenated nitrile (HNBR) for stator material,
which is further chlorinated to reduce friction and heat generation;
• An “Equal Wall Stator” that has a uniform rubber thickness all around each
cavity. This is different than a conventional stator, where the elastomer has a
varying thickness. seepex claims this design allows higher pressure boosts,
dissipates heat more efficiently, and allows the pump to handle higher GVFs. It
is also much shorter than a conventional stator, which is an advantage, espe-
cially on an offshore platform where space is a premium.

seepex pumps run at a maximum of 350 rpm.

seepex makes its PCP cases out of various materials, including carbon steel,
stainless steel, and duplex. The rotors can be made out of these same materials, plus
titanium. For a high GVF, seepex recommends coating the rotors with its
proprietary hard chrome material, called Duktil. In addition to being good at a high
GVF, the Duktil material offers good abrasion resistance. Additionally, if pumping
high GVF streams, seepex recommends chlorinated HNBR as stator material to
reduce friction.
Having an elastomeric stator and metal rotor, the seepex basic pump designs are
conventional. However, the seepex “Equal Wall Stator”, described in Testing
Facilities in Section 673 and Equal Wall Stator in Section 6144, is unique in the
industry.

Chevron has had success with PCPs but only in services that have approximately
40 percent GVF or less. Historically, Chevron has had field problems pumping
higher GVF streams due to poor lubrication between rotor and the stator. As
mentioned before, the MTBF of PCPs in multiphase service is typically 2 years to 3
years.
680 Manufacturers—Electric Submersible (ESP) MPPs
681 General
Most ESPs are installed downhole. They must fit in the well casing, and for this
reason, the diameter of the pump’s body is small, and there are many pump stages to
develop the head. Thus, the pumps are very long and thin. The ESP consists of
several impellers in series with sleeve type bearings between each impeller (in the
more robust designs), all lubricated by the fluid being pumped. An ESP assembly
for a multiphase application, from bottom to top, consists of a motor, a protector
(also called the seal), usually a gas separator, sometimes a gas handler, and the ESP
itself. A cross section of an ESP is shown in Figure 600-20.
The two most common ESP manufacturers are Centrilift and Schlumberger (Reda).
Each has been active in trying to capture additional MPP business. Both claim that
their units can handle high concentrations of gas, 90 to 100 percent, with gas
separators and gas handlers. Even with these devices, ESPs should not be used if the
GVF exceeds 60 percent. With slugging, the MPP will experience a 100 percent
GVF stream at times, and if this is anticipated, ESPs should not be used.
682 Centrilift
Applications
Centrilift ESPs are common throughout the oil industry. Centrilift defines an MPP
as one that has a GVF of over 20 percent. More than 25 percent of Centrilift
installations meet this definition.
The maximum capacity of Centrilift MPPs is 14,000 bpd, with a maximum pressure
boost of 5,000 psi. These maximums, based on a fairly low GVF, decrease as the
GVF increases. Note that, since an ESP is usually installed downhole, 14,000 bpd
represents the production from a very large well.
An installation list, such as the one that describes Schlumberger ESPs in

Figure 600-39, is not available. Centrilift ESPs are deployed as MPPs in at least
three Chevron installations: in Thailand, New Mexico, and California. The
installation in New Mexico involves 27 ESPs that pump an unknown quantity of gas
but probably less than 40 percent. They have been operated for only 2 years, but to
date, no failures have occurred.
Centrilift applies its MPPs in oil wells that are onshore, on offshore platforms, and
subsea. Centrilift has installed a few onshore and on platforms in the horizontal
orientation. As with many MPPs, Centrilift MPPs have been used to increase
production by decreasing wellhead backpressure.
If applied subsea, the pumps typically are installed in a subsea well(s) and pump the
well’s fluid to the ocean’s surface. To decrease the pressure boost, the ESP can
discharge into a skid on the seabed floor. The skid collects fluid from several ESPs
or wells and uses another ESP to send the fluid to the surface. In these cases, they
offer a complete design for the subsea installation.
Centrilift has the ability to apply its pumps at temperatures up to 400°F (204.4°C).
Their MPPs are sized as any other multistage ESP. Centrilift uses a proprietary
computer program, AutoGraphPC, that analyzes the fluid volume in each stage as
the fluid progresses up the pump.

As stated in Section 681, Centrilift ESPs can be positioned vertically or horizontally
in a well or horizontally on the surface. All Centrilift pumps are electrically driven.
Testing Facilities
Centrilift has a test facility in Claremore, Oklahoma, where it can test ESPs up to
10,000 bpd and vary the GVFs up to a maximum of 50 percent.

Centrilift usually runs its pumps at 3,600 rpm but has the ability to run them up to
4,800 rpm, using an ASD to increase the electrical frequency. If sand or other
abrasives are present, Centrilift coats its impellers and bearings with a surface
treatment called Armor I or Armor X. Note that the bearings are lubricated by the
produced fluid. If the bearings were uncoated, the sleeve bearings between each
impeller (in the robust design) would wear quickly in the presence of sand.
Centrilift claims the Armor X is harder than Ni-Resist or even silica sand. Armor X
is almost as hard as tungsten carbide. The intent is to crush the sand getting into the
bearings and flush it through the bearings with the pumped stream without damage.
This becomes problematic as the GVF of the stream increases, since gas will not
effectively flush particulate away or provide lubrication.

Centrilift offers a wide variety of materials, depending on the production fluid
chemistry and the environment. As mentioned in Materials and Coatings in
Section 682, Centrilift coats its bearings with Armor X to extend the life of the
bearing when sand is present.
Design Improvements
Centrilift claims it has a unique MPP impeller design that avoids gas locking, and
this alone can pump up to 70 percent free gas. Gas locking is common for
centrifugal pumps and, especially, ESPs, if fed a high GVF. To avoid gas locking,
Centrilift impellers deploy a steep vane angle with large balance holes.
Centrilift offers two types of gas separators: rotary and vortex. One or the other is
used and attached ahead of the inlet to their ESP. The separated gas is vented into
the annulus between the pump or tubing and the casing, above the liquid level.
Centrilift claims that its pumps can handle 90 percent free gas with the vortex
separator.
683 Schlumberger (Reda)

Applications
Schlumberger ESP MPPs are installed downhole in an oil well. They are not
deployed horizontally on the surface.
Their maximum flowrate is 9,000 bpd, with a maximum pressure boost of 4,000 psi.
As with Centrilift, these numbers decrease as the GVF increases.
As of 2004, Schlumberger deployed 17 MPPs, as shown in Figure 600-39.
Fig. 600-39 Significant Schlumberger ESP MPP Installations (Information Courtesy of Schlumberger) (1 of 2)
Onshore
Colombia Hocol Onshore Apr-02 1 538 - - 60 Electric -
New Mexico Devon Onshore Aug-03 1 538 - - 65 Electric -
New Mexico Devon Onshore Apr-04 1 538 - - 65 Electric -
Midland, TX Aethon Onshore Mar-04 1 400 - - 30 Electric -
New Mexico Devon Onshore Mar-04 1 400 - - >60 Electric -
Midland, TX Oxy Onshore Apr-04 1 400 - - - Electric -
Garden Grove Oxy Onshore Aug-04 1 400 370 3,000 - Electric 250
Garden Grove Oxy Onshore Aug-04 1 538 - - - Electric -
Garden Grove Oxy Onshore - 1 538 - - - Electric -
New Mexico Devon Onshore - 1 400 - - - Electric -
Fig. 600-39 Significant Schlumberger ESP MPP Installations (Information Courtesy of Schlumberger) (2 of 2)
Onshore
Qatar QPD Offshore - 1 538 - - - Electric -
Colombia Hocol Onshore - 1 538 - - - Electric -
Elk Hills, CA Occidental Onshore Feb-05 1 - - - >50 Electric -
Schlumberger has installed MPPs in oil wells that are onshore, with at least one
associated with an offshore platform. Schlumberger has not deployed any in subsea
locations. The pumps are used primarily to decrease a well’s downhole
backpressure, thereby increasing production. Schlumberger has also replaced
conventional ESPs with its MPP version, using their gas handler, after the previous
conventional ESP gas locked.
Schlumberger ESPs are manufactured by its subsidiary, Reda. Schlumberger also
owns Framo, who makes helico-axial MPPs. (Refer to Section 623 for a discussion
on helico-axial pumps in general and Section 062 for specific information on
Framo). As a result, Schlumberger offers a gas handler that is based on Poseidon
technology. It is a helico-axial pump called the Poseidon ESP that is 6.3 meters long
and is attached ahead of the suction of a Reda ESP. With this helico-axial gas
handler, Schlumberger claims it can pump a GVF up to 75 percent. Schlumberger’s
first gas handler unit was installed in 2003, designed to handle approximately
60 percent gas. Since 2003, Schlumberger has installed 16 similar machines at
various locations, none of which involve a GVF as high as 75 percent.
Schlumberger claims that its Poseidon gas handler can pump such a high GVF
because it homogenizes the mixture, puts some of the gas back in solution, and
increases the fluid pressure, reducing the gas volume and bubble size as the stream
enters the conventional ESP, the next section after the Poseidon.
Schlumberger has several MPPs with the Poseidon gas handler installed in fields
that experience a GVF from 45 to 65 percent. One of them listed in Figure 600-38,
in Colombia, involved its Poseidon MPP replacing a conventional ESP.
Schlumberger claims that the new MPP eliminated gas locking and increased
production by 100 percent, pumping a fluid with a GVF of 60 percent. Another
installation in Elk Hills included a Poseidon MPP installed in the vertical portion of
a well that also includes a horizontal segment. Schlumberger claims that the MPP
increased production by 40 percent.
Finally, a Poseidon pump was installed in the Garden Grove field. This field is a
CO2 injection field, and the Poseidon pump was installed to replace a conventional
ESP that gas locked from CO2 breakthroughs. Schlumberger claims the new MPP
increased production by 130 percent.
Schlumberger uses its proprietary software to size their MPPs. During this process,
it derates the pump to offset a high GVF.

As mentioned, Schlumberger pumps are deployed downhole in wells in the vertical
and horizontal positions. The pumps have not been deployed on the surface. All
Schlumberger pumps are electrically driven. Schlumberger MPPs are controlled to
shut down on high motor temperature, as opposed to the conventional approach of
using motor current. If the GVF is high, motor current is not a reliable indicator of
motor problems.
Packaging
Schlumberger packages its pumps itself. It packages them to be installed in pods,
under packers, and in series (two ESPs in one well).
Testing Facilities
Schlumberger has a testing facility in Bartlesville, Oklahoma, and another in
Rosharon, Texas. Schlumberger test facilities include two test wells that can test
pumps requiring up to 1,500 HP. Schlumberger does not test with natural gas, using
air and water instead. Schlumberger also tests bearings without liquid lubrication to
evaluate conditions with slugs of gas.

Schlumberger tries to mitigate the erosive affects of sand with a proprietary hard
material called 5530. Schlumberger can run its pumps up to 80 hz or approximately
4,800 rpm, using an ASD to control the frequency.

The materials used in Schlumberger MPPs depend on the fluid chemistry and the
environmental conditions in the well. It offers a wide variety of materials, including
5530 metallurgy. If CO2 is present, Schlumberger offers a material called
“Redalloy”. In addition, Schlumberger uses a silicon carbide for the bearings.
Schlumberger claims that this silicon carbide bearing material is self-lubricating,
offsetting the effect of poor lubrication from the gas-laden multiphase fluid.

ESPs are considered to be reliable for applications having GVFs of approximately
40 percent. However, even liquid only ESPs have MTBFs of only 2 years to 3 years.
It is difficult to get information on ESPs that demonstrate even this reliability when
the GFV exceeds 60 percent. Despite claims by Centrilift and Schlumberger that
they can do better, actual proven reliability to date is not impressive.
As mentioned earlier and Centrilift agrees, the average ESP is pulled every 2 years
to 3 years for maintenance. Some of the reasons are described in Driver in
Section 625. ESP reliability will decrease further if the pumped fluid is a multiphase
fluid having a substantial amount of particulate and a GVF over 60 percent.
Based on operating experience throughout Chevron, ESPs should not be used

horizontally, above grade, in place of an API 610 pump. Chevron has several
examples of this orientation, and the reliability has been poor, requiring frequent
maintenance, with availability sometimes being only a few months.
690 Sizing of the MPP, Its Driver, and Associated Facilities (Upstream
Applications)
691 Twin Screw MPP Design Guide for Upstream Applications

Figure 600-40 shows the MPP design guide for upstream applications.
Fig. 600-40 Twin Screw MPP Design Guide for Upstream Applications
692 Sizing
The following are required to size an MPP:
• Average and maximum anticipated gas volume fraction (GVF);
The GVF is simply the percent gas in the total flow. (The units for gas flow are
standard cubic feet per day or actual cubic feet per day, with the suction
pressure and temperature provided.) GVF percentage will determine if a screw
sealant is necessary. Screw sealant is discussed in Section 633. Gas oil ratio
(GOR) can also be provided in place of GVF, but the liquid water must then be
provided.
• Liquid oil flowrate (bbl per day);
• Liquid water flowrate (bbl per day);
• Constituent breakdown of the liquid (e.g., percent composition of water, oil,
natural gasoline, gas/oil fraction, and sand);
• Suction pressure to pump;
• Suction temperature to pump;
• Discharge pressure;
• Liquid viscosity;
• The sand or particulate quantity by weight and volume, size distribution, its
composition (i.e., is it quartz or any other hard mineral), and its shape (sharp or
rounded edges);
This affects the size of the inlet screen mesh, the types of coatings for the
internal parts, and the pump’s internal clearances. (Refer to Section 637.)
• Wax or paraffin content and cloud point;
Solid wax particulate in the inlet stream will plug up the MPP suction strainer
or filter. Either a duplex strainer or duplex filter is always recommended to
catch particulate and debris in the inlet stream, including applications in which
wax is not expected. If wax is present, a duplex filter (not the coarser mesh
duplex strainer) should be used. The filter must be sized to have a larger area
and a greater collapse differential pressure than the duplex strainer. The duplex
design is always recommended to enable switching screens while operating.
Specific recommendations are given in Suction Strainer in Section 639. For
example, refer to Section 6135, Princess Field Lessons Learned, item B and
Section 6136, Main Pass 313 Lessons Learned, item F.
• The flow regime (in particular, the tendency for slugging);
Slugging may not be a significant problem if the pump is located near the
well(s). However, slugging could be severe if the pump is located a distance
away with the suction line traversing hilly terrain. Slugging is discussed in
Section 635. It is often recommended that the suction and discharge pipe be
simulated to determine the extent of slugging. Refer to Section 639 for
additional information on this subject on Mitsue Field, Lessons Learned.
• Type of driver: electric motor direct drive (with or without an adjustable speed
drive), a diesel engine driver, natural gas engine driver, hydraulic turbine drive,
belt drive, etc.;
– An electric motor direct drive runs the MPP at the speed of the driver, but
the speed can be varied if an adjustable speed drive (ASD) is installed. An
ASD allows the operators to change speed to compensate for a field’s
changing flow conditions or to increase the speed to maintain the same
production rate as the pump wears or to simply make up for unanticipated
conditions that show up after the pump is installed.
– By design, the diesel engine, natural gas engine, and hydraulic turbine
drives are all adjustable speed drives. These drives can all be used to vary
the speed of the MPP by operating the pump speed off a specified set
suction pressure.
– Hydraulic drives are available and sometimes very appropriate, especially
if a field is being water flooded.
– Belt drives, though not recommended due to maintenance problems, are
much cheaper than the others, and with them, speed can be changed in
steps by replacing sheaves and belts. Note that belt drives are limited to
approximately 250 HP.
– Additional criteria for choosing the type of drive are the size and weight
allowed by the application (especially if downhole, on a platform, or
subsea), its reliability, and the type of fuel available (e.g., diesel, natural
gas, or electric power).
• Estimated variation of flow conditions over the design life of the field or the
pump, realizing that there are many unknowns that cannot be determined prior
to purchasing the MPP, such as:
– What production rate will be possible as the field ages?
– What composition changes (e.g., percentages of water, gas, particulate,
H2S, and crude) can be anticipated in an existing field after the installation
of an MPP? (Changes in the percentage of H2S will affect the pump’s
material selection.)
– An estimate of the recycle, slip, and flashing factor. The pump’s recycle
liquid is usually hotter than the feed to the MPP. Therefore, it may flash
into gas, occupying the capacity of the pump. If this is expected, the pump
may need to be designed larger. Flashing is discussed in Section 638.
Recycle, slip, and flashing are discussed in Section 695.
Due to these uncertainties, the MPP should be designed with flexibility in mind. If
the MPP is electric motor driven (and most are), an ASD is strongly recommended
such that the speed of the MPP can be adjusted to compensate for potential
production increases or decreases. If practical for additional flexibility, an API 682
Seal Flush Plan 32 is recommended, rather than relying on a Seal Flush Plan 11 or
31. Section 6100 also offers some assistance in estimating future production flow
conditions.
693 Pump Sizing Calculation

MPP suppliers rate their pump capacities in barrels per day of mixed flow at suction
conditions. Mixed flow means the combined flow of oil, water, particulate, and gas
in barrels per day. Normally, gas is described as either standard cubic feet per day
(SCFD) or actual cubic feet per day at suction conditions (ACFD). Sometimes, it is
described by the GOR.
The following equation shows the conversion of the flow of gas to equivalent
barrels per day (BPD) at suction temperature and pressure:
Actual cubic feet per day (acfd) =
(Gas rate, scfd) (14.7/Actual suction press in psia) (Actual suction temp/520°R) Z,
(Eq. 600-1)
where:
Z is the compressibility factor and 520°R = 460°F + 60°F (Z is often
taken as 1).
After the gas rate is converted to actual cubic feet per day, it must be further
converted to bpd using the following conversion:
Actual barrels per day (abpd) =
(Actual gas flowrate, acfd) (7.481 gal/ft3) (1 bbl/42 gal)
(Eq. 600-2)
Example:
Given: oil production rate = 1,000 bopd
GOR = 400 scf/bbl
Water cut (wc) = 25% of the liquid rate (water and oil)
Suction pressure = 50 psig
Suction temperature = 150°F
Discharge pressure = 350 psig
API gravity = 24
Gas compressibility, Z = 0.98
Steps:
1. Determine the gas rate as follows:
(1,000 bpd) (400 scf/bopd) = 400,000 scfd
150°F + 460°F = 610°R
50 psig + 14.7 = 64.7 psia
Actual gas flowrate = (400,000 scfd) (14.7 psia/64.7 psia) (610°R/520°R) (0.98)
= 104,870 acfd
Actual bpd of gas = (104,870 acfd) (7.481 gal/ft3) (1 bbl/42 gal)
= 18,680 bpd of gas
2. Determine the water rate:

Water rate = (wc/(100  wc)) (oil rate)
= (25/(100  25)) (1,000 bopd) = 330 bbl of water per day (bwpd)
3. Determine the total liquid rate:
Total liquid rate = 1,000 + 330 = 1,330 bbl of liquid per day (blpd)
= 40 gpm
4. Determine the total mixed flowrate of all substances:
Total flowrate to the pump suction = (18,680 + 1,330) bpd
= 20,010 bpd (at 50 psig and 150°F)
694 Gas Volume Fraction (GVF)

Suppliers use the term gas volume (or void) fraction (GVF). This is simply the
actual gas flowrate (by volume, determined at suction conditions) divided by the
total mixed flowrate (by volume). For the above example:
GVF = 18,680/20,010 = 0.93 (or 93%)

(Eq. 600-3)
695 The Recycle, Slip, Flashing Factor for Twin Screw MPPs
An MPP must be sized to handle more that just the design capacity obtained from
Section 691 because of both external and internal losses. These losses can come
from external liquid recycle, which may flash into gas or internal slip (flow from
discharge back to suction). Flashing is discussed in Section 638. Refer to Recycle,
Slip, Flashing Factor in Section 638 for specific values for the recycle, slip, flashing
factor.
As mentioned in Recycle, Slip, Flashing Factor in Section 638, one simply
multiplies the design capacity obtained in Section 691 by the recycle, slip, flashing
factor to get the actual pump capacity.
In the above example, if the factor is 1.15 percent, the design flowrate of the pump
would be:
Design flowrate = 20,010 bpd (1.15) = 23,011 bpd or 670 gpm
(Eq. 600-4)
Section 638 recommends that once the MPP flow diagram is determined, the
streams be simulated using HYSIM or PRO II. This simulation will provide a good
estimate of the amount of flashing and, thus, a better estimate of the increased pump
capacity one needs to add. Section 6910 shows and describes a typical flow
diagram.
696 Pump Efficiency for Twin Screw MPPs

“Slip” for a twin screw pump is defined as fluid that flows backwards through the
screw clearances. As slip increases, the pump efficiency is reduced. Slip increases
as the pump differential pressure increases or if the internal pump clearances
increase due to wear. Slip decreases as the viscosity of the fluid increases and as the
pump speed increases.
If pumping a pure incompressible liquid, the volume of the pumped fluid does not
change, and the increasing pressure difference across each lock is constant. (Refer to
Figure 600-41.) In this pure liquid application, the pump’s efficiency is maximized.
Fig. 600-41 Pressure Distribution when Pumping Only Liquid
DISCHARGE PRESSURE
INCREASING PRESSURE
SUCTION PRESSURE
DIRECTION OF FLOW
SLIP
SCREW
LOCKS
LIQUID ONLY
If pumping a multiphase fluid with a substantial GVF, however, most of the pressure
rise occurs in the final stages of the pump. (Refer to Figure 600-42.) This is due to
the compressibility of the large volume of gas at the inlet, where the gas volume is
substantially reduced, with relatively low pressure increases.
The low pressure rise across the first few locks means that, in relative terms, a
reduced amount of slip exists there, with an increased amount of slip across the
latter locks (those near the outlet). This reduced amount of slip across the first few
locks is somewhat offset by the fluid’s lower viscosity there, which is due to the
relatively large amount of gas in the stream. This reduced viscosity allows the fluid
to slip easier through the first few locks, versus later in the pump, where the volume
of gas is greatly reduced, and the multiphase fluid becomes more viscous. The
bottom line is that the pump efficiency will be lower if pumping a multiphase fluid
than if pumping a pure liquid.
697 Overall Mechanical Efficiency Calculation

Mechanical efficiency is defined as the theoretical horsepower required for pumping
liquid and gas, divided by the actual horsepower delivered to the shaft.
This is shown in the following equation:
Fig. 600-42 Pressure Distribution when Pumping a Multiphase Fluid
DISCHARGE PRESSURE
INCREASING PRESSURE
SUCTION PRESSURE
DIRECTION OF FLOW
SLIP
GAS PHASE LIQUID PHASE
Overall mechanical efficiency =

(Theoretical gas HP + Theoretical liquid HP)/shaft HP
(Eq. 600-5)
where:
Theoretical liquid hp = [liquid rate (gpm)] [pump DP (psi)]/1,714
Theoretical gas HP = (Inlet press, psia) (Gas rate, acfm) (ln [(1+ pump DP/inlet
pressure)/229.7])
Actual shaft HP = [equivalent mixed flowrate (gpm)] [pump DP (psi)]/1,714 (the
mechanical efficiency factor)
The mechanical efficiency factor is usually 1.1 to account for
inefficiencies due to bearings, gear boxes, seals, etc.
For the above example:
Theoretical liquid HP = (40 gpm) (300 psi)/1,714 = 5.0 HP
Theoretical gas HP = (64.7 psia) (73 acfm) [ln(1+ (300 psi/64.7 psia))/29.17] = 35 HP
(Note that both equations are rounded to nearest 5 HP.)
Actual shaft HP = [(670 gpm) (300 psi)/1,714] (1.1) = 130 HP
Overall mechanical efficiency = (Gas HP + liquid HP)/shaft HP = (5 + 35)/130 = 0.31 or 31%
698 Volumetric Efficiency Calculation

Volumetric efficiency is the delivered pump capacity divided by its theoretical
capacity. Mathematically, it is expressed as:
Volumetric efficiency, Ev = Qc/Qt

(Eq. 600-6)
where:
Ev = volumetric efficiency
Qc = actual delivered capacity, gpm
Qt = theoretical capacity, gpm (provided by the vendor after a given
model is specified)
699 Pump Sizing Spreadsheet for Twin Screw MPPs

An example of the Excel twin screw pump sizing spreadsheet is shown in
Figure 600-43.
The following example problem is included on the Figure 600-43 spreadsheet to
demonstrate the use of this tool. As mentioned in the example spreadsheet,
information that has to be input into the spread sheet is in the gray shaded area (red
font color), whereas the calculated quantities are blue and yellow font colors.
The spreadsheet calculates:
• Water rate (bwpd)
• Total liquid (blpd)
• Formation or produced gas rate (mscfd)
• Z (the gas compressibility factor of approximately 1 for suction pressures less
than 100 psig)
• Total gas rate (mscfd)
• Delta P (psi)
• Gas rate (actual bpd)
• Gas rate (acfm)
• Total pump throughput (abpd, gas, and liquid)
• Total pump throughput (gpm, gas, and liquid)
• Gas fraction (%)
• Theoretical liquid HP
• Theoretical gas HP
• Theoretical total HP
• The recycle, slip, and flashing factor
• Equivalent pump capacity (gpm and bpd, including the recycle, slip, and
flashing factor)
• Required bhp
• Mechanical efficiency
• Power (kW)
Fig. 600-43 Example of a Multiphase Pump Design Sizing Spreadsheet
March 2009
Pump Manual
Input data into the gray shaded fields. The spreadsheet calculates the blue and yellow colored numbers. Project Name
Yellow shaded columns are calculated capacity and power required to meet given case flows and Engineer
conditions. Date
Gas
Oil Water Water Total Form. Lift Com. Total Pump Suction Pump Delta
API Rate Cut Rate Liq. Gas Gas Fact. Gas Suction Temp. Dsch. F
Case Gravity (bopd) (%) (bwpd) (blpd) GOR (mscfd) (mscfd) (Z) (mscfd) (psig) (°F) (psig) (psi)
1 24 1,000 25 330 1,330 400 400 100 0.98 500 50 150 350 300
2 24 1,000 25 330 1,330 400 400 0 0.98 400 50 150 350 300
3
4
5
6
7
Recycle,
Total Total Slip, Equivalent
Case Gas Gas Pump Pump Gas Theoretical Theoretical Theoretical Flashing Pump Req Mech Power
Rate Rate Thruput* Thruput* Fraction Factor Capacity
abpd, gpm, ***
abpd* acfm* gas & gas & liquid gas hp** total hp (%) GPM BPD BHP Effcy kW
liq liq hp**
1 23,360 91 24,690 720 95% 5 45 50 15% 830 28,390 160 31% 120
2 18,680 73 20,010 580 93% 5 35 40 15% 670 23,010 130 31% 95
3
4
5
6
7
*Calculation is based on the pump suction conditions listed. Contacts: Bob Heyl
**Calculation is based on the fluid theoretical horsepower to pump the liquid or gas fraction only.

***Leistritz believes the bhp is the same as pure liquid when pumping a mixture of gas and liquid. Updated:
600-89
The recycle, slip, and flashing factor was developed from field trials. It is used to
increase the total pump throughput to account for slip in the pump and gas flashing
out of the solution from the recycle stream. Minimum recommended value for this
factor is 15 percent. The spreadsheet uses the following guidelines for calculating
the factor to be used:
• 15 percent for API gravities less than or equal to 25 and having any GVF
percent;
• 25 percent for API gravities greater than 25 and GVFs less than 50 percent;
• 33 percent for API gravities greater than 25 and GVFs greater than 50 percent.
For example, the background and need for the recycle, slip, and flashing factor is
explained thoroughly in Section 6134 Lessons Learned, item J.
6910 Piping and Instrumentation Diagram (P&ID) for a Twin Screw MPP
Early in the decision process, a “draft” P&ID of the MPP and its associated
equipment should be developed. Figure 600-44 shows a typical diagram.
The type and number of alarms, shutdowns, and various devices required for a twin
screw MPP installation should follow the requirements of appropriate Chevron and
industry specifications. Each installation is somewhat different, and each should be
reviewed to determine the proper devices. The diagram in Figure 600-44 is
considered a typical representation of such devices for a typical electric motor
driven MPP installation:
Fig. 600-44 Piping and Instrumentation Diagram (P&ID) of an MPP Installation (Courtesy of Leistritz Corporation)
The various devices shown in the P&ID are:

1. Suction strainer or filter: The dual strainer or filter installed in the suction line
can be switched from one basket to the other while continuing to run the MPP.
A high differential pressure alarm and shutdown are required. The shutdown is
required to be set lower than the collapse pressure of the basket strainer/filter.
Each basket’s surface area in the dual strainer is required to have an area
150 percent of the inlet pipe area. A mesh size of 1/8 inch square is
recommended. If wax is present, a dual filter should be used instead of the dual
strainer with each basket’s filter having an area 200 percent of the inlet pipe
area. (Refer to Section 6135 Lessons Learned, item B and Section 6136
Lessons Learned, item F.)
2. Instruments on the suction line: A high temperature and a high/low pressure
alarm should be installed in the suction line, immediately ahead of the pump.
3. MPP instruments: The MPP should have an RTD or thermocouple touching the
outer race of each antifriction bearing. It should be connected to a monitor to
alarm at 200°F (93.3°C). As an alternate, a RTD or thermocouple could be
installed in the lube oil reservoir set to alarm at 180°F (82.2°C). Refer to
Lubrication in Section 622. Additionally, an accelerometer should be installed
on the case, over each bearing, to be connected to a vibration monitor.
4. Pressure safety relief valve: A pressure safety relief valve is required on the
pump discharge line ahead of any valve to avoid overpressuring the positive
displacement MPP and the discharge piping. The safety valve should relieve
back to suction entering the suction line upstream of the alarms/shutdowns
described in item 2. Chevron prefers the safety valve relief to be routed to a
suction tank or as far upstream as possible. A separate safety valve located on
the discharge piping is required, as opposed to a safety valve built into the
pump.
5. Instruments on the discharge line: A high temperature alarm and a high
pressure alarm/shutdown on the pump discharge are required upstream of any
block valve.
6. Liquid knockout boot or vessel on the discharge: If the GVF is ever expected to
be above 95 percent, a liquid knockout boot or vessel is recommended. It is to
be located on the MPP discharge piping upstream of any block valve. It is
required to supply screw sealant and seal flush fluid to the MPP. Note that, if
slugging conditions are expected, the GVF will exceed 95 percent, and the
knockout boot or vessel is required. The knockout boot can be a wide spot in
the line, or it can be a separate vessel. Either one should include a high/low
level gage with a low level alarm. (Refer to Section 633 and Section 635.)
7. Instruments on the screw sealant/seal flush line: A flow indicator with a low
flow alarm on the screw sealant/seal flush line from the knockout boot to the
MPP inlet. If the crude oil is light, more than 30 degrees API, a cooler (not
shown) is recommended to minimize flashing inside the MPP. Flashing will
occupy much of the MPP capacity and will limit the amount of production from
the field. This cooler is not shown in Figure 600-44. (Refer to Section 6134
Lessons Learned, item K.)
8. Bypass/recycle line: Piping should be installed between the suction of the MPP
and its discharge to allow the multiphase fluid to bypass the pump before
startup to relieve the wellhead shut-in pressure. It may also be necessary to use
it during startup of the system to control the amount of fluid the MPP is
pumping through the entire discharge line, reducing the discharge backpressure
that the pump is pushing against. This bypass line should be blocked during
normal operation.
9. Electric motor instrumentation: RTDs are required for the motor stator
windings, two per phase. Additionally, the motor should have vibration sensors,
either proximity probes for sleeve bearings or an accelerometer on the motor
bearing cases for antifriction bearings. The sensors should be connected to a
vibration monitor. Finally, an RTD or thermocouple should be mounted on the
outer shell of each sleeve bearing or touching the outer race of each antifriction
bearing. The vibration sensor and temperature indicators for the bearing are not
shown in Figure 600-44.
10. ASDs: ASDs are highly recommended with an MPP electric motor driver to
allow the MPP to vary its speed. A separate computer with a variety of alarms
and indicators is deployed to monitor the ASD and the motor driver.
11. Seal drain tank: A seal drain tank with associated piping is shown in the P&ID
to collect mechanical seal leakage. It should have a high level alarm.
6911 Separator Sizing for Twin Screw MPPs

The following is a guide for sizing the downstream separator:
• If the GVF is less than 95 percent and no slugging is expected, the externally
supplied screw sealant is not needed, and a downstream separator is not neces-
sary. (Note that, if no slugging is expected, but an API Seal Flush Plan 32 has
been chosen, an external source [a source other than the MPP pumped stream]
of liquid, such as water or oil is required.)
• If the GVF will exceed 95 percent, the screw sealant needed is 4 percent of the
pump capacity using the following formula:
Amount of screw sealant (bpd) = 4 percent of the MPP capacity (bpd)
(as calculated in Section 693.)
If the pump’s final capacity is 23,011 bpd, the screw sealant would be 4 percent of
that or 920 bpd.
If the MPP suction line is simulated using Olga and Olga indicates that a gas slug
will occur for 1 hour, the downstream separator’s capacity would be:
Separator capacity = 950 bpd/24 hr per day,
or approximately 38 bbl/hr (1,600 gal/hr or 27 gal/min).
A 38 bbl storage volume would be a vessel approximately 6 feet in diameter by
8 feet in height. This is a fairly large vessel, but its size can be decreased if the MPP
is moved closer to the well(s). If this closer location is practical and another
simulation is performed, Olga will indicate a reduced possibility for slugs, and the
slugs will be of a shorter duration. This, of course, allows the downstream vessel to
be reduced in size.
6100 Typical MPP Application Opportunities
6101 General
There are many opportunities for MPP applications in upstream and a few in
downstream.
In upstream, an MPP is often compared to a conventional system that consists of a
production separator vessel, transfer pumps, a gas compressor, possibly a test
separator, associated metering/instrumentation, and usually two pipelines—one for
the gas and another for liquid.
By contrast, an MPP installation is simpler, consisting of the MPP itself and a single
production pipeline to transfer both gas and liquid. MPPs boost pressure without the
need to separate, costing less to purchase and install, plus saving space and weight.
6102 Upstream Application Opportunities—Listing

The following is a list of the opportunities for upstream applications. Each
opportunity is detailed further after this listing:
1. Reducing the back pressure on the well(s) or taking suction from an oil well(s)
to increase the production and extend the life of the well(s);
2. Replacing old, worn out conventional facilities;
3. Bringing abandoned wells back to life;
4. Producing marginal fields;
5. Transporting production several miles through one pipeline;
6. Producing subsea reservoirs with a more cost effective installation and ease of
retrieval;
7. Overcoming the static pressure from a deep subsea well;
8. Reducing emissions;
9. Producing into a high pressure header.
Advantage
MPPs also have a particular advantage in many applications—saving weight and
space compared to a conventional system, especially on a platform.
6103 Upstream Application Opportunities—Details

The following describes in more detail each of the MPP upstream application
opportunities:
1. Taking suction from an oil well(s) to increase production and extend the life of
the well(s);
The value of increased production is used to pay out the MPP installation.
An MPP can be installed to take suction from one or more oil wells. This is
probably the most common of all MPP applications. The MPP reduces the
wellhead backpressure.
In addition, an MPP can be economical when it takes suction from the
discharge of almost any type of artificial lift, including gas lift, electric
submersible pumps (ESPs), progressing cavity pumps (PCPs), and rod pumps
(where the flow line is tied to the casing). (An MPP taking suction from a
well(s) that has a downhole progressive cavity pump [PCP] installed will not
usually increase production.)
– Gas Lift
Gas lift involves natural gas injected into the tubing string at intervals from
the bottom to the top. Gas bubbles decrease the fluid density, which helps
move the fluid to the surface. Installing an MPP at the surface will increase
production, because it essentially lowers the bottom hole flowing pressure
by the same amount as the wellhead backpressure. An MPP could also
reduce the amount of gas lift required, saving energy costs.
The quantity of gas involved in gas lift is often considerable. The MPP has
to process this gas, along with any increase in production from the well.
This often requires the MPP to be large and, perhaps, uneconomical. As
stated many times, each installation must be evaluated on its own merits.
– Electric Submersible Pumps (ESPs)
The capacity of an ESP depends on its speed and the diameter of its
impellers. It follows a normal centrifugal pump performance curve of
differential pressure (head) versus capacity. Production could be improved
by replacing the ESP with a larger capacity ESP. The replacement pump
would have to be one that operates at a higher speed or has larger
impellers. Increasing its speed may not be feasible or, at least, expensive.
Increasing the impeller size is often not feasible, because the impeller
diameter is limited by the casing. An MPP at the surface reducing the
discharge pressure of the ESP will decrease the ESP differential pressure.
This moves the ESP operating point on its performance curve and
increases capacity.
– Rod Pumps
A rod pump consists of a motor, speed reducer, and “chicken head” on the
surface next to the wellhead. The “chicken head” moves a downhole
positive displacement pump up and down at the bottom of a long rod,
usually 1,000 feet or so long. Since the pump is a positive displacement
type, theoretically its capacity remains constant for any discharge pressure.
For this statement to be entirely true, slip has to be ignored.
The downhole pump is connected to the well tubing string. The casing
surrounds the tubing. Crude oil, water, etc., enters the casing annulus
through perforations. Fluid in the casing annulus feeds the downhole
pump. The pump can be sped up to produce more, if the level remains high
enough. If the level gets too low, the rod pump is shut down or slowed
down to wait for more fluid to enter the casing.
If one can lower the pressure in the casing annulus, the level will rise and
the rod pump can be sped up to produce more. This is often accomplished
by installing a casing gas collection system. A casing gas collection system
consists of a vacuum pump connected to a piping network that is connected
to several well casings at the surface. It keeps the casing pressure low and,
thus, the downhole casing level as high as possible. An MPP connected
immediately downstream of a rod pump with this type of system will not
appreciably increase production. It merely lowers the discharge pressure of
the downhole pump without affecting the level in the casing.
Occasionally, if rod pumps are spread out with a considerable distance
between them, a casing gas collection system is not installed, and the
casing at the surface is connected to the flow line. Here, the casing
pressure is fairly high, and the downhole level in the casing is substantially
lower. In this case, an MPP taking suction from the discharge of a rod
pump will increase production, because the casing gas pressure is reduced
along with the rod pump discharge pressure. Note, however, to get more
production out of the well, the rod pump has to have the capability to
increase its speed.
The logic that an MPP will not increase production if the rod pump has a
casing collection system is not entirely correct. Production will increase a
little because the MPP lowers the downhole pump discharge pressure.
These pumps wear out and are operated almost until complete failure.
During a large part of their operating life, they circulate much of their
capacity back to suction (“slip”). Slip increases as they wear. An MPP
installed at the surface lowers the discharge pressure of these pumps,
which in turn, increases their capacity by decreasing slip. The pump life is
thereby increased. Connecting an MPP to several wells like this extends
the Mean Time Between Failures (MTBFs) of these pumps, saving a
considerable amount of maintenance cost and down time.
2. Replacing old, worn out conventional facilities;

A cost effective MPP can replace old, worn out conventional facilities and
often keep a marginal field operating longer. Its cost is almost always less than
replacing the conventional facility. For example, several Bornemann MPPs
were installed in older German oil fields taking suction from wells in different
fields. The MPPs increased the production, extending the life of the fields.
Without the MPP, the oil field would have been shut in.
3. Bringing abandoned wells back to life;
MPPs have brought back to life wells that were shut in, because they were
considered too uneconomical due to the amount of gas and water produced. In
these cases, the cost of the conventional system was too expensive to make
reopening the wells profitable. The lower cost MPP made these wells profitable
again.
4. Production from marginal fields;
A common application for an MPP is to move multiphase fluid from a remote,
marginal, or satellite oil field to a processing facility several miles away.
Without the MPP pressure boost, the wellhead pressure is often not enough to
transport enough fluid to make the remote marginal well profitable.
A more expensive conventional system could do this, as well. However, the
MPP is less expensive and requires only one pipeline, whereas the conventional
system usually requires two.
a. In California, the Midway-Sunset field needed to increase production from
some remote wells. A twin screw pump was installed, and the production
increased dramatically. Refer to the pump installation in Figure 600-45 and
the strip chart results in Figure 600-46.
As can be seen in Figure 600-46, the total oil and water production
increase due to the installation of the MPP was 1,750 bpd, that is, from
1,000 bpd to 2,750 bpd. The wells involved were located in a remote
location and by installing the MPP, the back pressure on the wells was
reduced significantly, causing the increase in production. Note that, when a
second pump was installed, another increase of 650 bpd was seen (from
2,750 bpd to 3,400 bpd). The production increase was so cost effective that
Midway-Sunset now has four twin screw MPPs in operation. (Refer to
Section 6138.)
b. In Duri during the field testing of the helico-axial MPP versus the twin
screw MPP, the production increased from 10,900 bopd to 12,500 bopd
when the twin screw pump was put online. When the helico-axial MPP
was put online, the increase, though smaller, was still significant,
increasing from 10,900 bopd to 11,900 bopd. (Refer to Figure 600-47 and
Figure 600-48.)
Fig. 600-45 Bornemann Twin Screw Pump at Midway-Sunset
Fig. 600-46 Strip Chart Showing the Increase in Production at Midway-Sunset
Fig. 600-47 Duri, Indonesia, MPP Field Testing
Fig. 600-48 Strip Chart Showing the Increased Oil Production with MPPs during Field Testing in Duri, Indonesia
5. Transporting production several miles through one pipeline;

An MPP can pump fluid to an existing processing facility several miles away,
through a single pipeline. This is attractive compared to a conventional facility,
which would require more equipment to perform the same task, including a
possible second pipeline. For example, the Mitsue pump, mentioned several
times in this manual, was moved to the Princess Oil Field in Canada, where it
moved multiphase fluid from several wells to the processing facility, 18 miles
away, through one pipeline. (Refer to Section 6135 for further details.)
The capability to pump mixed fluids through miles of pipeline can also be
important for a new offshore platform, eliminating the need for a second
pipeline to shore.
Another good application for an MPP is to boost pressure from subsea wells
that are far from an existing platform. Often, preliminary analysis shows that
the recoverable reserves for the new field are too small to justify a stand-alone
full production platform that is closer to the wells. A subsea MPP has the
capability to eliminate the need for a new platform.
6. More cost effective subsea installation and ease of retrieval;
For a subsea installation, space is valuable, and an MPP requires a much
smaller, simplified seabed support structure than a conventional system. Lifting
an MPP pump from the seabed for maintenance is easier and less expensive
than a conventional pump and compressor, because it weighs less than their
combined weight.
7. Overcoming the static pressure in a deep subsea well;
Deep subsea oil wells often have their production limited because of the static
pressure needed to transport the fluid from the wellhead to the surface. This
becomes worse as the well depletes, and its wellhead pressure decreases. If
located a mile below the surface, the pressure required to move fluid to the
surface would be approximately 2,300 psi. A subsea MPP, taking suction from
such a wellhead, can increase the fluid pressure significantly to get the fluid to
the surface, reducing the wellhead backpressure and greatly increasing
production.
8. Reducing emissions;
MPPs reduce pollution by eliminating the emissions from the many pieces of
equipment needed for a conventional system. For example, in California, the
Midway-Sunset field needed to add a few new wells, but the permitting process
for the vent for the atmospheric collection tank had stalled the field’s
production. An MPP, which required no emission permit, was installed to pump
the gas and liquids to the existing permitted collection facilities several miles
away.
9. Producing into a high pressure header.

Occasionally, a well’s pressure is not sufficient to transfer its production into a
distribution header supplied by higher pressure wells. An MPP, taking suction
on the well and boosting its pressure, allows the lower pressure well(s) to
produce into the higher pressure line, thus increasing the field’s total
production. Refer to the Section 6125 example for Main Pass 59A, Gulf of
Mexico (2007).
Advantage
MPPs also have the following particular advantage, which is extremely useful in
many applications:
Reducing offshore platform costs by reducing weight and space. An MPP weighs
less and involves a smaller footprint than a conventional system. This is important
for an offshore platform, where the additional weight and space would be costlier if
a conventional separation system were installed. An MPP occupies approximately
25 percent of the area and weighs approximately 25 percent as much compared to a
conventional system.
6104 Downstream Applications

There are several downstream applications that are important to consider:
1. Flare gas knockout drum pumps.
MPPs have successfully replaced flare gas KO drum pumps in refining. Since
the pumps can handle gas, water, particulate, and oil, the pumps are well suited
for this application. They are particularly more reliable than the old
reciprocating pumps often used in these situations. The MPPs also use
mechanical seals instead of the packing required by the recips, thus reducing
emissions. In new installations, the old required flare gas pits are no longer
required, since unlike the recips that required the pit to provide their high
NPSHr, the twin screw MPPs require little or no NPSHr. Since the KO pits can
contain and confine the often hazardous vapors leaked by the recip packing,
locating the pump at ground level where vapors are not confined, allows for a
much less expensive installation, providing a safer work place, as well.
2. Pumping high vapor containing streams or boiling hydrocarbon streams.
Since twin screw MPPs have little or no requirement for NPSHr, they can
handle these streams very well. Twin screw MPPs are especially well suited in
this regard and will soon be covered by API 676 standards when the 3rd edition
of that standard is published (expected 2009).
6110 Technology Qualification Process (TQP)
6111 Introduction
MPP technology is being used in applications with ever more challenging
environments and process conditions. To determine the risk involved in using MPPs
in such applications, the pumps must be evaluated for operability and reliability.
Chevron has developed a TQP to assist in making this determination. Chevron’s
system is a process initially developed by Det Norske Veritas (DNV).
6112 Chevron TQP

The following is taken from the Chevron TQP home page.
TQP is a versatile and scalable process to formally assess new (and existing)
technology and make high quality, risk based decisions regarding the technology’s
further development or deployment and utilization on projects. TQP also provides a
common language for the understanding and communication of a technology’s
“readiness for use”. The logic diagram in Figure 600-49 gives an overview of the
major process elements.
Fig. 600-49 Logic Diagram
“… We will have demonstrated leadership in deploying technology to create

value across the company, both high impact and where multiple opportunities
for application create high impact. It will be measurable and widely recognized.
Leaders will put a priority on enterprise behavior so that we can take full
advantage of the intellectual capital of an integrated company, quickly
deploying best practices and technology.” - Dave O’Reilly, The Next Five
Years.
The link to the home page of TQP is provided for further reference:
http://etc.chevron.com/team-tqp/default.asp
Contact Chris Chaplin of the TQP team for additional information:
[email protected].
6113 Definition of Qualification

Qualification is the process of providing the evidence that a technology will
function within specific operational limits with a specified level of confidence.
6114 Machinery TQP

Mark Weatherwax of the ETC Machinery and Power Systems team has developed a
general overlay for the general TQP discussed above. The overlay is intended to be
used as a guideline for evaluating the specific risks involved with a specific set of
project conditions. The following logic diagrams, Figure 600-50 and Figure 600-51,
taken from the overlay, are logic diagrams of how the process should be used to
evaluate risks associated with machinery technologies.
At the time this MPP manual is being published, the machinery TQP is in
development.
Figure 600-52 is an example of the use of this machinery TQP in flowsheet format.
The example is for a General Electric LM6000PF gas turbine mechanical drive
application.
For more information contact Mark Weatherwax by email at
[email protected].
6115 Technology Development Stages (TDSs) for MPPs

MPP technology is composed of four different pump types: twin screw, helico-axial,
progressing cavity pumps (PCPs), and electric submersible pumps (ESPs). Each
pump type is developing at its own rate, and each has expanding application and
performance envelopes, where pump manufacturers are increasing current limits on
capacity, differential pressure, gas volume, reliability, and operating environment
(onshore, on platforms, or subsea). Each has a history of proven experience onshore
and in association with platform installations. Each pump type design and current
limitations have been covered in this manual, but regarding subsea in particular,
each pump type is developing rapidly.
Fig. 600-50 Logic Diagram—Machinery Technology Qualification Process (Part A)
Typically the potential equipment selection is a specific supplier model or train configuration.
Other cases may include a process specification and development of equipment to meet the
Potential
process conditions. In either condition, the process operating conditions and any unusual
Equipment
ambient conditions should be identified during the qualification process.
Selection
Used Within All Sub

In Chevron YES Existing YES Components YES
Fleet? Chevron Within
Operating Experience?
Parameters?
NO Do Lessons
NO Learned
Exist?
NO
Are Operating
Used Within YES Parameters
Industry? Within Existing YES
Industry Identify
Operation? Individual
Components
That Are
NO Step-outs
(TC4)
YES
YES
Review of Industry Are Lessons
Experience NO Learned
NO
Incorporated?
Is Industry
Experience YES
Considered
Acceptable?
All Sub
Components
Within YES
Experience?
Document
NO References -
Classified as a
(TC3) YES Document TC1
References -
(TC2) Classified as a
NO TC2
Do Lessons
Learned Exist?
YES
NO
Are Lessons
YES Learned
NO
Incorporated?
(TC4)
NO
A Threat
Assessment is
Required.
Fig. 600-51 Logic Diagram—Machinery Technology Qualification Process (Part B)
The level of effort for the Technology Qualification Process should be based on magnitude of
the step-outs being considered. The attendees at the Risk Assessment Workshop should also
be determined based on the magnitude of the step-outs.
Potential outcomes from the Risk Assessment Workshops include determining that the risks
exceed the potential benefits and therefore abandon the qualification process.
1. Identify individual risks and create an overall Threat Register.

2. Create a Threat Risk Table for each component identified.
Risk Assessment 3. Create action items to address individual risk items.
Workshop 4. Identify alternate solutions or risk mitigation options.
5. Use a Risk Matrix to rank each risk.
6. Agree on Acceptance Criteria for identified actions.
Note: Equipment risks can range from operational reliability and equipment failures to
schedule risks from testing or manufacturing delays.
NO
Technology This plan is a detailed document that addresses the risks and action items identified during the
Qualification Plan Risk Assessment Workshop. The plan will become the basis for performing the Technology
Qualification.
Are the risks fully

addressed by the
actions identified?
During this phase all the activities outlined during the Qualification Plan will be addressed. For
equipment packages, this typically entails the following types of activities:
rd
NO YES
1. Analytical analysis by Suppliers or 3 Parties (Torsional, Lateral Rotordynamics, Stress
Calcs, CFD, FEA, etc.)
2. Model testing of components.
3. Full scale testing of components.
4. Etc.
Technology
Qualification Individual activities may be completed at various stages in a project depending on the
Execution associated risks.
Perform a GAP
Analysis. Was
the
Performance
Criteria
Satisfied?
YES
Document Qualification
Process
Fig. 600-52 General Electric LM6000PF Gas Turbine Mechanical Drive Application
Performing a rigorous TQP evaluation to develop a TDS or Maturity Rating for a

specific project requires the engineer to have a specific set of operating conditions.
Since, in this manual, there is no specific application to judge against, it is not
possible to provide definitive numbers or ratings. In each specific MPP application,
a TDS assessment should be made of the specific pump type(s) to be considered. If
the stage number assessment indicates the development stage is lower than TDS 9, a
TQP evaluation should be initiated.
The general TDS definitions, associated with the TDS numbers, are provided in
Figure 600-53.
Fig. 600-53 Technology Development Stage (TDS) Definitions
Onshore and Platform Installations

As has been described in earlier sections of this manual, all four MPP pump types
are well established, as represented by multiple surface installations both onshore
and on platforms, and the technology can be assigned a TDS number of TDS 9
(field proven).
The current limits of MPP technology, proven in project applications, were
presented (Section 650, Section 660, Section 670, and Section 680), and certain
aspects of the commercially available MPP technologies were compared. (Refer to
Section 621, Figure 600-7 and Figure 600-8.)
Subsea Applications
Subsea applications are developing very rapidly for all the MPP designs. As was
stated in the manual, these applications are in various stages of development and are
for various environments and operating conditions. While ESPs in dry tree
applications (pump is downhole subsea, but the driver is on a platform deck) have a
significant history, those applications where the pump and driver are all on the
seabed floor are in a much lesser stage of development (TDS 5). Wet tree helico-
axial pumps also have a reasonable history (TDS 9), while wet tree twin screw
pumps to date have a very limited number of subsea applications (TDS 9). Wet tree
applications of PCP MPPs are still being developed (TDS 5).
Table of TDS Ratings for Various Pumps for Various Applications

Figure 600-54 shows in tabular form some general TDS number guidelines and
important qualification limits for the various MPP designs. Note that these
guidelines are based on each technology’s development at the time of publication.
Fig. 600-54 TDS Ratings for Various MPP Pumps for Various Applications
Pump Type Application TDS Number Important Limitation(s)
Twin screw Onshore TDS 9 Differential pressure limited to 1,400 psi, flowrates
limited to 370,000 bpd, GVF unlimited.
Platform TDS 9 Differential pressure limited to 1,400 psi, flowrates
limited to 370,000 bpd, GVF unlimited.
Subsea Seabed TDS 9 Differential pressure limited to 750 psi, flowrates
limited to 170,000 bpd, GVF unlimited, water depth
limited to 5,500 ft.
Helico-axial Onshore TDS 9 Differential pressure limited to 1,200 psi, flowrates
limited to 290,000 bpd (Framo), 560,000 (Sulzer) GVF
unlimited.
limited to 290,000 bpd (Framo), 560,000 (Sulzer) GVF
unlimited.
limited to 290,000 bpd, GVF limited to 90%, water
depth limited to 3,000 ft.
PCPs Onshore TDS 9 Differential pressure limited to 900 psi, flowrates
limited to 60,000 bpd, GVF limited to 40%.
Platform TDS 9 Differential pressure limited to 900 psi, flowrates
Subsea Seabed TDS 5 Considerable development required before commercial
application in a subsea environment.
ESPs Onshore TDS 9 Differential pressure limited to 4,000 psi, Flowrates
Subsea Seabed TDS 9 Dry tree applications: differential pressure limited to
4,000 psi, flowrates limited to 10,000 bpd, GVF limited
to 60%, water depth limited by dry tree applications.
Population Chart
The population chart in Figure 600-55 is provided to show the current common
usage range of the various types of MPPs.
Fig. 600-55 Population of Various MPP Types and Their Operating Regions
6116 TQP Summary

In general, all of the MPP technologies can be considered as proven for applications
involving surface onshore and surface platform installation. It is only in subsea
seabed applications that some of the technologies are currently considered less than
proven, and limitations are still significant. Having said this, it is still important to
remember that, even though a technology is proven, it doesn’t necessarily mean that
the technology has a high level of reliability for a specific project application. Each
pump type must be evaluated for the specific environment and range of operating
conditions, and its reliability must be estimated in order to make the proper
selection that provides the lowest cost of ownership for the application.
6120 Economics
6121 General
In the past, Chevron has often avoided MPPs because the risks associated with a
“new application” were considered too great to offset the benefits. The decision
makers often installed conventional complex and costly facilities instead, consisting
of a separator, compressor, pump, etc., as described more fully in Section 614. This
was done even though the MPP alternative was more economically attractive. Such
concerns are no longer valid. As shown in Section 6110, MPPs should no longer be
considered a “new product”. Hundreds of MPPs now function reliably throughout

the world, operating with great success.
6122 Existing Facilities

Options
Often Chevron is faced with existing facilities that require extensive maintenance
and repair. In such cases, Chevron needs to consider no longer replacing these
facilities in kind but to install an MPP instead. There is a basic process that can be
used to help the decision. The process involves comparing the economics of the
following options:
• Continuing with maintenance and repairs;
• Installing direct replacement conventional facilities;
• Installing an MPP.
Calculations Associated with the Options

1. Continuing with maintenance and repairs;
List the value of the current revenue, expenses, and profits for the gas and oil
produced or recovered and the cost to maintain and repair. Prepare a balance
sheet of annual income versus expenses, and determine if it is economical to
continue in this mode of operation.
2. Installing direct replacement conventional facilities;
List the benefits of installing a conventional facility that includes the cash
generated. Also, list its capital cost and operating expense. Compare this to
option 1 to obtain a rate of return, payout period, and NPV for the conventional
system.
3. Installing an MPP.
List the benefits of installing an MPP, including the cash generated. Also list its
capital cost and operating expense. Compare this to “No Improvement” case to
obtain a rate of return, payout period, and NPV for the MPP. Compare this to
the Conventional Facility case to decide which is the most attractive.
The alternative that has the best economic outcome, taking into account risk, is the
one that should be chosen.
6123 New Facilities

For a new facility, compare the MPP case to the conventional facility case.
6124 Factors to Consider

The following factors need to be considered in determining the economic payout of
an MPP installation:
• MPP cost, equipment weight, foundations, possible support systems, spare
parts inventory, and space—A paper study conducted several years ago
confirmed that an installed MPP is less expensive than a conventional system,
in that the MPP costs approximately 70 percent less, occupies approximately
25 percent less area, and weighs approximately 25 percent fewer pounds. The
MPP has the following advantages:
– No separation vessel (twin screw MPP);
– No compressor;
– Fewer instruments;
– One foundation;
– One baseplate grouted (if at an onshore facility) into place or one skid
installation (if on a platform or subsea facility);
– One electrical power supply line, one electrical connection, one starter, and
other electrical components;
– Required spare parts for the MPP versus those required for the conven-
tional system’s pump and compressor;
– One multiphase mixed flow discharge line.
• MPP energy consumption—If compared to a conventional facility, an MPP will
likely cost less, but it may require more energy to operate. If power costs are
high, such as at some onshore locations, and the application has a high GVF, a
conventional system may be more economical. Offshore (especially if gas has
no value and is intended to be flared), the MPP is likely to be more cost effec-
tive.
• Productivity index (PI)—A PI is used to measure the ability of the well to
produce, and it is defined as the ratio of the total liquid rate to the pressure
drawdown. The PI should be a valid measure of the well productivity potential
only if the well is flowing under pseudo-steady state conditions. During a tran-
sient period, the productivity index will vary, depending on the measurement of
the flowing bottom hole pressure (Pwf).
PI is given by the following expression:
PI = Q/(Pr  Pwf)
(Eq. 600-7)
where:
Q = Total fluid rate, stb/day
PI = Productivity Index, stb/day/psi
Pr = Reservoir static pressure, psi
Pwf = Flowing bottom hole pressure, psi
• For example, if the bottom hole flowing pressure can be reduced by 50 psi and
a well has a PI of 0.1, a production increase would be expected to be 5 bpd
(0.1 x 50 = 5).
When considering a new MPP to reduce a well’s backpressure, one must
assume the bottom hole pressure is reduced by the same amount as the back
pressure. This is not always true. However, most of the time, the assumption is
accurate enough.
When applying PI, make sure its definition is understood. Some people define
PI as an increase in crude oil production, while others define it as the total
liquid produced, meaning that the PI could include both crude oil and water.
• Pressure reduction limit—An MPP may become large if the existing wellhead
pressure is low. Gas occupies a very large volume at a low pressure. Therefore,
to reduce the wellhead pressure further, the MPP has to be very large and
quickly becomes uneconomical. This can better be explained in the following
example:
The Mitsue pump handled 25,000 bpd of gas at 50 psig suction pressure and
10,000 bpd of liquid. If the pump were designed for a suction pressure of
0 psig, the gas volume would rise to 110,000 bpd, while the liquid rate would
stay essentially the same. The pump would have to pump approximately
120,000 bpd or almost 5 times the original capacity. Under these conditions, a
potential MPP may be too large to pay out from the expected production
increase.
Many oil fields, especially older ones, have maximized their production by
reducing the wellhead pressure. It is not uncommon to see a 10 psig or 20 psig
wellhead pressure. If one wishes to install an MPP under these conditions, the
size may be too large to be economical. Though this is generally true, there are
exceptions, and each installation needs to evaluate its own economics.
Note that most subsea wells have a high wellhead pressure because they need to
overcome the static pressure to get the fluid to the ocean surface. Because the
initial wellhead pressure is high, an MPP could reduce it significantly, making a
huge impact on crude oil production, all with a reasonably sized pump.
6125 Examples
Both of the Chevron MPP field tests (Main Pass 313 offshore platform and the
Mitsue Field in Alberta, Canada) completed in early 1996 decreased wellhead
backpressure and increased crude oil production. The economics for these tests are
described as follows:
Main Pass 313 Offshore Platform (1993)

Extended well performance information was collected on the best performing well
on the platform, well A-23D. The well showed a boost in production from 70 bopd
before the MPP to more than 600 bopd with the pump. Later, the pump maintained
an average increase of approximately 200 bpd.
The installed cost of the MPP was approximately $150,000 (in 1993 dollars). The
economics on an annualized basis are listed in Figure 600-56. It is assumed that well
A-23D was always lined up to the pump with an operating factor of 0.9 (which
means that the pump is operational 90 percent of the time). The economic
calculations are shown in Figure 600-56.
Fig. 600-56 Main Pass 313 Economics

Revenue/Cost Calculation $/Yr
Oil Revenue (200 bpod) x (330 days/yr) x ($15.00/bbl oil) $990,000
Gas Revenue (200 bpod) x (1000 scf/bbl oil) x (330 days/yr) x ($0.12/bbl fluid) 132,000
Liquid Treating Costs (200 bpod) x (1 bbl fluid/0.6 bbl oil) x (330 days/yr) x ($0.12/bbl (13,200)
fluid)
Energy Consumption (56kw) x (2730 btu/hr/kw) x (24 hr/day) x (1/scf/ 1000 btu) x (330 (8,100)
Costs days/yr) x ($0.06/mscf)
Added Gas Compression (200 bpod) x (1000 scf/bbl oil) x (330 days/yr) x ($0.06/mscf) (4,000)
Cost
Net Revenue $1.097 MM/yr
After Tax Profit (Takes into account royalties, operating expenses, and $570M/yr
depreciation.)
NPV $2.8 MM
Payout 3 Months
Note that the analysis in Figure 600-56 was done in 1996. Assumptions used then
would be quite different today. The assumptions used then were:
• Oil price is $15.00/bbl;
• Operating factor is 0.9;
• Gas price is $2.00/mscf;
• Tax rate is 38 percent;
• Liquid treating expense is $0.12/bbl fluid;
• Royalty rate is 13.0 percent;
• Depreciation life is 7 years;
• Efficiency of gas engine power generators is 30 percent;
• Discount rate is 11.0 percent;
• Inflation rate is 3.0 percent;
• Gas compression cost is $0.06/mscf;
• Project life is 10 years;
• Composite water cut is 40 percent (for calculation of total liquid rate from
known oil rate).
Mitsue Field, Slave Lake, Alberta, Canada (1993)

At approximately the same time as the Main Pass 313 pump was installed, another,
larger MPP was installed in the Mitsue Field, Slave Lake, Alberta, Canada. In this
situation, the economic analysis was based on accelerated production, not new
production. The resulting rate of return, based on Canadian economics, was
11 percent, making it a marginal project. The project went ahead anyway, partially
because it was a field trial to obtain important technical information. Later, this
pump was moved to the Princess Field, also in Alberta, Canada, at a cost of
approximately $100,000 (U.S.). The economics of that installation are not currently
available. However, the project was economical, and a conventional system was not
installed. Besides being more expensive, the lead time for the conventional system
components was much longer, costing considerable production.
Escravos 3B Platform, Nigeria (2004)

During the Escravos 3B Flares Out project in Nigeria, the Malu, Opolo, Ewan, and
Isan platforms were evaluated for MPPs as opposed to a conventional separation
system. Initially, based on economics, it was determined that MPPs would be
installed on all the platforms. Unfortunately, it was only then realized by the project
that the gas lift gas volume would also be handled by the MPP. This significantly
increased the gas volume to 95 percent GVF and resulted in the MPPs becoming
uneconomical on all but the Opolo platform. On this platform, the capital
expenditure savings alone was approximately $6 million. However, operations
wanted to have similar facilities on all the platforms, and the MPP for Opolo was
dropped.
Main Pass 59A, Gulf of Mexico (2007)

GOMBU’s Main Pass 59A platform handled a field consisting of 18 wells, some of
which produced at a higher pressure than others. The higher pressure wells
prevented the lower pressure wells from producing through the tieback pipeline.
When they were able to produce, the low pressure wells produced oil, water, and
gas (95 percent GVF). After researching several options, including larger diameter
lines, other piping configurations, and the cost of a conventional system, the largest
Leistritz MPP manufactured was purchased for $3 million (installed for $5 million).
It is driven by a 1,700 HP natural gas engine to take advantage of the low cost gas
available on the platform. (Refer to Figure 600-57 and Figure 600-58.) The MPP
was chosen over a conventional separation system based on capital expenditure cost
savings and a smaller footprint. Production has increased by 1,000 bpd so far (in
2007) and is expected to increase further as more of the low pressure wells are
brought online. Based on only the increased production seen thus far, the simple
payout is 0.17 year with oil valued at $80/bbl. Refer to Comparison #2 in
Section 6126 for additional information.
Fig. 600-57 Main Pass 59A Leistritz MPP on Test (Courtesy of Leistritz Corporation)
Fig. 600-58 Leistritz MPP Installed on Main Pass 59A
6126 Cost Comparison Study for an exceedingly high GVF of 98 Percent—

MPPs versus a Conventional Separation System
Comparison 1
During the late 1990s, a cost analysis (paper study) was made to determine the
conditions in which MPPs should be used and in which a conventional separation
system should be used. Conclusions from that study are:
1. The capital cost of an MPP facility is typically approximately 70 percent that of
a conventional system. This was the same for all flowrates and pressures.
2. An MPP facility weighs approximately a quarter of a conventional system, and
it occupies approximately a quarter of the space of a conventional system.
3. A conventional system usually uses less energy, especially if the GVF is
exceedingly high. Under these conditions, there could be an economical payout
for a conventional system.
The following operating conditions were used in this study:
• Three flow cases were considered: 40,000 bpd, 80,000 bpd, and 120,000 bpd.
• Three, six, and nine wells were assumed to be involved in the three flow cases,
respectively.
• The test separator and multiphase metering capacity are adequate for one well
at a time.
• Gas composition of the multiphase mix is (GVF) 98 percent by volume at pump
suction conditions. This equates to a GOR of 500.
• Three electrical rates were considered: $.03/kW-hr, $.05/kW-hr, and
$.07/kW-hr.
• All machinery is electrically driven.
• No spare equipment capacity was included, except for the crude transfer pumps
associated with the conventional system. Here, two pumps are included.
• Both systems operate at a suction pressure of 15 psig. The discharge pressure
varies, depending upon the three cases considered for differential pressure:
100 psi, 400 psi, and 800 psi.
• Oil gravity is 30 degrees API.
• Gas gravity is 0.7.
• Wellhead temperature is 100°F (37.8°C).
• Additional deck costs are based on $5,000/st and 0.065 st/sq.ft. Note that a
conventional system occupies more space, and its cost on an offshore platform
would be greater.
• The gas and liquid separated in a conventional system are re-combined into one
discharge pipeline.
To simplify the analysis, a few items were not considered. They are the utility costs
(air, water, etc.), lights, safety equipment, office or shop costs, and transportation.
At the high GVF (98 percent), the energy costs played a significant role and were 25
to 50 percent lower for the conventional separation system, compared to the
multiphase energy costs. The 25 percent figure was associated with the 100 psi
differential case, while the 50 percent figure was associated with the 800 psi case.
Therefore, the higher the differential pressure, the greater the difference in energy
consumption for the MPP. This is due to the decreasing efficiency of the MPP with
the combination of high GVFs and high differential pressures. The higher additional
energy costs associated with a multiphase system can be used to pay out the more
efficient conventional system. This could be a factor for onshore applications in
which energy costs are high, but it may not be a factor for offshore applications in
which the electrical costs are low due to power generated on the platform itself.
Comparison 2
In 2006, a potential MPP application existed on the Main Pass 59A platform in the
Gulf of Mexico. An analysis was performed by Greg Sinclair of GOMBU to
compare the cost of the conventional system versus an MPP. The application
involved taking the discharge of several wells and increasing the pressure of these
wells to match the discharge pressure of other wells served by the platform. The low
pressure wells needed boosting so that they could produce an economic flow, while
using the same tieback line to the coast.
The analysis showed the following results:
1. Several scenarios were evaluated. The MPP solution was the most economic
for the Main Pass application, even when compared to numerous other alternate
solutions.
2. In spite of the 95 percent GVF, the MPP solution had a simple payout of
0.17 year. (Note that, at the time these calculations were performed, all the low
pressure wells had not been brought online yet. Based on only those wells
already brought online, the simple payout was still found to be only 0.17 year.)
3. In the economic analyses performed for platforms, MPPs have the advantage of
less weight and a smaller footprint than conventional separation systems.
Appendix I contains Greg Sinclair’s presentation at the 2007 Mechanical Equipment
Round Table (MERT). His presentation was given after the startup of the Leistritz
twin screw natural gas driven MPP. This pump is the largest twin screw MPP in the
Gulf of Mexico. The second presentation is a shortened version of the first, sticking
to the comparisons and the economic details.
As mentioned several times in this document, each potential installation must be
evaluated economically on its own merits, using assumptions and conditions related
to that specific installation. The sample analysis is good only for general
considerations and should not be considered for anything else.
6130 Selected MPP Installations (including Lessons Learned)

6131 General
As mentioned in Section 2, almost all of Chevron and legacy-Texaco MPP
experience has been with twin screw pumps. Chevron has experience with one
helico-axial pump that was part of a 1 month Indonesian trial in Duri (refer to
Section 6138). Since 1990 and as of the printing of this manual, Chevron and
legacy-Texaco have installed at least 94 MPPs. A list of these installations is
included in Figure 600-1.
The installations presented in this section were carefully monitored by specialists to
learn as much as possible about MPPs. Though many of these installations have
been referred to throughout this manual, this section is specifically devoted to
covering these installations in detail and discussing the “Lessons Learned” from
each application.
6132 Humble’s Fluid Flow Test Facility MPP

This motor driven Leistritz L4MK 82-40 MPP was tested in the legacy-Texaco
Humble Flow Loop MPP test facility from 1992 through 2002 on heavy and light
oil, with different water, sand, and gas volume fractions. The maximum flowrate
was 120 bpd, with a designed maximum of 225 psi differential pressure. The pump
performed admirably, and its success was the driver for other applications
throughout the company.
Lessons Learned
Twin screw MPPs were proven to be flexible in handling different ranges of oil,
water, gas, and particulate.
6133 Trinidad
The Leistritz twin screw pump was installed in 1992 at this facility. The pump was
one of the first twin screw pumps installed on an offshore platform by legacy-
Texaco. The pump took suction from one or more wells that were essentially shut-
in, and the MPP brought them back into production. The MPP successfully pumped
heavy crude oil, water, and gas with a substantial quantity of particulate.
Lessons Learned
MPPs can pump heavy crude oil, water, gas, and particulate at the same time. MPPs
are effective in bringing shut-in wells back into production economically.
6134 Mitsue Field, Slave Lake, Canada (later moved to Princess Field in Canada)
In 1992, a 700 HP, Leistritz L4H twin screw pump was tested for a joint industry
project at the Texaco Humble Fluid Flow facility. The test was sponsored by
Chevron, legacy Texaco, and several other major oil companies. It successfully
demonstrated that the MPP could pump multiphase fluids in the controlled
environment of the test facility.
The next phase was to determine if an MPP could successfully pump multiphase
fluids in an actual oil field environment. Chevron purchased this pump in 1993 and
installed it in the Mitsue oil field (Slave Lake, Alberta, Canada). This MPP was
installed at a pipeline junction for the purpose of increasing production from sev-
eral wells by lowering wellhead backpressure. Figure 600-59 shows a simplified
flow diagram of this arrangement.
Fig. 600-59 Mitsue MPP Field Trial
10,000 FEET
FROM TO
SATELLITE BATTERY
6 WELLS
600,000 SCFD
400 BPD
100 FEET
FR
FROM
SATELLITE
TEST SEPARATOR
2 WELLS
LC
80,000 SCFD
130 BPD
FR
MULTIPHASE
PUMP
CHEVRON MULTIPHASE METER LOOP (CMML)
FIC
RECYCLE
At Mitsue, flow from several wells fed the pump simultaneously. One was free
flowing, one used an ESP, and the remainder used rod pumps. As shown in the
figure, most of the production came from wells approximately 2 miles upstream of
the MPP. The suction piping ran through several hills and valleys that caused severe
slugging as described in Lessons Learned example D. Note from the figure that the
MPP discharged into a dedicated vessel in which the gas and liquid phases were
separated and measured. Liquid was recycled from this vessel to supply screw
sealant for the pump.
The 800 psi pressure boost capability of this pump was not required for this oil
field. However, because one of the goals was to push the pump to its design limits,
the pump discharge was throttled to achieve this differential pressure.
The process conditions for the Mitsue field trial are summarized as follows:
• At 1,800 rpm, flow was 25,000 bpd (485,000 scfd, 845 bpd of liquid);
• At 3,000 rpm, flow was 42,000 bpd (845,000 scfd gas, 1,765 bpd of liquid);
• Pump differential = 800 psi;
• Pump discharge = 850 psi (upstream of pressure letdown valve);
• 96 percent GVF to 97 percent GVF;
• Wellhead pressure drop = 150 psi;
• Increased oil production = 100 bpd @ 3000 rpm;
• Crude oil gravity = 40 degrees API to 43 degrees API.
The duration of the field trial was to be 6 months, but it ran for 10 months. The
pump was then shut down for inspection.
Goals and Results

A summary of goals and results follows, and additional details are provided in the
“Lessons Learned” section.
Goal 1. The primary goal was to demonstrate the technology in an actual oil field,
its robust design, and its reliability.
Results. Many unanticipated challenges were corrected to allow the pump to
operate reliably, with no mechanical failures for the pump’s entire run. After
10 months, the pump was shut down and inspected. The pump showed almost no
wear, and it was estimated that the pump could have operated reliably for at least
another year.
Goal 2. Another goal was to develop pump selection techniques.
Results. The Chevron MPP sizing method was greatly improved due to experiences
with this application.
Goal 3. The final goal was to confirm the pump’s economics.
Results. Crude oil production was expected to increase by 130 bpd to 270 bpd.
Even with this additional production, the financial justification for a “commercial”
application was marginal. The project broke even, with the value of money near
11 percent.
While the pump was being operated, Chevron Canada Resources was unable to
determine any incremental crude oil production, and the pump was shut down and
moved to a more profitable location, the Princess Field.
Lessons Learned
1. To accommodate the higher temperatures associated with pumping a high
GVF fluid, the pump clearances should be increased to accommodate a
300°F (150°C) rise above the highest suction temperature.
During the 1992 JIP test, rotors on the MPP seized while pumping a high
percent gas. Because all MPPs are more inefficient while pumping high
percentages of gas than pure liquid, the rotors heat up and expand.
During the Mitsue test, the screws did just that—they expanded, rubbed, and
seized. Leistritz increased the clearances between the two screws and also
between the screws and the case, such that the pump could expand without
rubbing. It was designed to accommodate a 392°F (200°C) rise. Later, after
being deployed at Mitsue, the MPP saw a 284°F (140°C) rise.
2. Require the pump manufacturer to design the twin screw pump rotors
with adequate stiffness.
The rotor design for a twin screw pump that pumps pure liquid allows the
rotors to bend and touch. The liquid acts as a cushion and lubricates each
screw’s interface with the other screw and with the case.
In an MPP with a high GVF fluid, hardly any lubrication is available to

lubricate the screws. Since the screws touching each other or touching the case
would result in seizure of the screws, the screws in an MPP have to be stiffer
than in a liquid pump. This condition is even more critical at higher differential
pressures, since a higher discharge pressure causes the screws to deflect more
than a lower discharge pressure.
Based on this experience, the pump should be designed such that the maximum
rotor deflection is less than half the internal screw clearance.
3. A screw sealant should always be provided if the gas content could rise
above 95 percent.
A screw sealant should always be provided if slugs are anticipated, even if the
average GVF is less than 95 percent. An externally supplied screw sealant is
one from a wide spot in the discharge piping, a separator vessel, or from an
outside source. The Mitsue pump experienced many severe liquid slugs.
Between these slugs, it pumped 100 percent gas for up to 2 hours, and a screw
sealant became absolutely necessary.
MPP installations might average less than 95 percent gas but still require a
screw sealant. If slugs are possible, the alternating liquid and pure gas streams
can occur while the overall average stream remains less than 95 percent. If pure
gas is fed to the pump during slugging conditions, the pump must have an
externally supplied screw sealant to prevent it from vapor locking, overheating,
and failing.
4. Twin screw pumps, properly designed, will operate reliably under severe
slugging conditions.
The Mitsue MPP was fed from a satellite 2 miles away. Flow from several
wells entered this satellite where the streams were tested. Fluid was then sent to
the MPP through a common suction line traveling through hilly terrain. Before
the MPP started, liquid collected in these pockets at the bottom of each hill.
After the pump started, this liquid was flushed out and then partially filled
again, with the cycle repeating itself. The net result was alternating slugs of
pure gas and liquid.
The most severe slugs occurred during the first 6 hours after startup. This initial
slugging is called “terrain slugging”. During terrain slugging, the MPP was fed
pure gas for 2 hours. The pump’s discharge temperature rose by 284°F (140°C).
Note that the pump was designed to accommodate discharge temperatures of
392°F (200°C).
After feeding the pump with gas, a liquid slug hit the pump, and the sound for
the motor changed dramatically (sounding like wood being cut by a table saw).
At the same instant that the liquid hit the pump, the suction pressure dropped by
approximately 30 psi, while the discharge pressure rose by approximately
2,000 kPa because of a downstream pressure control valve nearby. (Later, this
control valve was removed, and the discharge pressure surges became less
severe.) Next, the separator level rose significantly. The liquid, being cooler,
caused the temperature in the reservoir to drop by as much as 212°F (100°C).
After approximately 6 hours, the terrain slugs diminished but were replaced
with hydrodynamic slugs. This type of slug occurred more often but was less
severe. A hydrodynamic slug was arbitrarily defined as one that slowed the
pump down by approximately 20 rpm. There were 17 of these occurrences
counted in 5 minutes, which equates to approximately 5,000 slugs per day.
Throughout all this, the pump operated reliably. After 10 months of operation,
the pump was shut down. The inspection confirmed that the pump was in good
condition, and it was determined that the pump would have operated for at least
another year before requiring any maintenance.
The pump remained undamaged by the slugs, because as the slugs enter a twin
screw pump, they are split into two equal parts. These parts enter the screws
from the opposite ends at exactly the same time, cancelling out the resulting
forces.
5. Do not locate a downstream restriction (control valve, orifice, etc.) close to
the pump and preferably avoid having any, if possible.
A control valve, orifice, etc., if necessary, should be located at least 10 feet
downstream of the pump discharge flange. This distance provides a cushion to
dampen any surges. In the Mitsue installation, the MPP’s discharge pressure
spiked at 2,000 kPa when the control valve was in place. When the control
valve was removed, the pressure spikes were almost nonexistent.
6. Locate the MPP as close as possible to the wells to reduce the severity of
slugging.
The severe slugging experienced by the Mitsue pump could have been reduced
if the pump had been located closer to the wells, instead of 2 miles away, with
the suction line traveling through hilly terrain.
7. Model the MPP inlet and discharge piping to identify potential slugging
conditions and to design related equipment (piping, separator, etc.).
For future installations, the multiphase simulator, Pipephase, should be run on
the inlet and on the discharge piping from the pump to the downstream
separator. If Pipephase indicates slugging, a transient simulator, such as OLGA,
should be used to predict the size and frequency of the slugs.
With the results from these two programs, the facilities engineer will know if
slugs are likely and understand the length of time one could expect the
alternating 100 percent gas and 100 percent liquid slugs. This information will
also be required to size the associated system equipment (piping, separator,
etc.).
8. Size the downstream facilities to accommodate the MPP flowrate as if it
were pure liquid.
As mentioned, the Mitsue MPP experienced severe slugging, operating for at
least 2 hours on 100 percent gas and then several minutes on pure liquid. While
feeding pure liquid, the downstream separator could not dump the liquid fast
enough. As a result, the MPP shut down due to a high separator level.
If a liquid slug enters an MPP, the MPP pumps the liquid at its theoretical
flowrate. For the Mitsue pump, this was 40,000 bpd, or approximately
1,200 gpm for the length of time the liquid slug was present. The downstream
separator in combination with the discharge flow line needed to be designed to
handle this amount of liquid. However, in this case the discharge piping was
originally designed for only 100 gpm, relying on a small separator vessel for
surge capacity. Both had to be enlarged to accommodate the 1,200 gpm rate.
9. For electric motor drivers, check the motor’s available torque against the
pump’s required torque at all speeds but especially at the minimum
operating speed.
The motor size selected to accommodate the pump at normal conditions was
not large enough to supply adequate torque at the pump’s minimum speed. The
motor horsepower had to be increased substantially to handle the minimum
speed condition.
10. Use HYSIM or PROII to simulate the process conditions around the MPP
and to identify where flashing will occur. Flashing liquid into gas decreases
the pump’s total volumetric capacity.
One of the most important lessons learned from the Mitsue trial was that flow
from the Mitsue field amounted to only 60% of the MPP’s theoretical capacity.
The decreased capacity was due to flashing of the hot screw sealant or crude
oil, recycled from the downstream separator. Mitsue crude oil (40 degrees API
to 43 degrees API) is light and flashes easily. The flashed gas volume is much
greater than the liquid volume and occupies several times the pump capacity.
Figure 600-60 was developed from the process flow simulator, HYSIM, to
demonstrate where the flashing occurred. Referring to Figure 600-60, most of
the flashing occurred across the flow control valve, where the recycle stream
pressure dropped from 414 psi to 50 psi, flashing 64 bpd of liquid into
4,171 bpd of gas. This stream mixed with the crude oil coming from other
wells, and because the recycle was hotter, another 31 bpd of liquid flashed into
2,958 bpd of gas. The difference, in volume of over 7,000 bpd, represents
approximately 27 percent of the pump’s theoretical capacity of 26,200 bpd at
2,000 rpm. The 7,000 bpd of gas occupied pump capacity and backed out the
same quantity from the field.
The HYSIM simulation also showed more internal slip than originally
anticipated, approximately 12 percent of the theoretical pump capacity. The
same pump during the Houston JIP test showed only an 8 percent slip, but that
testing was done with a heavier crude oil (26 degrees API versus the 40 degrees
API to 43 degrees API crude at Mitsue). Adding all losses together (flashing,
internal slip, and the liquid that is recirculated), the internal losses amount to
43 percent. Thus, while the pump actually pumped 26,200 bpd, it was only
pumping approximately 57 percent of that amount or 14,867 bpd from the
Mitsue field, with the remainder being recycled within the pump.
Fig. 600-60 Hysim Results, Mitsue MPP Test
As mentioned earlier, MPPs handling GVFs above 95 percent need to be

designed larger to compensate for these high losses. An additional 15 to
35 percent of the field requirements should be added in sizing an MPP: for
heavy crude, use 15 percent, and for light crude, use 35 percent. The increased
pump capacity is determined from the recycle, slip, flashing factor, fully
described in Recycle, Slip, Flashing Factor in Figure 638.
11. Install a cooler to decrease anticipated flashing and to increase the actual
production flow.
One possible approach to decrease flashing if MPPs are used in light crude oil
service is to install a small cooler in the recycle stream. The HYSIM process
simulator indicated that this would have increased the actual pumped
production from the Mitsue field by approximately 10 percent. The cooler in
this case would have been small, approximately the size of a radiator in a diesel
truck.
12. If slugging is anticipated, size the driver HP as though the MPP were
pumping pure liquid.
Finally, because of the Mitsue experience, MPP sizing criteria was reviewed
and changed. It is recommended that, if slugging is anticipated, add up all the
fluid volumes (liquid and gas) at suction conditions, and calculate the
theoretical horsepower for that entire quantity, as if it were all liquid.
6135 Mitsue Pump Moved to the Princess Field

As mentioned earlier, the Mitsue pump did not pay out economically. It was,
therefore, moved in 1997 to the Princess Field, also in Canada. There, it pumped
crude oil, water, and gas from several wells through a pipeline approximately
18 miles long to an existing process facility.
The cost of moving the pump to the Princess Field was approximately $100,000,
which was considerably less than installing a conventional separation facility
(separator, liquid pump, compressor, etc.) at the Princess Field.
The process conditions for the Princess Field were as follows:
• 40,000 bpd (crude oil, 3,000 bpd, 70 percent GVF, 70 percent water cut) at
3,000 rpm;
• Pump differential = 725 psi;
• Pump discharge = 850 psi;
• Wellhead pressure = 125 psi;
• API gravity of crude oil = 23.
Note that the API gravity in the Princess Field was heavier (23 degrees) than at
Mitsue (40 degrees to 43 degrees). Therefore, the Slip, Recycle, Flashing factor at
the Princess Field was only 15 percent.
The Princess Field was still being developed when the MPP was installed. As new
wells were brought online, the pump capacity became inadequate. To gain more
pump capacity, the MPP’s speed was increased to 3,600 rpm, the highest speed
within the Chevron organization. The pump has run reliably under these conditions
for several years.
Lessons Learned
1. Leistritz or Bornemann twin screw pumps can operate reliably at
3600 rpm.
A Leistritz or Bornemann pump, if properly applied and installed, using the
information presented in this manual, can run reliably at 3,600 rpm. The
Princess Field application proved this.
2. Install a dual filter or strainer on the MPP suction.
If wax is present, a filter should be used. If a more coarse material is
anticipated, a strainer should be used. The filter/strainer should include a
differential alarm/shutdown where the filters can be changed while the MPP
runs. Design the shutdown setting such that the filter collapse pressure is
greater than the shutdown setting.
The crude oil at Princess Field was very waxy. Thus, the inlet filter plugged
several times, automatically shutting down the MPP. A dual filter was
eventually installed with the capabilities mentioned above.
3. Directly contact the mechanical seal supplier and carefully review the
seal’s design.
The pump manufacturer should not be relied on completely to communicate
accurate data to the seal supplier.
For the Mitsue pump, it was discovered that the pump supplier had planned to
install a seal with an unacceptably low pressure rating. The pump supplier
determined that the “static” rating of the seal, which was above the design
requirement, was adequate. However, the “dynamic” rating, which is the rating
of the seal while rotating, was actually higher, and therefore, the seal design
was not acceptable.
6136 Main Pass 313, Offshore Platform in the Gulf of Mexico

In 1993, a second Chevron MPP was installed on the offshore platform, Main
Pass 313, in the Gulf of Mexico, as shown in Figure 600-61.
Fig. 600-61 Leistritz Pump at Main Pass 313
This site included an unused well test header and test separator that could be
dedicated to the MPP. The field trial equipment arrangement is shown in
Figure 600-62. With this arrangement, any well or combination of wells could be
directed to the MPP, discharging into the dedicated test separator. The separator
measured the gas and liquid phases accurately. The discharge pressure of the pump
was regulated by the back pressure controller on the gas stream on the outlet of the
test separator.
The field trial at Main Pass 313 required a low pressure boost of 125 psi. The size of
the pump in relation to the well output dictated handling only one well at a time.
Most of the wells were gas lifted.
Fig. 600-62 Main Pass 313 Flow Diagram
GAS TO LP SYSTEM
PUMP BYPASS
OIL/WATER TO LP SYSTEM
FR
TEST SEPARATOR
LC
FROM TEST HEADER
MULTIPHASE
PUMP
CHEVRON MULTIPHASE METER LOOP (CMML)

FR
FC
LIQUID RECYCLE
This field trial used a Leistritz twin screw pump (pump capacity: 26,000 bpd at
1,800 rpm). The process conditions changed, depending upon which well was
tested. One of the wells on the platform produced the following:
• 500 bpd of liquid (mostly oil);
• 98 percent GVF amounting to 385,000 scfd or 25,500 bpd of gas at suction
conditions of 30 psig, 100°F (37.8°C);
• 26,000 bpd total of mixed flow (500 bpd of liquid and 25,500 bpd of gas).
The well normally produced into a separator at 100 psig. By reducing the wellhead
pressure to 30 psig with the MPP, incremental oil production was expected to
increase by 200 bpd.
Goals and Results

Goal 1. Demonstrate that the pump can operate reliably.
Results. During its run of 10 months, the double mechanical seals failed and were
replaced with single seals, flushed with liquid from the downstream separator. Also,
debris (weld slag) damaged the pump before a suction strainer was installed. After
these changes, the pump operation was considered reliable.
Goal 2. Confirm the economics.
Results. By decreasing the back pressure on various wells, the crude oil production
increased. Well A-23D increased production by 200 bpd when the backpressure
decreased from 100 psi to 60 psi. The economic payout was 3 months.
Goal 3. Validate the selection and application process.

Results. The pump and driver sizing criteria were revised based on the test.
Lessons Learned
1. Specify that the mechanical seals meet Chevron specification PMP 4662
that modifies API 682. (Using this standard will avoid many of the problems
detailed below.) If possible, use the less expensive, and perhaps more reliable,
single seals with an external flush.
Most of the problems encountered with the Main Pass 313 pump were related
to the four pressurized dual mechanical seals. The pressurized dual seal barrier
fluid must operate at a higher pressure than the seal chamber. Pressure reversal
occurs if the seal chamber pressure gets higher than the barrier fluid pressure.
The original Main Pass 313 seals could not tolerate any pressure reversal.
When this occurred, they failed.
Chevron specification PMP 4662 that modifies API 682 was not used for the
seals on this MPP. API 682 requires a pressurized dual seal to be able to
withstand 40 psi pressure reversal without damage.
The dual seals, because of poor reliability, were replaced with single seals with
a flush from the downstream separator. With this change, the seals were easier
to operate and proved to be reliable.
2. Specify balanced mechanical seals.
The seals at Main Pass 313 were not balanced. Hydraulic forces closing the seal
faces are lower with balanced seals, generating less heat and improving
reliability. API 682 and Chevron PMP 4662 call for using balanced mechanical
seals.
3. For pressurized dual mechanical seals, specify an external barrier fluid
pump, not a shaft driven unit.
On Main Pass 313, the dual seals were supplied with barrier fluid that was
delivered by a shaft driven pump. Therefore, until the pump started, the barrier
fluid to the dual seals was 0 psi. Note that, in a pressurized dual seal
arrangement, the barrier fluid must always be greater than the pressure in the
seal chamber, which is slightly above the suction pressure. A hand pump was
used to raise the barrier fluid pressure to a level above the suction pressure
before starting the MPP. An accumulator, a small vessel containing barrier fluid
in one end and process fluid at suction pressure in the other, was installed to
keep the pressure above suction while the pump operated. The two fluids were
separated with a rubber bladder. With this device, suction pressure was
communicated to the barrier fluid controller to keep the barrier fluid above
suction pressure.
Since this test, the MPP manufacturers have deployed a more reliable
pressurized dual mechanical seal that eliminates the need for a hand pump and
accumulator. Modern seals are designed with an external barrier fluid pump,
with its discharge pressure controlled high enough such that the seal chamber
pressure stays constant and is always higher than suction pressure.
An external barrier fluid pump also allows the barrier fluid system to operate if
the MPP is shut down, preventing corrosion. Finally, the operator can be
assured that barrier fluid is circulating at the proper pressure before attempting
to start the MPP.
4. Design the mechanical seals for an additional 200 psi to accommodate
pressure spikes.
At Main Pass 313, high pressure rushed into the pump when the inlet valve was
opened, causing a pressure surge. The pressure was communicated to the seal
system via the accumulator. The pressure differential across the twin screw
pump caused it to act like a motor and turn. This motion rotated the shaft driven
barrier fluid pump, producing a significant pressure spike. This added pressure
was enough to blow a gasket in the seal oil filter. It is now recommended that
the seals be designed for unanticipated spikes by adding another 200 psi to their
design.
The pressure spikes were caused by the shaft driven barrier fluid pump, which
increased pressure faster than the control system could react. An external
barrier fluid pump would decrease the possibility of spikes. However, even with
the external pump, requiring additional pressure for the seal design is
recommended.
5. For pressurized dual seal designs, make sure the barrier fluid cannot leak
into the lubricating system. Design for 5 gpm barrier fluid rate into the
seal chamber of each seal.
Lubricating oil was used for the barrier fluid, instead of the preferred mixture
of water and ethylene glycol. This water and ethylene glycol mixture was
preferred for the barrier fluid, because it dissipates heat better than the
lubricating oil. As stated, this mixture was not used, because it could have
leaked from the outboard mechanical seal and contaminated the lubricating
system. For future designs, the pump should be designed such that leakage of
barrier fluid cannot contaminate the lubrication system, allowing the barrier
fluid to be a mixture of water and ethylene glycol.
The barrier fluid flowrate to each mechanical seal was originally 1 gpm per
seal.
During the purchase order stage, the flowrate was raised to 5 gpm to allow for
better dissipation of heat.
6. Install a strainer in the suction line to the MPP.
The Main Pass 313 pump ran fine until early January 1995, when it began to
rub. The pump was inspected and repaired at the Leistritz facility in New
Jersey, where damage to the rotors and the pump case bore was found. The
damage was caused by a small piece of weld slag (1/4 inch) entering the
suction of the pump. The damage could have been avoided if a suction strainer
had been installed.
MPPs are usually installed downstream from a crude oil well. It is common for
well debris (weld slag, metal parts, etc.) to be flushed into the MPP. Therefore,
a dual strainer is absolutely necessary.
6137 Humble, Texas, Production Field

This motor driven, Leistritz L4NG 126-98 MPP operated from 1997 until 2004,
when a well rework overpressured the pump, destroying it. This pump’s maximum
rate was 13,000 bpde at 92 percent GVF, with a differential pressure of 75 psi.
Though the pump increased the production of the field, its main purpose was to give
legacy-Texaco some operating experience with MPPs. Before being destroyed by
the well rework, the pump had been the longest continuously operating MPP in the
U.S. The pump was also used as a test pump for a new grease canister seal flush
system installed, which is revolutionary in its development and application. (Refer
to Grease Canister Seal Flush System in Section 6146.) Figure 600-63 shows a
Leistritz pump at the Humble, Texas, field being removed for maintenance.
Fig. 600-63 Humble Field, Texas, Leistritz Pump Being Removed for Routine Maintenance
Lessons Learned
1. Twin screw MPPs can have a long mean time between failure (MTBF).
2. Mechanical seals operating at 6 psig suction pressure can be flushed with
grease from an intermittently feeding grease canister.
This system can be very useful if the MPP stream is continuously high in GVF
or particulate or both.
6138 Duri Trial, Indonesia

In 1998, a Bornemann twin screw pump and a Sulzer helico-axial pump were
installed side by side in Duri, Indonesia. They operated under the same conditions,
pumping crude oil from 33 wells. The purpose of this trial was to determine the best
pump type (twin screw or helico-axial) that should be used in the future light oil
steam flood (LOSF) project, also in Indonesia. The pump is shown in
Figure 600-64, and its simplified flow diagram is shown in Figure 600-65.
Fig. 600-64 Test Facility at Duri, Indonesia
Fig. 600-65 Simplified Flow Diagram of the Duri Test Facility
Duri is a typical steam flood oil field, where steam is continuously injected to
reduce the viscosity of the heavy crude oil and pressurize it into the producing
wells. Each Duri producing well used a rod pump to lift the crude and send it on to a
processing unit. The MPPs were installed between the wells and the processing unit.
As mentioned, both pumps were designed to meet the same conditions. They are:
• Total multiphase fluid flowrate = 125,000 bpd;
• Composition = 20 percent water, 5 percent API 22 gravity crude oil, 75 percent
gas, and 90 bpd sand; (Sand concentration was 0.3 percent by volume or
0.5 percent by weight.)
• Inlet temperature = 250°F (121°C);
• Inlet pressure = 50 psig;
• Pressure Boost = 150 psig.
Twin Screw MPP Test Results

The twin screw MPP was a Bornemann pump, model MW 9.5zk-67, coupled
directly to a 500 HP electric motor with speed controlled by an adjustable frequency
drive (AFD). This allowed the pump to operate from 650 rpm to 1,950 rpm,
although 1,800 rpm was used throughout the trial. The pump bearings were partially
submerged in four self-contained lube oil reservoirs, cooled by a water jacket.
Four single mechanical seals were installed in the pump at the end of each screw.
The seals were flushed with 5 gpm of fresh water (API Plan 32). The water went
through the mechanical seal chamber into the pump through a close clearance throat
bushing. The pressure drop across the bushing eliminated the possibility of the
pumped fluid being released to the atmosphere. During a seal leak, only the water
flush was released. The flush not only cooled and lubricated the seal as flush, but it
also acted as screw sealant for the pump.
The Bornemann pump incorporated a discharge containment chamber that separated
gas and liquid, retaining the liquid and letting the gas flow on. It also included a
recycle valve to send a variable amount of retained liquid back to suction. The
recycle valve was closed during the trial. However, because of a groove in the
valve, it still transferred 3 to 4 percent of its capacity back to suction. This, plus the
external seal flush, was more than adequate to supply the screw sealant needs.
After operating 24 hours, the pump was shut down and partially inspected to
determine if any wear occurred in the recycle valve or if any sand had collected in
the containment chamber. There was no sign of wear or sand accumulation in either.
The pump ran a total of 1 month and was shut down for a thorough internal
inspection. During the first 25 days, the pump was fed crude oil from the wells. On
the 26th day, the crude oil feed was stopped and replaced with steam. This was done
to simulate steam breakthrough, a condition expected at the LOSF.
The pump’s slip was calculated daily, and it showed no significant wear during the
first 25 days of operation. However, the slip increased during the steam injection
test, indicating significant wear. After 3 hours, the steam was discontinued, and
Fig. 600-66 Twin Screw Pump Slip versus Time
crude oil was again fed to the pump. The slip stayed high but constant, indicating
that no additional wear was occurring. This is shown in Figure 600-66.
The exact reason for the different wear patterns is not known. It was theorized that
wear did not occur during the first 25 days because the sand in the crude oil stayed
in suspension and passed through the pump without touching the rotors or case. It
did not separate out and, therefore, did not cause erosion. Crude oil also lubricated
and protected the metal.
On the 26th day, during the steam injection test, the pump suffered significant wear.
It is believed that steam condensed in the suction line and/or inside the pump. The
pump then processed steam, condensate, and sand. The sand in the steam and/or
condensate centrifuged out and wedged between the screws and the casing, causing
significant wear.
As mentioned above, the pump ran a month and was then shut down and inspected.
It confirmed the slip data that showed significant wear patterns. They are
summarized as follows:
• The screws were worn an average of 0.012 inch. Note that they were originally
carbon steel, coated with Praxair Super D-Gun 2015 at 0.003 inch thick.
• Much of the coating on the screws was worn off.
• A groove occurred at each balance hole in the screw OD, through the coating
into the parent metal, extending to the outside of the non-leading edge of each
screw.
• The leading edge of each screw was rounded on the OD. This is shown in
Figure 600-67.
Fig. 600-67 Twin Screw Wear Patterns
Helico-Axial MPP Test Results

A Sulzer helico-axial pump was tested for a month, after the twin screw test. As
mentioned earlier, this pump was tested under the same process conditions. The
pump, Model MPP 7/7, was a seven stage machine on one shaft, driven by a
700 HP electric motor. Its speed was controlled by an ASD between 2,200 rpm and
3,600 rpm. During the test, the motor ran most of the time at 3,450 rpm.
The Sulzer helico-axial pump was more complicated than the Bornemann twin
screw pump. It contained two mechanical seals, one single and the other a
pressurized dual seal. The single mechanical seal was located at the outboard end
and a pressurized dual mechanical seal at the motor end. A radial bearing was
located outboard of the single seal on the outboard end, while another radial bearing
was positioned between the two seals in the pressurized dual seal configuration. A
thrust bearing was located on the motor end, outboard of the mechanical seal and
the radial bearing.
Unlike the twin screw pump, on which thrust was balanced, here, thrust from slugs,
etc., was not balanced, and the helico-axial pump required a thrust bearing. With
severe slugging conditions, the thrust force would become too large and overload
this bearing. Therefore, a buffer tank is usually required to be installed ahead of the
pump to smooth out the slugs and minimize pump thrust. Because of the flat terrain
at Duri, severe slugging was not anticipated, and the buffer tank was not installed.
This turned out to be the correct decision, because severe slugging did not occur.
The pump ran for a month and showed little or no apparent process problems or
wear. On the 26th day, a steam injection test was attempted. During the initial phases
of the test, the pump shut itself down several times, refusing to pump high gas
fractions. For example, at 94 percent gas, the pump produced only 20 psi
differential pressure instead of the desired 150 psi. The steam test was aborted after
several attempts lasting 40 minutes, and the pump wear from a possible steam
breakthrough could not be determined.
The helico-axial MPP was shut down after running 1 month because of a dual
mechanical seal failure. Additionally, a malfunctioning solenoid valve in the
lube/barrier oil system caused the lube oil to overheat, creating several leaks. The
seal and the solenoid valve were repaired, and the pump started again approximately
1 month later. It ran another 2 months, and the seals failed again. While the pump
was shut down, an internal inspection was not attempted, because the LOSF project
management had already decided to use the twin screw pump. The Bornemann twin
screw pump was less expensive, less complicated, more efficient, could handle
slugs better, and had a wider range of flowrates and pressures. Detailed reasons are
presented in Figure 600-7 and Figure 600-8 in Figure 621 and Figure 600-68 in the
following Lessons Learned section.
Lessons Learned
1. Twin screw pumps can pump crude oil containing large quantities of sand,
with little or no wear, if the crude oil is heavy or viscous enough. Sand will
pass through the pump and will not cause wear or settle out.
This is probably the most important lesson learned from the Duri test. Sand at
90 bpd was pumped by the Bornemann pump without any wear, and it did not
settle out inside the containment chamber. This statement is based on the
findings of an inspection after 24 hours of operation, the process data obtained
throughout the test, and finally, the complete inspection after a month of
operation.
Also, the Sulzer helico-axial pump that operates at a much higher speed did not
show signs of wear while operating. However, it was never opened and
inspected for sand, as was done on the Bornemann pump.
2. Pump wear from sand erosion will likely occur during a steam
breakthrough or if a lighter, less viscous crude oil is fed to the pump.
Duri’s crude oil was 22 degrees API, which is fairly heavy. It is believed that, if
it were less viscous, more wear would have occurred. Water is less viscous, and
the twin screw pump suffered significant wear while attempting to pump steam,
steam condensate, and sand. Additionally, a study by the University of
Erlangen in Nuremburg, Germany, with sand in water, confirmed significant
wear, destroying a Leistritz twin screw pump in only a few hours.
To better define the viscosity in which wear would occur, a computer
simulation program was run. It showed significant wear from sand in water and
light crude oil. There was virtually no wear when the crude oil was heavier than
an API gravity of 30 degrees.
3. The Bornemann twin screw pump is a better choice than the Sulzer helico-
axial pump for most surface applications.
Figure 600-68 compares the Bornemann twin screw pump to the Sulzer helico-
axial pump.
Fig. 600-68 Comparison—Typical Twin Screw MPP versus the Helico-axial MPP
Item Twin screw Helico-Axial
Purchase price $400,000 $600,000
Ability to handle slugs Excellent without additional Requires additional equipment, a buffer
equipment. vessel.
Lube oil system Self-contained, each bearing. Requires a circulating system (pump,
cooler, filter, etc.).
Mechanical seals Four single seals with API Plan 32. One single and one pressurized dual
seal with a seal oil circulating system.
Ability to withstand sand erosion Good because the sand was Unknown
included in a heavy (22 degrees
API gravity or lower), viscous (330
cp) crude oil.
Flow range (bpd) 4,000 to 125,000 (design = 110,000 to 140,000 (design = 125,000)
125,000)
Suction pressure (psig) 50 (80 initially) 50 (80 initially)

Discharge pressure (psig) 100 (175 design) 100 (175 design)
Design speed 1,920 rpm 3,450 rpm
Electric motor size 500 HP 700 HP
Efficiency during test 45% 22%
In Figure 600-68, it is obvious that the twin screw pump was the better choice
for the future LOSF project in Indonesia. The twin screw pump was less
expensive, less complicated, more efficient, better at handling slugs, and
pumped a wider range of flowrates and pressures.
It is difficult to extrapolate this conclusion to other applications, although this
conclusion is probably correct for most surface facilities, such as onshore
installations or offshore platforms. Helico-axial pumps are installed extensively
subsea and perhaps, there, they might be a better choice. Each installation must
be evaluated on its own merits.
4. If sand is anticipated, consider the following to minimize wear:
– Using a removable liner;
– Rounding the leading edge of each screw;
– Coating the liner with satellite or tungsten carbide, and boride or nitride the
screws; (Refer to Section 640.)
– Keeping the original design speed as low as possible;

– Using an ASD to allow the speed to increase if wear is indicated;
– Considering additional screw turns or locks to develop the head required.
Balance holes should not be drilled on the OD of the screws.
6139 COB Facility, El Tigre Field, Venezuela; Boscan Field, Venezuela; Kome,
Miandoum; and Belobo Fields, Chad
COB Facility, El Tigre Field, Venezuela
This field had four twin screw transfer pumps that pumped a multiphase stream of
crude oil, water, and gas from El Tigre to Hamaca. The manufacturer of the pumps
was Flowserve, a manufacturer not recommended by Chevron. These pumps were
purchased by Chevron’s Venezuelan partner. The four pumps were motor driven
NPS 14 HP pumps, capable of pumping 75,000 bpd, with 10 to 20 percent GVF at
800 psig differential pressure and 1,800 rpm, using 1,600 BHP motors. The pumps
were installed in 2002 and immediately became a continuous problem.
These pumps were out of service so frequently that a machinist specializing in twin
screw MPPs was hired full time to keep them operational. The pump’s primary
failure mode was its mechanical seal system, but rubbing of the rotor against the
stator bore was also common.
The suction line was a significant distance from the suction tank and had a 30 foot
high loop in it immediately prior to the suction flange of the pump. As a result, gas
slugging was a problem.
In 2005, the pumps were replaced by five Bornemann MW 10.6zk pumps. These
pumps have been operating since then without problems. They are shown in
Figure 600-69.
Fig. 600-69 Bornemann Pumps at El Tigre Field, Venezuela (Courtesy of Bornemann

Corporation)
With the installation of the Bornemann pumps, the following corrective actions
were taken:
• The pumps were located much closer to the suction tank.
• The suction line was kept in the horizontal plane, with as little rise as possible
anywhere along the line.
• The Bornemann pumps had stiffer rotors, which prevented contact between the
rotor and the bore.
Boscan Field, Venezuela

Chevron had similar problems with Flowserve twin screw pumps in the Boscan
field, Venezuela, prior to this application. Here, too, the pumps were purchased by
the Venezuelan partner.
Kome, Miandoum, and Belobo Fields, Chad

Chad is another example where Flowserve pumps failed. The MPPs repeatedly
failed vibration limits on the Flowserve test stand. The order was cancelled, and
Leistritz pumps were purchased instead. Today, the Leistritz pumps are operating
reliably. The installations are shown in Figure 600-70 and Figure 600-71.
Fig. 600-70 Leistritz Pumps in Kome Field, Chad (Courtesy of Leistritz Corporation)
Lessons Learned
1. Purchase reliable MPPs—do not purchase pump designs that are not
recommended.
As mentioned above, the pumps at the COB facility experienced continuous
maintenance problems, averaging an extensive repair each week. To solve the
problem, all four pumps were replaced with reliable Bornemann pumps.
As mentioned earlier, the two manufacturers of twin screw pumps
recommended are Bornemann and Leistritz. As of this writing, no other
Fig. 600-71 Leistritz Pumps in Miandoum Field, Chad (Courtesy of Leistritz Corporation)
manufacturers are recommended. Perhaps, in the future, the other

manufacturers will improve their design and have a reliable pump to offer.
Flowserve’s original pump design allowed for contact between the two rotors
and between each rotor and the case. Each shaft was allowed to deflect under
the discharge pressure. This design was originally for liquid only service but, in
these applications, was being supplied for multiphase service. We have seen
some of the El Tigre shafts fatigue at the root of one of the screws, where the
screw joins the shaft. The fatigue crack eventually propagated to the point
where the working diameter of the shaft became so reduced that the shaft failed
in torque.
Flowserve has also provided extensive rework of its pumps at the Chevron
Boscan, Venezuela, location. The work was unprofessional, and the pump soon
failed. (Photos are available.) Flowserve has admitted problems with its
workmanship and has vowed to redesign its pump, fabrication methods, and
QA/QC methods. To date, no successful applications of Flowserve MPPs are
known.
2. Keep the suction line as short as possible, and avoid vertical loops.
The pumps at the COB facility were located several hundred feet from the
suction tank. The suction line looped approximately 30 feet up and down to get
over an elevated walkway. The pumps were unreliable by themselves, but the
poor routing of the suction line aggravated the situation. The new Bornemann
pumps mentioned above were installed closer to the suction tank without any
vertical loops.
61310Main Pass 59 A
In 2007, a Leistritz MPP, Model 4HK-365-175, was installed on this offshore,
unmanned platform in the Gulf of Mexico. It is driven by a 1,700 HP natural gas
engine, the largest of any current installations. Its speed is approximately 1,200 rpm.
The installation is shown in Figure 600-72.
The MPP was designed to meet the following process conditions:
• Flowrate = 150,000 bpd with 95 percent GVF;
• Pressure boost of 640 psi.
The pump increased production from 18 wells by decreasing wellhead pressure
from 350 psig to 150 psig, while the pump took over boosting the pressure into the
higher pressure discharge line. Before the pump’s installation, each well had to
match the wellhead pressure from adjacent wells, such that the 18 wells could flow
into a common production line. After various options were considered, the MPP
was shown to be the most economic solution to the situation. The unit went online
in September 2007. It is the only Chevron application driven by a natural gas
engine. This driver was chosen since excess gas was being produced, and the gas
was then “essentially free” for the application. (Refer to Figure 600-72.)
Fig. 600-72 Leistritz MPP Installed on Main Pass 59A
Lessons Learned
1. If properly sized, natural gas engines can be used for driving MPPs.
Though natural gas engine drivers are not constant torque machines and twin
screw pumps are, these engines can effectively be used, if sized for the
maximum torque requirements of the pump.
2. Use “free” fuel when it is available.

If “excess” or uneconomical or hard to recover natural gas is available, such
gas should be considered for use as fuel for the MPP driver.
3. Make certain that the mechanical seals are rated for the shut-in pressure
of the well.
In the Main Pass 59A installation, it was determined after installation of the
MPP that the mechanical seals were not designed to withstand the well’s shut-
in pressure. The suction piping to the well had to be modified by installing a
pressure control valve in the suction line and using it and the bypass line to
allow the shut-in pressure to be handled by the pump. (See Figure 600-73.)
Fig. 600-73 Main Pass 59A—Addition of the Suction Control Valve
6140 New Developments (as of 2008)
6141 General
This section describes technological developments from various MPP and
mechanical seal manufacturers. Some are being researched by the manufacturer and
are not yet commercially available. Others can be purchased but have not yet
established a track record for reliability.
These new developments, before being applied, should be thoroughly investigated,
and all the information that is possible should be obtained from the manufacturer
and the users, including a “user” list with names, phone numbers, and Internet
addresses, such that the development’s reliability can be substantiated. Finally, an
ETC expert should be involved before any of these new developments are used in
any application.
6142 Twin Screw Pump Developments

Downhole Twin Screw Pump
This type of twin screw pump is designed to be installed vertically in an oil well.
Thus, its diameter is small, as are its screws. The larger the diameter of the well, the
more likely this pump will be able to be used. CAN K, a company in Edmonton,
Alberta, Canada, manufactures such a twin screw pump. Chevron installed one in
2002 at the Rangely field, but the results were poor, partly because the well’s
process conditions were not adequately transmitted to CAN K. CAN K has had
other failures, as well, although most of the failures were because the application
was poorly defined by the user. (Technology development stage TQP rating of
TDS 6.)
CAN K has a few other downhole pumps operating elsewhere, driven by a long
shaft from a driver on the surface. The reliability of the shaft driven units is
unknown. (Technology development stage TQP rating of TDS 7.)
Flowserve and Colfax also have plans to develop a downhole pump and are looking
for a partner with whom to develop the pump. (Technology development stage TQP
rating of TDS 3.)
Bornemann and Leistritz, the two major manufacturers of twin screw pumps, are not
actively developing a downhole pump.
As of this writing, a reliable downhole twin screw pump is not available. This concept
should be avoided until its reliability has been proven by actual field applications.
Digressive Screw Pitch

Bornemann recently developed a “digressive screw pitch” for its twin screw pump
rotor design, and it is commercially available. A digressive screw pitch rotor is one
in which the pitch decreases as the flow moves from the inlet to the outlet.
Bornemann claims the new pitch improves the pump capacity, efficiency, and
reliability.
Since 2007, Bornemann has produced several, and it has made the digressive screw
pitch standard for its MW 8.5, MW 9.5, and MW 10.5 pumps.
The digressive screw pitch concept makes technological sense. Screw pitch
decreases towards the discharge, where the pressure increases and the volume of gas
decreases. Thus, the pump efficiency should improve. Manufacturers of helico-axial
MPPs have included this concept in their designs for years. Because of this and
because Bornemann has decided to make the digressive screw pitch standard on
several of its MPPs, the digressive screw pitch should be considered for future
Chevron applications. (Technology development stage TQP rating of TDS 9.)
Subsea Twin Screw Pumps

Subsea MPPs are usually economically attractive, because they can dramatically
increase production by overcoming the huge static pressure needed to move the well
fluid to the ocean’s surface. However, installing them such that they are quickly and
economically retrievable and increasing their reliability such that they will not need
to be retrieved for a number of years is a difficult task. A reputable packaging
company must be used, usually one other than the MPP manufacturer. Chevron
should consider installing subsea pumps if the economics show they could be
attractive. The Chevron TQP should be used to determine the equipment’s
suitability for the specific application.
There is currently only one commercial twin screw subsea pump installation,
consisting of two Bornemann pumps, installed for the British Petroleum King Field
in the Gulf of Mexico. The packaging was done by Aker Kvaerner.
British Petroleum claims the pumps increase production by 20 percent and overall
recovery by 7 percent, extending the life of the field. The pumps are on the ocean
floor under 5,500 feet of water. The pumps are driven by subsea electric motors
controlled by an ASD on a host platform. They supply 6.6 kv to the subsea pumps
15 miles away. (Technology development stage TQP rating of TDS 9.)
A Leistritz subsea twin screw pump is expected to be operational soon, off the coast
of Brazil. That pump, SBMS-500, will pump from the ocean floor (2,100 feet below
sea level) to the surface, requiring 870 psi. This pump will experience an average
GVF of 87 percent but is designed for 100 percent, with an external supply of screw
sealant. Slugs are expected, because the pump is located 1.5 km from the well.
Lubricating oil is supplied to the pump and motor from a tank located on the host
platform. This system has been extensively tested for years onshore at facilities in
Brazil. It is unique and patented. (Technology development stage TQP rating of
TDS 8.)
Finally, Flowserve initiated an effort to develop a subsea twin screw pump with
Shell and an independent contractor, SubSea 7. The relationship with SubSea 7 has
since been discontinued, and the future of the project is currently unknown.
GE-Vetco Gray is developing a new subsea twin screw pump, as well. It has been
designed to overcome many of the current subsea MPP problems. It has not yet been
field tested. (Technology development stage TQP rating of TDS 5.)
6143 Helico-Axial Pump Developments

High Boost and/or Hybrid Pumps
Framo has developed a “high boost” pump that extends its operating parameters.
Framo has also developed a “hybrid” pump that extends the operating parameters
even further. The hybrid pump is a helico-axial pump followed by a section of
conventional centrifugal pump impellers. This concept is similar to the
Schlumberger ESP gas handler, in which the helico-axial portion homogenizes the
mixture, returns some of the gas into solution, and increases the fluid pressure to
reduce gas volume and bubble size before the fluid enters the conventional
centrifugal stages. (Refer to High Boost and/or Hybrid Pumps in Section 6143.)
Framo claims its hybrid pump will develop a differential pressure up to 2,900 psi
with a GVF of 50 percent. (Technology development stage TQP rating of TDS 6.)
Sulzer and Aker Kvaerner are also developing a “hybrid” pump, similar to the
Framo pump, with the first stages being helico-axial, followed by more stages of
radial flow impellers. Their plan is to develop this pump to be installed subsea at
deeper depths than are currently available. (Technology development stage TQP
rating of TDS 6.)
Wet Gas Compressor

Framo has recently developed a “wet gas compressor” that can pump a 90 percent
GVF at a differential pressure of 600 psi. Framo is now testing its first unit at its
Norway test facilities. The compressor has not yet been deployed in any field
application. Therefore, applying one of these in a Chevron facility is not
recommended until the concept has proven its reliability. (Technology development
stage TQP rating of TDS 4.)
Dresser has also recently developed a wet gas compressor. However, its wet gas
compressor separates the liquid and compresses only the gas.
6144 PCP Developments

Equal Wall Stator
seepex recently started to sell its PCP with an “equal wall stator” that has a uniform
elastomer thickness. This is different than a conventional stator, in which the stator
wall has a varying thickness. seepex claims this design delivers a higher pressure
boost, dissipates heat more efficiently, and allows the pump to handle higher GVFs.
Finally, the “equal wall stator” is shorter than a conventional stator, which is an
advantage, especially on an offshore platform where space is a premium.
One known installation is in the Northern UK near the Shetland Islands, for Conoco.
seepex is proud of this installation and wrote a paper entitled, “The Next Generation
of Progressive Cavity MPP, Use of a Novel Design Concept for Superior
Performance and Wet Gas Compression”, by Kamran Mirza of seepex.
This concept appears technologically solid, and Chevron should consider using the
“equal wall stator” if a PCP is being considered the most attractive MPP alternative.
(Technology development stage TQP rating of TDS 9.)
seepex is also currently developing a wet gas compressor using the “equal wall
stator” concept to handle 100 percent gas indefinitely. The seepex unit has not yet
been deployed. (Technology development stage TQP rating of TDS 5.)
Metal to Metal PCP

Kudo has developed a PCP with a metal rotor and a metal stator. Kudo has
approximately 40 operating PCPs but, as yet, has not sold any for multiphase
service. In 2005, Kudo Industries developed this high temperature PCP for liquid
phase only and has deployed all its pumps in downhole, single well applications.
Kudo calls its PCP the mm-PCP or PCM Vulcan.
The main advantage of this type of PCP is that it can tolerate high temperatures, up
to 570°F (300°C) that occasionally exist in oil wells. Conventional elastomeric
stators are limited to 300°F (150°C). Also, the mm-PCP can handle wide swings in
temperature that can be created from a steam flood operation.
The mm-PCP also has promise in pumping particulate. To date, wear caused from
pumping particulate has not been quantified, although the manufacturer claims this
pump should be able to handle up to 5 percent (by weight) sand.
The mm-PCP has been field tested in France and Canada. During both tests, the
pump wore with time, opening its clearances. This wear was seen by observing the
reduction of its volumetric efficiency that dropped from 70 to 90 percent, to as low
as 35 percent, although the time it took for this efficiency drop is not known. During
another test, it ran effectively for over a year without failure—a period considered a
success by the pump manufacturer.
The mm-PCP has not yet been applied in multiphase service, and therefore, the
concept should be approached with caution, and this pump should not be applied in
multiphase service until the concept has proven reliability. (Technology
development stage TQP rating for multiphase of TDS 1.)
6145 ESP Developments

Helico-Axial Gas Handlers
As mentioned in Section 683, Schlumberger owns Framo, the manufacturer of
helico-axial pumps and Reda, who makes ESPs. As such, it has developed a helico-
axial gas handler to be installed ahead of the suction to its ESP. (Refer to
Section 623 for a discussion on helico-axial pumps in general and Section 662 for
specific information on Framo). The gas handler is 6.3 meters long. The
combination of gas handler and ESP is called the Poseidon ESP. With this
configuration, Schlumberger claims the ESP can pump a GVF up to 75 percent,
because it homogenizes the mixture, returns some of the gas into solution, and
increases the fluid pressure to reduce gas volume and bubble size to the inlet of the
conventional ESP. Schlumberger’s first gas handler was installed in 2003, designed
to handle approximately 60 percent gas. Since 2003, users have installed 16 similar
machines.
Gas handlers are fairly common. However, this application of a gas handler is
unique. It should be considered for use only after the subject has been thoroughly
investigated, which includes contacting users to determine the pump’s reliability.
(Technology development stage TQP rating of TDS 9.)
6146 Mechanical Seal Developments

Diamond Faced Seals
Mechanical seals with diamond faces on the rotating and non-rotating surfaces have
been developed by John Crane, Flowserve, and Burgmann. This type of seal face
material may have promise for MPPs that handle sand or other types of abrasives.
The John Crane diamond face material is called JCDiamond. It uses a proprietary
diamond growth process that is a crystalline pure diamond film, consisting of ultra-
nano sized crystals on a lapped surface of the seal face. The diamond facing
material is very hard and smooth, building less heat between the faces. John Crane
claims that, compared to a conventional seal face material, silicon carbide, the
JCDiamond film’s wear resistance is higher, its power consumption is lower, its
need for lubrication is reduced, and its operating range (temperature and pressure) is
wider.
JCDiamond is now available for use, but the material has not yet been applied
commercially. It does not yet have a track record for reliability. John Crane is
looking for someone to test the material.
The Burgmann diamond faced seals are similarly grown diamond crystals on a base
material.
These seals have the potential to reduce the need for the large seal flush streams
required today, the necessity to have a constant seal flush stream, and the potential
to handle flush streams containing particulate. These seals could make a significant
impact on MPPs located in remote areas, pumping a stream containing particulate.
Diamond seal technology is new and not yet deployed. Although the concept sounds
good, a potential application should be approached with caution until it has proven
reliability. (Technology development stage TQP rating of TDS 5.)
Grease Canister Seal Flush System

Chevron developed a grease canister seal flush system, as yet unpatented, that ran
reliably for several years in a Leistritz twin screw MPP in the Humble production
field. The system used grease canister lubricators to feed a stream of grease between
the inner seal and the outer lip seal. The canister held the grease and used an
extremely small integral pump with which to pump the grease from the canister to
the seal faces, as required. So far, the seal system has been tried on only one MPP
with an operating suction pressure of 6 psig. This system also has the potential to
make a significant impact on MPPs located in remote areas, pumping a stream
containing particulate.
This seal should be tried in a future application, in which other, more proven seals
are difficult to deploy. ETC should be involved in the application. (Technology
development stage TQP rating of TDS 9.) Contact Bob Heyl at ETC:
[email protected].
6150 Definitions and Acronyms
6151 Definitions
Adjustable Speed Drive (ASD), also called Variable Speed Drive (VSD)—An
electrical speed controller that varies the speed of an electric motor driver, thereby
controlling the flowrate of the driven pump. A Variable Speed Drive (VSD) is one
common subcategory type of an ASD.
Armor I or Armor X—A proprietary material used by Centrilift to coat bearings,
usually specified if pumping a high GVF fluid, sand, or particulate.
Centrifugal Pump—A pump classification in which a pump uses centrifugal force
to increase pressure at the pump discharge. Fluid flows into the eye of an impeller
and is thrown outward by its rotation. ESPs and helico-axial or rotodynamic pumps
are types of centrifugal pumps.
Conventional System—Separation equipment that can perform the same function
as a multiphase pump and consists of a separator, gas compressor, liquid pump,
separate sand or particulate handling system, and perhaps two discharge pipelines
(one gas and one liquid).
Chrome—A material overlay, installed to withstand sand or particulate erosion.
Duktil—A proprietary material used by seepex to coat the rotors in its pumps
designed for high GVF applications.
Electric Submersible Pump (ESP)—A centrifugal pump or combination helico-
axial/centrifugal pump usually installed in an oil well, consisting of multiple small
diameter impellers on one shaft. The pump assembly can be very long, consisting of
an electric motor, protector, gas separator, and the pump itself. When applied
downhole in a well, the motor is submerged in liquid.
Equal Wall Stator—A proprietary stator configuration by seepex that has a
uniform thickness around each screw cavity. This design differs from the
conventional stator in which the elastomeric thickness varies.
Flashing—The process in which a liquid near its boiling point quickly changes
phase to a gas, if its pressure is suddenly reduced. Note that the flashed gas occupies
several times the original liquid volume.
Gas Hardening—A metallurgical coating to harden the surface of a metal in order
to withstand erosion from sand or particulate. Two types are used in multiphase
pumps, nitriding, and boride gas diffusion, or boriding.
Gas Handler—An ESP subcomponent, usually installed between the gas separator
and the pump itself, to increase the pressure of the suction stream. The gas handler
homogenizes the mixture and re-liquefies some of the gas, lowering the GVF to the
ESP.
Gas Separator—An ESP subcomponent that uses centrifugal or rotary motion to
separate gas from the pumped fluid and vent it into the annulus of an oil well.
Gas Volume Fraction (GVF)—Volume of gas as a percent of the total volume of
all fluids (gas and liquid) at pump suction conditions, usually expressed as a percent
of total flow.
Helico-axial (Rotodynamic) Pump—A pump design having numerous impellers
and diffusers with characteristics similar to a centrifugal axial flow pump and which
is capable of pumping multiphase fluids.
Lock or Screw Locks—In a twin screw pump, the point at which the edge of the
screw meets the inside diameter of the case or case liner.
Material 5530—A proprietary hard material used by Schlumberger to withstand the
erosion of sand or particulate.
Multiphase Fluid—A fluid composed of two or more phases (e.g., produced well
fluids containing natural gas, crude oil, water, wax, natural gas hydrates, sand, and
particulate).
Multiphase Pump (MPP)—A pump that is capable of pumping at least two phases,
usually a liquid and a gas, but often includes a liquid, gas, and solids.
Productivity Index (PI)—A number that describes the expected production
increase from a crude oil well if the bottom hole flowing pressure is reduced.
Progressing Cavity Pump (PCP)—A positive displacement rotary pump
consisting of a single, serpentine screw that contacts a stationary liner, called a
stator. The stator is pressed into the case and usually is an elastomeric material.
Fluid is pushed along the shaft by the screw as the pressure is increased from the
pump suction to its discharge.
Protector (Seal)—An ESP subcomponent that separates the pumped fluid from the
coolant in the electric motor.
Recycle, Slip, Flashing Factor—A factor used to increase the calculated pump size
to allow for slip and flashing inside a twin screw pump.
Redalloy—A proprietary material used by Schlumberger in pumps designed for
applications that contain an appreciable amount of CO2.
Rotary Positive Displacement Pump—A pump classification in which a circular
rotating pump traps a fixed volume of fluid and then pushes it through the pump,
increasing its pressure along the way. Twin screw and PCPs are rotary positive
displacement pumps.
Rotodynamic (Helico-axial) Pump—A pump design having numerous impellers
and diffusers with characteristics similar to a centrifugal axial flow pump and which
is capable of pumping multiphase fluids.
Screw Sealant—In a twin screw pump, the fluid that is used to provide a liquid seal
between the edge of the screw and the case and between the edge of one screw and
the root of the adjacent screw.
Single Phase Pump—A pump that pumps only liquid.
Slip—In a positive displacement pump, fluid that passes backward across the edge
of the screw from a high pressure area to a lower pressure area, expressed as a
percent of total flow.
Slugging (Slugs)—Alternating of large pockets of liquids and gas.
Stellite 12 Weld Overlay—A weld overlay used to withstand erosion from sand or
particulate.
Subsea Seabed Installation—Pumps that are installed on the seabed floor.
SUME—A proprietary coating used by Sulzer to withstand erosion from sand or
particulate.
Surface Installation—Pump installations located on the surface, either onshore or
on an offshore platform.
Technology Qualification Process (TQP)—A Chevron process used to assess new

and existing technology, enabling risk based decisions to be made by rating the
appropriateness of a technology for a specific set of project conditions.
Tungsten Carbide—A metallurgical coating used to withstand erosion caused by
sand or particulate.
Twin Screw Pump—A positive displacement rotary pump, consisting of two
noncontacting screws installed side by side, in a case. One screw is driven by a
driver, while the second screw is driven by the first through a set of gears. The
screws are held apart and timed by the set of gears.
Variable Speed Drive (VSD)—An electrical speed controller that varies the speed
of an electric motor driver, thereby controlling the flowrate of the driven pump. A
VSD is a particular type of Adjustable Speed Drive (ASD).
6152 Acronyms
ASD—Adjustable Speed Drive
BHP—Brake Horse Power
ESP—Electric Submersible Pump
ETC—Energy Technology Company (Chevron)
FED—Facilities Engineering Department (Chevron)
GOMBU—Gulf of Mexico Business Unit
GOR—Gas Oil Ratio
GVF—Gas Volume Fraction
LOSF—Light Oil Steam Flood
MEPS—Machinery and Electrical Power Systems (Chevron)
MPP—Multiphase Pump
MTBF—Mean Time Between Failures
NPV—Net Present Value
OD—Outside Diameter
P&ID—Piping and Instrumentation Diagram
PCP—Progressing Cavity Pump
PD—Positive Displacement
PI—Productivity Index
PSA—Preferred Supplier Agreement
QA/QC—Quality Assurance/Quality Control
RTD—Resistance Temperature Detector
SS—Stainless Steel
SSU—Saybolt Seconds Universal
TDS—Technology Development Stage
TQP—Technology Qualification Process
VFD—Variable Frequency Drive
VSD—Variable Speed Drive
WC—Water Cut
6160 References
6161 Company Specifications

1. PMP-DG-4662 CRN, Data Guide for API 682, Data Sheets for John Crane
Mechanical Seal Selection
2. PMP-DG-4662 FS, Data Guide for API 682, Data Sheets for Flowserve
Mechanical Seal Selection
3. PMP-SC-4662, Shaft Sealing Systems for Centrifugal and Rotary Pumps
(Chevron’s exception specification for API 682)
6162 American Petroleum Institute (API)

1. API RP 11S, Operation, Maintenance and Troubleshooting of Electric
Submersible Pump Installations
2. API RP 11S1, Electrical Submersible Pump Teardown Report (ANSI/API RP
11S1-1998)
3. API RP 11S2, Electric Submersible Pump Testing
4. API RP 11S3, Electric Submersible Pump Installations (formerly API RP 11R)
5. API RP 11S4, Sizing and Selection of Electric Submersible Pump Installations
6. API RP 11S7, Application and Testing of Electric Submersible Pump Seal
Chamber Section
7. API RP 11S8, Electric Submersible Pump System Vibrations
8. API 610, Centrifugal Pumps for Petroleum, Petrochemical and Natural Gas
Industries (ANSI/API Std 610-2004) (also ISO 13709)
9. API 676, Positive Displacement Pumps—Rotary (includes Errata dated June
1994)
10. API 682, Shaft Sealing Systems for Centrifugal and Rotary Pumps
6163 National Association of Corrosion Engineers (NACE)

NACE MR 01-75, Petroleum and Natural Gas Industries—Materials for Use in
H2S-Containing Environments in Oil and Gas Production (also number as
ISO 15156)
Abstract
This section briefly describes several pump types, including artificial lift pumps
(electric submersible, sucker rod, and hydraulic-turbine pumps), jet pumps, air-
diaphragm pumps, regenerative pumps, and slurry pumps. It also lists references for
more detailed information on electric submersible and sucker rod pumps. Most of
the information on artificial lift pumps is reproduced from the Electric Submersible
Pumps Manual (Chevron USA, Denver).
Contents Page
710 Air-Diaphragm Pumps 700-2
720 Regenerative Turbine or Disk-friction Pumps 700-3
730 Jet Pumps 700-5
731 Performance Variations
740 Slurry Pumps 700-6
750 Artificial Lift Pumps 700-7
751 Electric Submersible Pumps
752 Sucker Rod Pumps
753 Hydraulic-Turbine Driven Pumps
754 Gas Lift Systems
760 Peristaltic Pumps 700-16
770 Multiphase Pumps (MPPs) 700-17
771 Two-Screw (Twin-Screw) Multiphase Pumps
772 Hydrodynamic (Helico-axial) Pumps
773 Progressive Cavity Pumps (PCPs)
774 Piston Pumps
700 Miscellaneous Pumps Pump Manual
710 Air-Diaphragm Pumps

Air-Diaphragm Pumps (Figure 700-1) are commonly used in utility and chemical
services. They meet neither ANSI nor API standards and are most often used
because of their portability and ability to run on compressed air.
Fig. 700-1 Cross Section of Air-Diaphragm Pump (Courtesy of Wilden Pumps & Engineering Co.)
Advantages of air-diaphragm pumps include:

• Portable and reusable.
• Operate on compressed air (piping or hose supply).
• Handle entrained air and solids.
• Can be run dry without significant damage.
• Simple and relatively dependable.
• No shaft seals or packing complications are involved.
• Non-sparking.
• Available in corrosion-resistant materials.
• Self-priming.
• Low cost.
Disadvantages include:
• They do not meet normal standards for continuous duty process service.
• Air-supply valves occasionally become plugged. (These valves require oil
lubrication from a self-contained reservoir. This is normally not a problem, but
the valves do need oil to operate.)
• Ball-check valves wear and occasionally stick.
Pump Manual 700 Miscellaneous Pumps
• Limited to relatively low temperatures (approximately 200°F) and pressures

(approximately 120 psi).
• Noisy operation.
720 Regenerative Turbine or Disk-friction Pumps

Disk-friction pumps are frequently called regenerative turbine-type pumps by manu-
facturers. They are also sometimes called Westco pumps, which was the name used
by one of the earliest manufacturers of this type of pump. The designation “turbine
type” is frequently confusing and not at all descriptive of the principle of operation
involved.
Externally, a disk-friction pump appears similar to a centrifugal pump, but the prin-
ciple of operation is quite different. Figure 700-2 shows several internal views of a
typical disk-friction pump. The pump consists basically of an impeller wheel or
disk, with vanes on the periphery, rotating in a concentric case. The case provides an
open passage around the impeller except at the cut-off point between the discharge
and suction nozzles. Liquid enters the pump at the periphery of the impeller and is
carried by the friction of the rotating impeller to the discharge port. The cut-off
point between the discharge and the suction is close-fitting to prevent excessive
internal leakage. Also, the plates on each side of the impeller must be close-fitting
to prevent excessive leakage along the sides of the impeller. A pump usually
contains several impellers.
Fig. 700-2 Cross Section of Disk-Friction Pump From Pumps for Chemical Processing by J.T.
McGuire, Marcel Dekker, Inc. Copyright 1990. Reprinted with permission.
The characteristics of disk-friction pumps make them principally suited for rela-
tively low capacities at medium-high heads. Figure 700-3 shows a typical perfor-
mance curve for a disk-friction pump at 1750 RPM. Such pumps are made in
capacities up to 150 gpm or more, but ordinarily these are not economical compared
to centrifugal pumps at capacities above 15 or 20 gpm.
Figure 700-3 shows that the total head falls off rapidly as capacity is increased.
Maximum brake horsepower is required at shutoff, with the horsepower dropping as
capacity is increased. Frequently, drivers are not sized for shutoff, but a relief valve
is installed to prevent overloading at low flow rates.
Fig. 700-3 Typical Performance Curve From Pumps for Chemical Processing by J.T. McGuire,
Marcel Dekker, Inc. Copyright 1990. Reprinted with permission.
As might be expected, the efficiencies of disk-friction pumps are never very high,
but for small capacity applications they are often more effective than centrifugal
pumps. For example, the efficiency illustrated in Figure 700-3 at 6 gpm for
170 ft-head is about 35%. The efficiency of a centrifugal pump for this rating would
be about 15%.
One advantage of the disk-friction pump over the centrifugal pump is that the disk-
friction pump can handle up to about 20% by volume of vapor along with the liquid
pumped. Disk-friction pumps are also self-priming, provided the case is filled with
liquid to act as a seal.
Disk-friction pumps are ordinarily used in clean, non-viscous services at flow rates
less than 20 gpm. For low flow rates, disk-friction pumps cost less than centrifugal
pumps of comparable capacity.
The efficiency of the pump falls rapidly as the close clearances between the rotating
disk and the case are increased by wear. Sand, mill-scale, or similar foreign parti-
cles in the liquid pumped may expand the clearances to a point where satisfactory
operation can no longer be obtained. This is the primary weakness of the disk-fric-
tion pump and has significantly reduced its application. Disk-friction pumps are best
used for condensate return or small boiler feed services or for LPG loading where
the liquids are usually clean.
730 Jet Pumps

A jet pump (Figure 700-4) produces a high-velocity jet of almost any fluid to pump
another fluid by entrainment with pressure recovery in a diffuser. The theory of jet
pump operation can be found in most engineering handbooks and therefore is not
presented in this manual. Since the performance of jet pumps is developed empiri-
cally from manufacturers’ tests, application of jet pumps to specific services is
usually a matter of clearly stating the requirements to be met.
Fig. 700-4 Typical Jet Pumps (Ejectors) (Courtesy of Schutte & Koerting Division, Ketema, Inc.)
Jet pumps are usually less efficient than other pumping devices and require a source
of high-pressure fluid for operation. This considerably limits their range of applica-
tion.
Jet pumps are primarily used to produce and maintain a vacuum by removing vapors
from a closed system. In this service they are usually called “ejectors.” Ordinarily,
steam provides the motive power. Jets for this purpose are frequently combined with
condensing equipment, and the entire system is purchased as a unit. Steam jets are
occasionally used for boiler feed water injectors where small capacities are
involved. Because jet pumps have no moving parts, they are sometimes used to
dewater sumps, especially where gritty or dirty liquids are handled.
Jet pumps are sometimes used for mixing liquids in a vessel. Some of the contents
are pumped back into the vessel through a submerged jet designed to entrain and
circulate the rest of the tank.
731 Performance Variations

Jets function properly over a relatively narrow range of deviation from designed
operating conditions. The effects of deviations can best be determined by jet
suppliers or from performance curves as applied to the specific service in question.
The following characteristics generally apply to steam jets in vacuum pumping
service:
1. Increasing motive steam pressure increases the flow of motive steam and
usually (but not necessarily) the pumping capacity. If motive steam pressure is
reduced, the opposite applies. If it is reduced too far, the jet discharge will start
to pulsate and suction may be lost.
2. If the pressure into which the jet pump discharges is higher than design,
pumping capacity (and thus vacuum) will decrease and suction may be lost.
3. Wet motive steam may impair performance; the extent of impairment depends
on the volume of liquid. Wet steam will cause increased erosion in the jet
nozzle and diffuser. Erosion is the common cause of poor performance.
740 Slurry Pumps

Slurry pump applications are exercises in exceptions and special cases. Slurry pump
engineering is a defensive practice of preparing for maintenance rather than
preventing it. Therefore, when selecting a slurry pump you should: 1) evaluate
pumping conditions carefully, 2) gather complete data about available pumps,
3) select the best pump for the application, and 4) be prepared for “excessive”
maintenance.
You should consider all of the following slurry characteristics. In addition, pump
engineers need to know how much the parameters will vary during operation, and,
what has worked best before in a similar service. Slurry fluid characteristics are as
follows:
• Capacity
• Suction pressure and NPSHA
• Differential pressure (head)
• % slurry by mass and volume
• Temperature
• Fluid corrosiveness to metals and elastomers

• Type and hardness of particulate material
• Size range and % breakdown of particulate material
• Tendency of particulates to agglomerate
• Shaft sealing requirements
• Clean sealing fluid availability
• Reliability requirements
Positive displacement progressive cavity type pumps have been most frequently
selected for slurry services in both the upstream and downstream sides of the
Company. This is because they lend themselves well to broad ranges of moderate
discharge pressures, flows below 100 gpm, and slurry percentages. Originally
designed for sewage sludge services, they usually are not optimally designed for the
oil and chemical industries.
Progressive cavities, like all slurry pumps, have finite lives. The manufacturer’s
emphasis is on quick disassembly and parts sales rather than precision design.
Beyond progressive cavities, very few other types of positive displacement pumps
have worked satisfactorily in slurries. One notable exception is Chevron Chem-
ical’s phosphate rock slurry pipeline between Vernal, Utah and Rock Springs,
Wyoming. It uses specially designed piston pumps to supply the high-discharge
pressures.
Centrifugal pumps are often the best selection for slurry services with flows above
50-100 gpm and heads below 150 feet, particularly with non-hazardous fluids.
Impeller speed/erosion relationships usually limit higher-head designs. The mining
industry employs a broad range of rubber lined centrifugal pumps which should be
considered for water slurries. Several pump companies supply horizontal and
vertical centrifugals to the cement and paint industries which resist wear with hard-
ened iron materials.
There are applications such as 750°F FCC bottoms pumps which demand features
of both slurry pumps and high-temperature API 710 designs. Lawrence Pump
Company has adequately addressed these demanding requirements. Their pumps
have been demonstrated at the Richmond Refinery.
750 Artificial Lift Pumps

Throughout a well’s producing life, reservoir pressure may be insufficient, and arti-
ficial lift may be required to maintain or increase production. The four major
methods for artificially lifting a well fluid include:
1. Electric Submersible Pumping
2. Sucker Rod Pumping
3. Hydraulic-turbine Pumping
4. Gas Lift
Each method has advantages and disadvantages when compared to each other.
Figure 700-5 defines the capability of each lift method under different operating
conditions. Selection of any method must be based on the well and reservoir
characteristics.
Fig. 700-5 Lift Methods

Operating Sucker Rod Hydraulic Electric Gas
Condition Pumps Pumps Pumps Lift
Scale Good Fair Poor Fair
Sand Fair Fair Poor Excellent
High Volume Poor Good Excellent Excellent
Flexibility Good Excellent Poor Good
Wax Fair Good Good Poor
Corrosion Good Good Fair Good
High Entrained Gas Fair Fair Fair Excellent
Deviated Well Poor Good Fair Excellent
Depth Fair Good Good Good
Simple Design Yes No No No
The following characteristics should be considered when choosing an artificial lift

system.
• Name and location of well and its proximity to other wells
• Reservoir data
• Bottom-hole pressure and productivity index
• Oil and water production
• Gas/oil ratio
• Depth of lift required
• Casing size, condition, and single or multiple completion
• Type and condition of existing lift equipment, if any
• Operating problems, such as scale paraffin, corrosion, etc.
• Availability of service, parts, and personnel familiar with the system
• Availability and cost of gas and/or electricity
• Degree of automation desired
• Ability of the system to meet changing production conditions, if required
• Safety
• Investment and operating costs
The most important selection criterion for an artificial lift system is to find the
system that will deplete the well most economically.
751 Electric Submersible Pumps

There are two types or designs for electric submersible pumps (ESP). One is
designed for oil or water production in downhole well installations. The other is
primarily a shallow-water well pump.
The downhole ESP (Figure 700-6) is common in oil field production. Typical
manufacturers are TRW/REDA, Centrilift-Hughes, and Oil Dynamics, Inc. ESPs
consist of four major sections:
1. Submersible motor and cable
2. Seal section
3. Pump (sometimes supplied with gas separator)
4. Topside transformer and electrical controls and instrumentation
Fig. 700-6 Electrical Submersible Pump System (Courtesy of the “Oil & Gas Journal”.
Reprinted with permission.)
These units run at 3600 RPM; each motor/transformer set is matched. The setting
depth, cable selection, and power rating of the motor determine the required topside
voltage and transformer rating. This is unique in every installation. The setting
depth can be as much as 13,000 feet and produce flows between 200 and
20,000 BPD. Motors are usually less than 200 HP.
The pump, seal section, motor, and tubing are designed to fit inside standard casing
sizes. Motors and pumps may be 5¼ inch or 7 inch diameter. Due to the restriction
in diameter, motors may be 20 or 30 feet long and pumps may have up to
400 stages. This also restricts the allowable diameter of the shaft, typically about
one inch.
The reliability of these pumps decreases as well temperature, deviation, and horse-
power increase. The average run life is between one and two years.
The advantages of an electric submersible pump system are:
• Economically produce high volumes of fluid.
• Work well in locations with minimal surface area, such as offshore
platforms.
• Low initial cost.
• Use a single tubing string and vent gas through the annulus.
The disadvantages of an electric submersible pump system are:
• Poor flexibility in lifting unexpected, rapidly changing, or low-volume produc-
tion.
• The electric power supply cable is affected by depth, corrosion, temperature, or
handling, which can cause cable failures.
• For wells with gas in addition to liquid, the electric submersible pump must
have good gas separation for efficient pumping.
• Abrasives, wax, or scale decrease the ESP run life. (This is true for all down-
hole pumps.)
• Relatively poor reliability.
• The inherent low efficiency of a centrifugal pump makes the overall efficiency
of ESPs poor.
• Pulling and repair costs are high.
For more information on ESPs, refer to “Electric Submersible Pumps,” from
Chevron USA, Northern Region (Denver), or refer to one of the following:
API RP-11R. Recommended Practice for Electric Submersible Pump Installations.
API RP-115, Recommended Practice for Operation Maintenance and Trouble-
shooting Electric Submersible Pump Installations.
Handbook for Oilfield Subsurface Electrically Driven Pumps. Centrilift-Hughes,
Inc. 200 W. Stuart Roosa Drive, Claremore, OK, 74017.
Shallow-water well pumps are very different from downhole ESPs. They are not
specifically designed for deep well installations and never used in oil production.
However, they are used in seawater lift services and other shallow, cool water
services. Shallow-water well pumps consist of a submersible motor and pump. The
seal is built into the motor; the pump does not require one. The electrical support
systems are standard and do not require special voltages or dedicated transformers.
The pump and motor are packaged by the pump manufacturer. Typical manufac-
turers of this type of submersible motor are TRW/Plueger, Hayward-Tyler, and
Byron-Jackson. The motors are larger in diameter, up to 14 inches and do not have
the pressure-balancing capabilities of the downhole ESP. The motors are much
shorter in length, usually about 10 feet and are capable of up to 700 HP. These
motors normally operate at 1800 RPM.
These motors can be installed onto almost anyone’s standard vertical turbine pump.
Typical manufacturers of the pump are Goulds, Peerless, Ingersoll-Rand, Bingham,
and Dresser-Worthington.
The shallow well ESP is usually more expensive than the deep well ESP, with little
overlapping coverage between them.
752 Sucker Rod Pumps

Sucker rod (walking beam) pumps are normally limited to downhole applications in
certain producing areas. These systems (Figure 700-7) are designed for lower
producing rates (up to 1,000 BPD) from shallow to moderate well depths (up to
11,500 feet). In a sucker rod pump system, a surface pumping unit converts the
rotary motion of a prime mover (gas engine, electric motor, etc.) into a recipro-
cating action. The reciprocating action is transferred to a positive displacement
pump located downhole.
The advantages of sucker rod lift systems are:
• They are the oldest and most widely used means of artificial lift, accounting for
85% of artificial lift in the United States. Consequently, sucker rod pumps are
the best understood artificial lift method.
• Low initial costs in shallow to moderate depth wells.
• Flexibility to handle changing production volumes.
• They use a simple tubing string and vent natural gas through the annulus.
The disadvantages of sucker rod pumping are:
• Increasing the depth and/or the produced volumes increases the system’s cost
and reduces the system’s production capabilities.
• Require a pulling unit for running and retrieving the downhole pump or
replacing sucker rods.
• Subject to wear, fatigue, or corrosion failures.
• Volumetric efficiency is reduced in wells with high GOR, solids, wax, H2S, or
corrosion.
Fig. 700-7 Typical Sucker Rod, or “Walking-Beam” Pump From Pump Handbook, (1976) Edited by Karassik, Krutzch,
Fraser & Messina. Used with permission from McGraw Hill.
References
The following references provide detailed information on sucker rod pumps and
their corresponding pumping units.
API RP-11L, Recommended Practice for Design Calculations for Sucker Rod
Pumping Systems (Conventional Units)
API BUL-11L4, Bulletin containing Curves for Selecting Beam Pumping Units
API RP-11AR, Recommended Practice for Care and Handling of Subsurface Pumps
API RP-11BR, Recommended Practice for Care and Handling of Sucker Rods
API SPEC-1B, Specification for Oil-Field V-Belting (Includes a design procedure
for power application of V-Belts)
API RP-11G, Recommended Practice for Installation and Lubrication of Pumping
Units
API SPEC-11AX, Specification for Subsurface Sucker Rod Pumps and Fittings
API SPEC-11B, Specification for Sucker Rods (Pony rods, polished rods, couplings,
and sub-couplings)
API STD-11E, Specification for Pumping Units
Rod Pumping Equipment: Surface and Downhole Analysis and Inbal Programs.
Chevron Oil Field Research Co., La Habra, CA. (Analysis of rod pumping system
performance, based on monitoring conditions at surface.)
Artificial Lift Sucker Rod Pumping. Royalty Enterprises, Inc., Garland, Texas. (This
book provides a detailed look at all aspects of sucker rod pumping systems.)
753 Hydraulic-Turbine Driven Pumps

Hydraulic-turbine driven pumps (Figure 700-8) can handle up to 12,000 BPD
production from shallow to deep well depths (up to 17,000 feet). In hydraulic-
turbine driven pump systems, a downhole hydraulic turbine receives its energy by
mean s of a fluid, usually oil or treated water. The power fluid is pumped downhole
by means of a surface pump.
Fig. 700-8 Hydraulic Pump From The Dictionary of Petroleum Exploration, Drilling and
Production by Norman J. Hyne, (Courtesy of PennWell Publishing.)
The advantages of this lift system are:

• Economics for hydraulic pumps improve with depth and produced-volume
requirements.
• Economical and flexible range in handling changing produced-volume require-
ments.
• The downhole pump can be run in and out of the well by circulation.
• Work in deviated, directional, or inaccessible well locations.
• Power fluid (either produced oil or water) can be a treated fluid to provide heat,
chemical, or dilutent.
• Hydraulic pumps work well in pumping crudes because the crudes are diluted
by the lighter power fluid.
• When using hydraulic powered pumps, a centrally located pump station can
serve a number of wells.
The disadvantages of hydraulic pumps are:

• Cannot vent gas with a single tubing string.
• When used with central station for multi-well system:
– Power fluid requires treatment.
– Long high-pressure lines for power fluid are required.
– Production testing is more difficult.
• Initial capital costs are high.
• When oil is used, the power fluid volume required may become very expensive.
• High maintenance cost.
• Sonic bottomhole pressure testing can’t be performed on subsurface comple-
tion types. A pressure bomb cannot be run with a reciprocating-type hydraulic
pump. The jet-type hydraulic pump allows pressure bombs to be run in the
chamber at the bottom of the pump.
• Power requirements are high due to the combined inefficiencies of power fluid
pumps, downhole pumps, and hydraulic turbines.
754 Gas Lift Systems

Gas lifting (Figure 700-9) can lift high volumes of fluid (up to 20,000 BPD on
continuous injection) from moderate to deep well depths (up to 14,000 ft.). In gas
lifting, the reservoir fluids are lifted from the wellbore by injection (intermittent or
continuous) of a high-pressure gas to supplement the reservoir’s energy.
Fig. 700-9 Gas Lift From The Dictionary of Petroleum Exploration, Drilling and Production by
Norman J. Hyne, (Courtesy of PennWell Publishing.)
The advantages of gas lifting are:

• Gas lifting is very economical where gas is readily available at high
pressures.
• Gas lifting handles changing volumes with economic flexibility.
• Gas lifting works in low productivity wells with high GORs.
• When using gas lift, abrasive material (sand, etc.) offers fewer problems.
• Gas lifting works in deviated, directional or small casing wells, or where
surface area is minimal, such as offshore.
• Gas lift valves are retrievable with wireline, eliminating the need for a work-
over rig in some instances.
The disadvantages of gas lifting are:
• Gas lifting requires a continuous high-pressure gas supply, operating costs will
increase as gas prices do.
• Gas lifting efficiency generally decreases with lower produced volumes and
shallower depths.
• Gas lifting cannot pump down a well because hydrostatic pressure from the
produced fluid column is still on the formation.
• Gas lifting can increase hydrate and paraffin accumulation due to the cooling
effect of gas expansion.
• Depletion of a low BHP well is difficult with gas lifting.
• For gas lifting, the casing must be completely intact (without leaks) and capable
of withstanding the lift pressures.
760 Peristaltic Pumps

Peristaltic pumps, also called hosepumps (Figure 700-10) are used for pumping
fluids such as waste sludges, lime and cement mortar, adhesives, and shear
sensitive fluids such as latex paints.
Peristaltic pumps are positive displacement pumps. The pumping action results from
alternate compression and relaxation of the hose, which is manufactured of Buna or
natural rubber. The hose is compressed within a circular housing by a rotor with
lubricated shoes that come in direct contact with the outer walls of the hose. The
pumped fluid is in contact only with the inner walls of the hose.
Peristaltic pumps are available in capacities to 300 gpm and pressures to 220 psig.
Advantages of peristaltic pumps include low fluid shear, no seals or packing, no
valves, and dry running. Disadvantages include limited hydraulic range, pulsating
flow, and limited life due to the component life of the hose.
Fig. 700-10 Peristaltic Pump (Courtesy of Watson-Marlow/Bredel Pumps)
770 Multiphase Pumps (MPPs)

In oil and gas production, the term multiphase flow refers to a flow having a
minimum of “two-phase flow.” “Two-phase flow” describes a mixture of hydro-
carbon liquid and hydrocarbon gas. If water is added as a separate immiscible phase,
the system is described as “three-phase flow.” If solids, such as sand, are also in the
stream, the stream is sometimes referred to as “four-phase flow.” Two-phase, three-
phase, and four-phase flows are all referred to as “multiphase flow.”
Oil wells produce a mixture of oil, water, inert gas, natural gas, and sometimes sand,
hydrates and wax. The transfer of such a mixture via a flow line to a central
processing facility is a multiphase production system. The traditional method of
handling this multiphase stream is to separate the phases and then use a pump for
the liquids and a compressor for the gas. Often the two phases are transferred using
separate pipelines.
Multiphase pumps (MPPs) are now available, accepted, and reliably used to move
the fluids and solids from Point A to Point B (multiphase flow) without first sepa-
rating the fluid into its particular phases.
Commercial multiphase pump technology is a mature technology, initiated during
the mid 1990’s. Approximately 1000 of these pumps are in service, almost all in
surface, land-based applications. An increasing number are used on offshore plat-
forms, and a few are found subsea.
Several types of multiphase pumps exist. The two most common are the two-screw
(twin-screw) pump and the hydrodynamic (helico-axial) pump. Single screw or
progressive cavity pumps have also been used in multiphase service, as have piston
pumps.
Multiphase pumps have their flow rates rated at pump inlet or suction conditions. A
100,000 bpd pump pumps 100,000 bpd of liquid and gas. The gas occupies space
within the pump, as well as the liquid, and its contribution to required flow capacity
needs to be calculated at pump inlet conditions. Since gas expands as the pressure
drops, the size of the multiphase pump gets larger as the suction or inlet pressure
drops. The pump also gets larger if the proportion of gas in the fluid is large. A
multiphase pump can become extremely large and quite expensive if the suction
pressure drops much below 40 psig or if the gas volume fraction (GVF) gets much
higher than 90%. In such cases, the conventional method of transporting multiphase
fluid (that is, separating the phases, pumping the liquid, compressing the gas, and
running each phase through its own pipeline), could be more economical. Each
application needs to be evaluated on its own merits.
Because the technology is fairly new and each application can encounter its own,
unique problems, it is highly recommended that a Chevron specialist assist in the
procurement and installation of all MPPs.
771 Two-Screw (Twin-Screw) Multiphase Pumps

The most common surface multiphase pump is a modified two-screw or twin-screw
pump. This pump is shown in Figures 400-3 and 400-8. It is a positive displace-
ment pump.
One of the two screws is driven by an electric motor, gas engine, or other type of
driver at speeds between 900 and 3,600 RPM. The driver is usually an electric
motor with a variable speed drive (VSD), but gas engines may also be used. The
other screw is driven by the first through gears. As each screw turns, trapped fluid is
pushed from inlet to outlet.
The fluid enters the case at the center and is split in half. Each of the two streams is
then directed to one bearing end of the screws. This theoretically eliminates any
thrust; however one of the bearings is designed to handle some thrust.
Multiphase pumps, when operated at a low speed, are also low-shear pumps, and, as
such, can be used to minimize the amount of emulsion formation that takes place
when pumping some oil and water combinations.
Mechanical Seals
The multiphase fluid is contained in the pump case by four mechanical seals, one on
each end of each screw. The most common approach is to use a single mechanical
seal at each location with a clean, cool external flush (See API Standard 682, Seal
Flush Plan 32). Produced water and crude oil from the pump discharge is often
used. Only clean crude with a low gas volume fraction (GVF) in an application
without gas surges can be used as a flushing fluid without using a downstream sepa-
rator. If the crude contains significant gas, or if the crude stream is likely to have
surges of gas, a downstream separator is needed to make sure the flush is 100%
liquid and that enough liquid capacity is available to flush the seals for the entire
time the pump experiences the gas slug.
The seal should incorporate an internal throat bushing to increase the pressure in the
seal chamber. The throat bushing minimizes foaming (gas evolution) at the seal
faces and therefore provides better lubrication to the seal faces.
Cyclone Separator
If the particulate level is minimal, the discharge from the pump can be directed to a
cyclone that removes the particulate and sends the particulate back to the pump
suction. Clean liquid is then fed to the multiphase pump seals as the flush fluid. (See
API Standard 682, Seal Flush Plan 31.) A cyclone separator application has many
disadvantages that are described in Section 833 of this manual under Seal Flush
Plan 31. If the crude contains any significant solids (0.5% is considered extremely
significant), it will not be a good source for the seal flush fluid.
Pump Screws
Besides the mechanical seal, the design of a multiphase twin-screw pump differs
from a conventional twin-screw pump with shorter (distance measured between
bearings) and more rigid screws. In multiphase service, the screws are designed not
to contact each other or the pump case, even if the pump is over-pressured.
Erosion and Wear Prevention

Multiphase service in an oil field often includes sand, usually suspended in the
crude oil. To decrease maintenance costs caused by erosion or wear, a machine-shop
replaceable liner is often installed in the case. It is less expensive to replace than
repairing the entire case. Also, the wetted parts of the pump can be coated by, or
specially treated, to produce a hard surface. Rotors can be nitrited or boride diffu-
sion coated. Both treatments involve hardening a thin surface layer of the rotor.
Stators and stator liners can be stellited with good success. Chrome plating is not
acceptable since it often looses its bond to the base metal and peels. If the sand is
associated with a crude oil that is highly viscous at flowing temperatures, the sand
will pass through the twin-screw pump with minimal wear or other damage.
Clearance Adjustments
There is a close clearance between the screws and between each screw and the
pump case. If sand or other small diameter solids are present, the clearances are
altered for the particulate size. If a high temperature fluid is being pumped, the
clearances will be opened to allow for the expansion of the rotors due to high
temperature. These clearances also allow a small amount of fluid to slip back to the
previous stage. This is called slip. As a percent of total flow, it increases as the
differential pressure across the pump increases, or as the viscosity of the fluid
decreases, or as the clearance increases. Minimal slip will occur if the pumped fluid
is a viscous, particulate and gas free, ambient-temperature liquid.
Particulates
Particulate size and hardness are important in designing the pump clearances as is
the flowing temperature and viscosity of the fluid being pumped. As a rule of
thumb, the particulate in the flowstream should be kept under 1% of total flow.
Capacity
Maximum capacity of a twin-screw pump is about 250,000 bpd at the pump inlet
conditions. Since twin-screw multiphase pumps do not have contact between the
rotors, and between the rotors and the pump case, they can pump 100% gas for short
durations, but GVFs of about 90% are usually a safe upper limit without taking
special precautions.
Artificial Lift Service

Multiphase twin-screw pumps for artificial lift services are being tested. In this
application, the pump is placed down-hole in the well bore. Surface or down-hole
motors can be used, depending on the conditions. Artificial lift is a new application
for MPP technology which shows some promise for crudes with:
• High viscosity, or
• High GVF or
• High temperature or
• High solids or
• High emulsion-forming conditions.
772 Hydrodynamic (Helico-axial) Pumps

The hydrodynamic or helico-axial multiphase pump is another common pump in
multiphase service. Although it looks like a multistage centrifugal pump, it is actu-
ally a multistage mixed-flow pump, with the lines of flow in much more of an axial
direction than a common centrifugal Electric Submersible Pump (ESP) (see
Figure 700-11). Helico-axial pumps are used in artificial lift applications.
The pump boosts pressure through several stages of open impellers. The multiphase
fluid moves from one stage to another axially or parallel to the shaft. Between each
impeller stage is a diffuser or set of vanes attached to the pump case that direct the
fluid from one stage to the next stage. The impeller openings become progressively
smaller with each successive stage compensating for the compression of gas.
Radial loads are supported by sleeve bearings when the speed exceeds 3,600 RPM.
Below that, the bearings are anti-friction. Unlike the twin-screw pump, this pump
feels a significant axial thrust that is countered with a tilting pad thrust bearing.
Mechanical Seals
To keep the multiphase fluid inside the pump, two mechanical seals are needed, a
single seal at the outboard end and a double seal at the motor end. A radial bearing
is located outboard of the single seal, while another radial bearing is positioned
between the two seals in the double seal configuration. The thrust bearing, on the
motor end, is outboard of the mechanical seal and the radial bearing.
Fig. 700-11 Hydrodynamic (Helico-axial) Pump (Courtesy of Sulzer Pumps)
Helico-axial Pump
MECHANICAL SEAL
BEARINGS
Lubrication
A combination lubrication/barrier fluid circulating system is required to lubricate all
bearings and to carry away the heat generated by the faces in both mechanical seals.
The barrier fluid system consists of a lube oil pump, a cooler, filters, a reservoir,
associated alarms/instrumentation devices and an accumulator to keep the lubrica-
tion/barrier fluid pressure at discharge pressure.
Speed and Capacity

These pumps operate at speeds from 2,000 to 7,000 RPM. They can be run from an
electrical motor, or a gas engine driver, or a hydraulic drive. A gearbox or an elec-
trical variable speed drive (VSD) is needed to reach the high speeds. Maximum
capacity is about 250,000 bpd, as measured at pump inlet conditions.
Erosion and Wear Prevention

As stated earlier, these pumps are mixed-flow centrifugal pumps. They are not posi-
tive displacement pumps. To minimize erosion, the pumps should be run at low
speeds. They are high-shear pumps, and, as such, tend to form emulsions.
The features of the helico-axial and twin-screw pump are compared and contrasted
in Figure 700-12.
Fig. 700-12 Comparison of Twin Screw and Helico-axial Multiphase Pumps For a Land Based Installation
Item Description Twin-Screw Helico-axial
Max Capacity 300,000 bpd 550,000 bpd
Relative Cost — 40-80% more expensive than the
twin-screw
Gas Limit • With liquid recirculation, 100% gas for For continuous service, 85% gas, except
(at suction conditions) an hour if special design features are
• Without recirculation, 100% gas for incorporated into the design for GVFs
about 20 minutes (confirm with vendor higher than 85%.*
for each particular model)
• For continuous service, 90% gas GVF
• Can be modified or specially designed
for gas slugs of some duration and for
95% and above GVFs*
Ability to Handle Slugs Excellent Usually requires a buffer tank ahead
of pump
Lube Oil System Self contained, each bearing Circulating system
Mechanical Seals 4 single seals requiring a clean, cool flush 1 dual and 1 single seal with a circulating
barrier oil
Ability to Process Sand Can handle moderate amounts but better if Poor
sand is suspended in a high viscous liquid
than if suspended in a low viscous liquid
Flow Range Wide range with speed changes Limited by the pump’s curve or envelope
Pressure Range Wide range-puts up whatever pressure is Limited by the pump’s curve or
required by system or as limited by driver envelope. Maximum differential pressure
provided. Maximum differential pressure is is about 2000 psi.
about 1500psi.
Relative Speed 900-3600 RPM Operates much higher RPMs than a twin-
screw
Relative Motor Size — Driver needs to be about 30% larger
Relative Efficiency 40-50% 20-30%
Relative Maintenance Costs Medium High
*Caution: It is difficult, almost impossible, to install a pump without some slugging. In a typical oil field, slugging
develops in the suction line from hilly terrain or from the oil well itself. Slugging causes the pump to momentarily
pump 100% gas, and without a steady supply of liquid to the pump, damage will occur. The pump may see an
average of 90% gas, but with slugging it will see 100% gas. It is always a good practice to feed liquid to the pump
from an outside source or from the discharge back to the pump inlet to insure that there is some liquid to the pump at
all times. If API Seal Flush Plan 32 is used for the mechanical seals, the flush fluid can be used to provide both seal
flushing liquid and the liquid required by the pump itself.
773 Progressive Cavity Pumps (PCPs)

The progressive cavity multiphase pump is a modified single-screw positive
displacement pump. See Figure 400-3 and Figure 700-9. The rotor or screw rotates
inside a case called the stator. Unlike the rotors in a twin-screw, with this pump, the
rotor contacts the stator and pushes the liquid from the inlet to outlet.
Stator Material
Traditional stators use an elastomeric material. The choice of the elastomeric stator
material is critical. If incorrectly chosen, the stators will fail because of entrained
gas from the multiphase stream permeating and swelling the elastomer. Recently
stators have been made of ceramic, but this also has shown to be a high mainte-
nance item. New elastomers are constantly being developed.
Note If this type of multiphase pump type is selected, an elastomer test should be
performed with the pumped fluid to determine the compatibility of the elastomer
with the fluid and the actual fluid temperatures and pressures.
Liquid Flush
Unlike twin-screw and helico-axial pumps that can run dry for short periods
providing the seals have an adequate source of liquid flush, PCPs can never be run
dry, even for a few seconds. It must be insured that the PCP always has liquid
entering the pump. This requires a downstream separator with sufficient volume to
supply liquid back to the pump inlet.
Capacity and Temperature

The maximum pump capacity is much lower than the twin-screw or helico-axial. It
is limited to around 20,000 bpd. PCPs can handle GVFs of about 65%, but not slugs
of gas without recirculating liquids, as discussed above. The permeability of the
elastomer to the gas in the pumped stream (as discussed above) is critical. The
pumps can handle particulate, but are often unreliable when they do. Their
maximum temperature depends on the elastomer chosen, but maximum tempera-
tures of 250 to 300°F are about as high as most manufacturers will guarantee. They
are used in surface and down-hole artificial lift applications with ESP motors or
surface motors.
Advantage
The one advantage of a PCP pump over the twin-screw or helico-axial pump is ease
of maintenance. The stator, when it fails, can be removed in the field saving a costly
and time consuming shop repair.
774 Piston Pumps

The piston pump uses a reciprocating piston design that incorporates two pistons
coupled out of phase by a common shaft. This design is a typical reciprocating
pump design in which the piston physically displaces the gas/liquid mixture on the
discharge stroke. Speed is usually very slow. Pump capacity is also very low.
Abstract
This section covers the fundamentals of mechanical seals, seal types and support
systems, seal application and selection, and troubleshooting procedures and
checklists. It briefly discusses packing. Much of this information is reproduced from
the Chevron USA (IMI) Mechanical Seal Manual.
Contents Page
810 General Information 800-3
811 Packing
813 Seal Leakage and Expected Seal Life
814 Recommended Services and Limitations
815 Converting Packed Pumps to Mechanical Seals
816 Seal Gaskets
820 Seal Types 800-9
821 Pusher Seals
822 Bellows Seals
823 Rotating and Stationary Seals
824 Balanced and Unbalanced Seals
825 Cartridge Seals
826 Inside Seals
827 Outside Seals
830 Support Systems 800-14
831 Flushing
832 Seal Flush Piping
833 Flush Plans
834 Quenching
February 2011 (E)  1999–2011 Chevron U.S.A. Inc. All rights reserved. 800-1
800 Mechanical Seals Pump Manual
835 Cooling
836 Pressurization
840 Dual Mechanical Seals 800-27
841 Introduction
842 Typical Applications For Dual Seals
843 Dual Seal Designs
844 Reliability and Experience
845 Costs
846 Buffer/Barrier Fluid Support Systems
860 Troubleshooting and Failure Analysis 800-52
861 Background
862 Elements of Successful Seal Performance
863 Mechanical Seal Failure Analysis
864 Seal Failure Troubleshooting
865 Solutions to Causes of Failure
Revision History 800-85
800-2  1999–2011 Chevron U.S.A. Inc. All rights reserved. February 2011 (E)
Pump Manual 800 Mechanical Seals
811 Packing
The most common source of trouble in centrifugal pumps is leakage of liquid along
the shaft. The stuffing box with soft packing, shown in Figure 800-1, is used to
prevent this leakage. The conventional stuffing box contains rings of a compressible
material such as woven carbon yarn that are held tightly against the shaft by
pressure from a gland. As illustrated in Figure 800-1, a lantern ring is frequently
placed near the center of the packing for introduction of a sealing or lubricating
medium. Usually, a removable throat bushing is positioned at the bottom of the
stuffing box which can be replaced if excessive clearance develops between the
bushing and the shaft. The bushing shown also has a bleed-off connection, used on
multistage pumps to keep the packing near suction pressure. Packing is now
considered obsolete for most applications.
Fig. 800-1 Conventional Stuffing Box

Mechanical shaft seals are frequently used instead of conventional packing and are
standard equipment for most centrifugal pump services. Basically, a mechanical seal
consists of a smooth flat surface mounted on the pump casing and a similar smooth
flat surface of different material rotating with the shaft but held closely against the
casing surface by spring and hydraulic pressure. Figure 800-2 is a cutaway view of a
typical single mechanical seal. Dual seal arrangements are also available (see
Figures 800-19 and 800-20).
Fig. 800-2 Typical Mechanical Seal
For satisfactory seal life the seal faces must be as flat and parallel as possible. The
surfaces are normally lapped to a flatness of one to three light bands (one light band
equals 11.6 millionths of an inch). One of the seal faces is held by a flexible
arrangement (usually springs) that keeps the two surfaces continuously in contact.
Enough pressure is maintained to exclude most foreign particles, but not enough to
exclude lubrication. The spring force also maintains face contact when the pump is
not running.
813 Seal Leakage and Expected Seal Life

Leakage from properly functioning mechanical seals is usually very slight—one or
two drops per minute at the maximum—but most commonly undetectable visually.
If mechanical seals operate satisfactorily when first installed, they will often do so
for a long time. A mechanical seal in continuous duty may be expected to operate
without requiring maintenance for several years and often for much longer. Seals in
intermittent services, as in bulk station plants, often operate for many years without
attention. The leakage that can be tolerated before repairing a seal depends on the
value and/or hazard of the liquid handled.
Mechanical seals are more vulnerable to premature failure when installed in a
system of new piping which has not been thoroughly purged of welding slag, mill
scale, and rust. Any system involving pumping equipment with mechanical seals
should be properly flushed and cleaned before initial operation.
814 Recommended Services and Limitations

In most services, mechanical seals have almost completely replaced packing. They
are particularly recommended for light hydrocarbon services, such as LPG, where
conventional packing has been a source of trouble. They are also highly satisfactory
in most clean hydrocarbon services. They require no adjustment or operator
attention until they begin to leak from excessive wear, and their normal life is
considerably longer than for soft packing. Because they are almost drop-tight,
mechanical seals are recommended for hazardous, corrosive, or costly liquids. They
are also preferred where atmospheric contamination must be avoided and where a
negligible level of leakage will reduce fire hazard.
Mechanical seals can be installed on pumps handling liquids containing abrasive
material in appreciable quantities if a separate clean lubricating fluid is injected over
the seal faces to flush the seal chamber so that abrasive material does not contact the
seal. Such flushing streams must be kept in operation whenever the pump is under
pressure, whether running or not. Services where loss of suction would allow the
seals to run dry, such as sump pumps, tank stripping, and similar applications, may
also require an external flush supply.
Mechanical seals are not recommended for stocks that become highly viscous or set
up on cooling unless special care is taken to adequately heat the seal area prior to
startup.
815 Converting Packed Pumps to Mechanical Seals

Consult the pump manufacturer before converting a packed pump to a mechanical
seal. Some older style packed pumps relied on the packing for additional shaft
support. If such a pump is not operating at its design point, the loss of support
around the shaft may result in excessive shaft deflection and radial vibration. When
converting packed pumps to mechanical seals, consider upgrading the pump to the
latest API requirements for shaft stiffness and stuffing box dimensions, by
purchasing a bearing bracket/stuffing box upgrade from the pump manufacturer.
To make the conversion from packing to a mechanical seal, the seal manufacturer
requires the following pump information. The Mechanical Seal Data Sheet in
API 682 may be useful in ordering parts when converting a packed pump to a
mechanical seal.
1. Serial number
2. Model
3. Size
4. RPM
5. Whether the case is split vertically or horizontally
6. Number of packing boxes
7. Shaft OD
8. Sleeve OD
9. Packing box depth
10. Maximum length of cartridge
11. Seal gland bolt circle
12. Number of seal gland bolts and location from 90-degree centerline
13. Size of seal gland studs
14. Relation of flush quench and drain port to vertical centerline
15. First obstruction
16. Whether pump loses suction
17. Whether pump cavitates
Items 6, 7, and 8 are necessary for the seal manufacturer to estimate the cost of the
seal. When the decision has been made to purchase the mechanical seal, the seal
manufacturer should have access to the original equipment manufacturer’s (OEM)
stuffing box drawings, but should fabricate the seal to accurate field dimensions.
Prior to fabricating the seal, it is good practice to compare the two and resolve any
discrepancies.
The end of the stuffing box (Figure 800-3) on an old pump is usually worn or bell-
mouthed from repeated packing replacements. To restore it to its original ID would
require welding and re-boring. It is usually less expensive to increase the OD of the
seal gland registered fit (Figure 800-4) by 1/8-inch over the dimension on the OEM
drawing. The stuffing box can be counterbored to fit the oversize seal gland
registered fit. This practice eliminates a welding repair.
Fig. 800-3 Bell Mouth Stuffing Box Fig. 800-4 Bell Mouth Repair
Items 9 and 11 can be established from the OEM drawings. Item 10 is determined
by the distance between the ends of the drive and driven shaft and must be taken in
the field. The first obstruction (item 15) is the distance between the stuffing box
face and the next object on the shaft, measured toward the bearing housing.
A packed pump will survive some instances of suction loss and cavitation (items 16
and 17), but either of these conditions will shorten mechanical seal life. When a
pump loses suction, the seal cavity empties and the loss of liquid between the seal
faces causes seal failure. Investigate the system for these problems prior to installing
a mechanical seal. If a suction loss problem does exist, it may be overcome with
low-flow or low-pressure instrumentation, which protects both the seal and the
pump. A close-clearance throat bushing and API 682 Flush Plan 32 can be used to
hold a positive pressure in the seal cavity and avoid seal failure during brief periods
of suction loss and cavitation.
816 Seal Gaskets

Leakage under the seal rings is prevented by gaskets. There are many seal gasketing
configurations using various materials, depending on the service. See Figure 800-5.
Fig. 800-5 Rotating Seal Member Gasket
API 682, and the petrochemical industry in general, has standardized on mechanical
seals that use only two of the gasket configurations shown in Figure 800-5, the
O-ring and the high temperature (flexible graphite) gasket. The other configurations
are designs that seal manufacturers devised to create successful gasket
configurations for Teflon (PTFE). Teflon has improved chemical resistance over
many elastomers, but causes corrosion fretting between the rotating seal member
gasket and the shaft sleeve O-ring. The Company has had poor experience with
Teflon secondary seals in any configuration.
When selecting a gasket, service temperature and chemical compatibility are the
predominant factors to consider. API 682 provides guidance for gasket selection
(see Figure 800-6). As shown, the basic gasket material selection for general
services is DuPont Viton. When chemical compatibility or temperature limitations
of Viton become a problem, DuPont Kalrez is recommended. In most hot services
(>300F) the Standard recommends flexible graphite gaskets. l
Fig. 800-6 API 682 Gasket Selections Courtesy of American Petroleum First
Service
Fluid Temperature Gasket Material Material Descriptions
Water <250F Fluoroelastomer (FKM) DuPont Viton
Sour Water <250F Perfluorelastomer (FFKM) ASTM D 1418, such as DuPont Kalrez
Caustic <180F Perfluorelastomer (FFKM) ASTM D 1418, such as DuPont Kalrez
Amines <180F Amine Resistant ASTM D 1418, such as DuPont Kalrez
Perfluorelastomer (FFKM)
Acids - H2SO4, HCl, H3PO4 <180F Perfluorelastomer (FFKM) ASTM D 1418, such as DuPont Kalrez
Hydrocarbon 20 to 300F Fluoroelastomer (FKM) DuPont Viton
Hydrocarbon containing 20 to 300F Perfluorelastomer (FFKM) ASTM D 1418, such as DuPont Kalrez
H2S or aromatics
Hydrocarbon containing 20 to 300F Perfluorelastomer (FFKM) ASTM D 1418, such as DuPont Kalrez
caustic
Hydrocarbon containing 20 to 300F Amine Resistant ASTM D 1418, such as DuPont Kalrez
amines Perfluorelastomer (FFKM)
Hydrocarbon -40 to 20F Nitrile B. F. Goodrich Hycar, Buna-N
Non-flashing hydrocarbon 300 F Flexible Graphite Union Carbide
(may contain H2S, Grafoil
aromatics, or amines)
Flashing hydrocarbon (may 300 to 500F Perfluorelastomer (FFKM) ASTM D 1418, such as DuPont Kalrez
contain H2S or aromatics)
Flashing hydrocarbon 300 to 500F Amine Resistant ASTM D 1418, such as DuPont Kalrez
containing amines) Perfluorelastomer (FFKM)
In common with many engineered components, the material composition and

manufacturing quality of O-rings varies greatly from one source to another. This is
of particular concern with O-rings because the various elastomers look similar. In
the past, attempts have been made to create a positive means to identify the different
O-rings such as through color coding. This was largely unsuccessful because there
was no standardization between suppliers and because, in some cases, the colored
material added to the O-rings had a detrimental effect on mechanical properties.
Figure 800-7 shows some of the current fluoroelastomer and perfluorelastomer
grades recommended for mechanical seals. This chart is based on Chevron Products
Company refining experience.
Fig. 800-7 Fluoroelastomer and Perfluorelastomer Grades

Gasket Material Description/Grade Comments
Fluoroelastomer DuPont Tracer Viton Can be positively identified under
ultra-violet light.
Perfluorelastomer DuPont Kalrez 4079 Greene Tweed Chemraz 505 is the
Greene Tweed Chemraz 505 standard for all services that require
perfluorelastomer gaskets at El
Segundo and Pascagoula refineries.
Amine resistant Perfluorelastomer DuPont Kalrez 1050LF or 3018 Greene Tweed Chemraz 505 is the
Greene Tweed Chemraz 505 standard for all services that require
perfluorelastomer gaskets at El
Segundo and Pascagoula refineries
The life of Fluoroelastomer (Viton) O-rings is not appreciably affected by

temperatures up to 350F. Above 350F, however, the life shortens quite rapidly.
Perfluorelastomer O-rings are selected for flashing hydrocarbon and water services
from 350°F to 450°F (API 682 indicates 300°F to 500°F). Generally, one of the
amine-resistant perfluorelastomer grades such as Kalrez 3018 is selected for these
services. Consult the seal supplier or an O-ring supplier for non-standard
applications.
820 Seal Types

821 Pusher Seals
Rotating pusher seals (Figure 800-8) operate as follows: As the rotating seal
member face (F) wears, the springs (D) push the face forward, also causing the
rotating seal member gasket (E) to be pushed forward on the sleeve or shaft. This
gasket, often called the secondary seal, is the seal between the rotating seal member
face and the sleeve or shaft. When the pump is operating, this gasket moves
constantly in an axial direction along the sleeve or shaft. Axial movement takes
place as the seal faces wear, or if the stationary face is not on a true 90-degree angle
to the pump shaft axis, in which case the gasket moves twice per shaft revolution.
The low cost of pusher seals makes them quite popular when the gasketing is not
affected by liquid temperature or chemical compatibility.
Fig. 800-8 Pusher Seal
822 Bellows Seals

In bellows seals (Figures 800-9 and 800-10), the rotating seal face is pushed
towards the stationary face by the combined force of the bellows and liquid
pressure. The rotating seal member gasket is located behind the bellows in a
position where it remains totally static. Unlike the pusher seal, the rotating seal
member in a bellows seal cannot hang up on the rotating seal member gasket.
Fig. 800-9 Bellows Seal (Elastometric) Fig. 800-10 Bellows Seal (Metal)
Bellows seals are somewhat pressure limited as compared to pusher seals. The seal
selection guide in API 682 limits selection of bellows seals to a sealing pressure of
250 psig. However, manufacturers generally design bellows for a pressure rating
exceeding 300 psig. Hence, 300 psig is more commonly regarded as the sealing
pressure limit for bellows seal selection.
Bellows seals are generally higher in cost than pusher seals. As a result, they are
generally selected for services where they provide extended life over pusher seals or
where pusher seals will not work. Metal bellows seals are particularly well suited to
high temperature service (350F to 800F). A flexible graphite gasket is usually
selected for the secondary sealing element in this temperature range.
Several materials have been used in the manufacture of metal bellows, including
316 SS, AM-350, and various Hastelloy and Inconel grades. The thin cross section
of the metal bellows requires a liquid corrosion rate of less than 2 mils per year.
Also, susceptibility to sulfide cracking, chloride cracking, or hydrogen
embrittlement are factors to consider in bellows material selection. The Company
has had poor experience with AM-350 metal bellows. API 682 has standardized on
two metal bellows materials, Hastelloy C for the Type B seal and Inconel 718 for
the Type C seal. These selections are based on industry experience more than on
cost. Cost is really a secondary factor because the money saved by selecting a less
expensive bellows is insignificant compared to the cost of a typical seal repair.
Fatigue is also a common failure mode in metal bellows. These failures often occur
in the first or last convolution in the bellows, when loss of lubrication between the
seal faces causes the faces to “grab”, resulting in high cyclic torque loads on the
bellows. It is for this reason that one avoids a metal bellows seal selection for
flashing services or services with an inadequate vapor suppression margin.
Appendix F contains a table of metal bellows failures documented in Chevron
facilities.
823 Rotating and Stationary Seals

Most mechanical seals are of rotating design (see Detail A, Figure 800-11). When a
rotating seal is operating at or above 5000 sfpm (surface feet per minute), some seal
distortion and dynamic balance problems may develop. These are eliminated with
the use of a stationary seal. The flexible seal member becomes the stationary seal
member. The rotating seal member in the stationary seal design is solid and not
subject to distortion or balance problems.
Fig. 800-11 Rotating and Stationary Seals
Stationary seals are most commonly used in pumps and compressors that operate at
high speeds. The stationary design is frequently used in high-temperature metal
bellows seals because they will accept more axial movement of the shaft and
vertical misalignment of the seal endplate. Also, the stationary bellows seal has
more clearance between its inside diameter and the shaft sleeve than does a rotating
seal. This allows better quenching and less seal hang-up from coking. During the
assembly of a rotating seal, the seal endplate that holds the stationary seal member
is often misaligned from a true 90-degree angle to the shaft. This misalignment
requires the rotating seal member face to move back and forth in an axial direction,
twice per shaft revolution. This movement can lead to premature failure of a pusher
seal’s rotating seal member gasket. It will also contribute to corrosion fretting of the
shaft sleeve.
When the misalignment of the seal endplate is excessive (greater than 0.003 inch),
the axial movement of the rotating seal member causes the rotating seal member
face to separate or float away from the stationary seal member face. This results in
excessive face leakage. In a stationary seal, the seal member is flexible, due to its
springs or bellows. Since it does not rotate, this flexibility compensates for the seal
endplate misalignment and allows the stationary seal member face to conform to the
true-running rotating seal member face.
824 Balanced and Unbalanced Seals

Seal cavity pressure on the back of the unbalanced seal member face in
Figure 800-12 applies a force to the opposite face. If this force is excessive, the heat
generated between the faces causes the face lubricant film to vaporize. If the back of
the seal member face is reduced in area, the same seal cavity pressure will apply
less force to the opposite face. This is done by removing material from the inside
diameter of the seal member face. The removed material is replaced by a step in the
shaft sleeve. This step is commonly termed the “balance shoulder.” The amount of
balance the manufacturer builds into the seal is determined by the operating
conditions and liquid characteristics.
Some users specify balanced seals for all pressures, speeds, and specific gravities.
This practice aids in parts standardization and is a precaution against installing an
unbalanced seal in a service where it will cause a premature seal failure. All of the
standard seal types in API 682 are balanced.
Off-the-shelf pumps used in low-pressure, low-speed, high specific gravity service
often have unbalanced seals. The added cost and delivery time needed to change the
unbalanced seal to a balanced seal cannot be justified.
Fig. 800-12 Unbalanced and Balanced Seals
825 Cartridge Seals

Cartridge seals (Figure 800-13) are required for all seals in accordance with
API 682. A cartridge seal is a complete seal assembly, including shaft sleeve, that
can be assembled external to the pump. The seal drive set screws are mounted
outside the seal cavity. Cartridge seals are available in all API 682 configurations,
including the dual seal configurations. Cartridge seals have several advantages:
• The cost of a seal repair is lower and less skill is required.
• The entire seal can be assembled in a shop environment where it is more likely
to be done correctly.
• The seal assembly can be given a bench pressure test with air (as required by
API 682) to ensure that the faces and gaskets are assembled correctly and are
functional.
• Cartridge seals can be qualification tested as required by API 682.
• The seal manufacturer has unit responsibility for the seal design.
Cartridge seals are generally recommended for almost all seal installations. Again,
they are required for all seals in accordance with API 682. This includes most
centrifugal and rotary pumps in accordance with API 610 and 676. In some small,
standard, or inexpensive pumps, cartridge seals may not be practical or their
incremental cost may not be justified. Also, some small and/or standard pumps are
close coupled or have short coupling spacers which do not allow for sufficient axial
space for installing a cartridge seal.
Fig. 800-13 Cartridge Seal
826 Inside Seals

An inside seal is any seal that is mounted inside the seal cavity. The hydraulic
pressure in the seal cavity has a closing effect on the seal face that tends to decrease
the leakage between the faces. Abrasives tend to be thrown away from the faces by
centrifugal force. The majority of mechanical seals are inside mounted.
827 Outside Seals

An outside seal is any seal that is mounted outside the seal cavity. The pressure in
the seal cavity has an opening effect on outside mounted seals. This opening effect,
combined with the centrifugal force of the liquid moving outward between the seal
faces, increases the leakage rate. Abrasives also tend to be centrifuged into the seal
faces, causing increased face wear. Leakage from a catastrophic outside mounted
seal failure will be more pronounced than leakage from the same catastrophic inside
seal failure.
Outside seals are generally not used in any service.
830 Support Systems

Mechanical seals work very well when constantly surrounded by a fluid that is
clean, cool, non-corrosive, and of the right viscosity.
Unfortunately, this seldom happens naturally, so it is often necessary to include a
support system to create a supportive environment for the seal. Usually this involves
some sort of flush to cool, purge, or pressurize the seal chamber and perhaps a
quench to retard the formation of solids on the atmospheric side of the seal faces.
The standard flushing arrangements have numbered flush plans as described and
illustrated in API 682. These plans describe the auxiliary equipment required to
ensure that fluid at the seal faces is maintained at the proper conditions.
Dual seals have buffer or barrier fluid support systems, which are more complicated
flush arrangements as required to support the outer seal. A buffer fluid is a fluid
injected in between dual mechanical seals, at a lower pressure than the pumped fluid
sealing pressure. A barrier fluid is also a fluid injected in between dual mechanical
seals, but at a higher pressure than the pumped fluid sealing pressure. In both cases,
the buffer or barrier fluid is injected between the primary and outer seals.
831 Flushing
A flush is a small amount of fluid that is introduced into the seal chamber close to
the sealing faces. The flush fluid improves the fluid conditions near the faces.
Usually, the main purpose of the flush is to suppress vapor formation at or near the
faces by heat removal and pressurization.
Guidance for selecting a flush plan is given in Section 850 and in Appendix B of
API 682. The following discussion is intended to provide a better understanding of
how flush systems are designed. This information should enable one to ask the right
questions for selections not covered in the guideline.
832 Seal Flush Piping

Controversy often exists about seal flush piping. Should it be stainless steel? Is seal
welding required? Where should the orifices be? The following recommendations
are a proven approach to seal flush systems.
For LPG or highly toxic services, or for services with pumping temperatures over
450F, an orifice (for restricting flow to the atmosphere in case of tubing failure)
should be provided at the discharge or suction nozzle connection. This orifice is in
addition to the orifice at the gland plate connection, and must be considered in seal
flush flow calculations. Flush and quench lines should be Type 316 stainless steel
tubing, ¾-inch by 0.095-inch minimum wall thickness or ½-inch by 0.065-inch
minimum wall thickness with stainless steel fittings.
833 Flush Plans

The following flush plans are shown and described in Annex D of API Standard
682, “Pumps-Shaft Sealing Systems for Centrifugal and Rotary Pumps.”
Flush Plan 01 (or ASME Plan 7301) takes liquid from the pump discharge, reduces
the pressure through an orifice and injects it into the seal cavity through an internal
passageway. The liquid then flows from the seal cavity to the back side of the
impeller. This prevents liquid stagnating in the seal cavity, reducing heat and the
buildup of contaminants.
Seal cavities for these pumps are often termed “internal seal cavities.” The internal
seal cavity is a large void directly behind the impeller; it does not have a throat
bushing.
Flush Plan 01 allows this pump to be built shorter with a more rigid shaft. If this
style of pump develops premature seal failure, the lack of a throat bushing makes it
difficult to control the environment in the seal cavity. To pressure or flush the seal
from an exterior source would require a larger seal chamber.
Pumps with internal seal cavities should be considered only for the following
services:
• Liquids with low vapor pressures
• Nonabrasive liquids
• Noncorrosive liquids
• Liquids with good lubricating ability
• Stable suction heads
• Flows that are always above the pump’s minimum flow
Pumps with internal seal cavities should not be considered for dirty, gaseous, or
abrasive service. This seal plan can be used with liquids that have increased
viscosity, or are at risk of freezing, at ambient temperatures.
Flush Plan 02 (or ASME 7302) has a dead-end seal chamber with no circulation of
flushing liquid, but it has plugged connections in the seal chamber for possible
future circulating liquid. Plan 02 allows the purchaser to specify a water-cooled
stuffing box.
If the liquid in the seal cavity is too close to its vapor pressure, this flush plan may
succeed in cooling it below its boiling point. However, Plan 02 should be used with
clean fluids with high specific heats, such as water, in relatively slow pumps.
Caution should be used if used in light hydrocarbon service where a slight buildup
in heat will cause vaporization.
In some older pumps, water circulating through the seal cavity water jackets
removes enough heat from the liquid in the seal cavity to keep the liquid below its
boiling point. Contact the pump manufacturer to establish the efficiency of the
pump’s water jacket.
Over time, contaminants will plate out and foul the water jacket walls. This fouling
will reduce heat transfer from the seal cavity liquid to the cooling water. The water
jackets should be inspected for fouling during routine plant shutdowns, and
chemically cleaned if necessary. The industry trend is away from water jackets.
Plan 02 is not very efficient at heat removal with pumps that have internal seal
cavities (API Plan 01), due to the lack of the throat bushing. The throat bushing acts
as a heat barrier between the cooled liquid in the seal cavity and the hot liquid
behind the impeller.
Plan 02 requires that the seal cavity be full of liquid. This requires installing a vent
at the top of the seal cavity.
Flush Plan 11 (or ASME Plan 7311) passes liquid from the pump discharge through
an orifice, through the seal chamber, and into the back side or low-pressure area of
the impeller. The amount of flush is usually determined by the seal or pump
manufacturer. If it becomes necessary to field-calculate the flush rate, the rule of
thumb is 0.75 gpm per inch of seal diameter for non-regulated emission services or
1.5 gpm per inch of seal diameter for services with regulated emissions. A
distributed seal flush system such as a circumferential or multi-port arrangement is
usually required to develop the higher flush rate.
Plan 11 avoids stagnation in the seal cavity, reducing heat and contaminant buildup.
Plan 11 is the most common Flush Plan, but it should be used only when seal
chamber pressure is less than discharge pressure. In single stage overhung pumps,
this requires the impeller to have balance ports and a back wear ring or pump-out
vanes.
An orifice is required in the flush line at the discharge connection for the following
reasons:
1. If the flush tubing fails, the amount of leakage to atmosphere is restricted by the
orifice.
2. Taking a pressure drop through an orifice located upstream at the flush source
decreases the velocity of the liquid entering the seal cavity.
Without the orifice, the high pressure drop across the orifice in the gland plate flush
port could cause excessive erosion of the seal parts or force the seal faces open.
It is poor practice to use orifices less than 1/8-inch because of plugging. If a
1/8-inch orifice is insufficient to produce the desired pressure drop, use multiple,
larger orifices instead of a single 1/8-inch diameter orifice.
The discharge flush also suppresses vaporization of the liquid in the seal cavity.
Installing a close-clearance throat bushing between the seal cavity and impeller area
restricts the flow from the pump’s discharge and raises the seal cavity pressure. This
pressure rise suppresses vaporization; however, note that reducing the flow may also
cause a reduction in cooling which may cause seal failure due to overheating.
Plan 11 is not used in dirty or abrasive service.
Flush Plan 12 (or ASME Plan 7312) is identical to Flush Plan 11 except for an
added Y-strainer in the flush line upstream of the orifice. The strainer’s purpose is to
trap any foreign material that could plug the orifice. If the strainer openings are too
small and catch particles, the strainer can plug and stop the flow of flush liquid to
the seal. If the stainer openings are too large, the stainer has little or no value for
protecting the seal.
Note This flush plan is not recommended.
If a strainer or filter must be used in a flush line, one of the following precautions
should be taken:
1. Establish a strainer blowdown schedule.
2. Instrument the flush system to alarm if flush flow is lost.
3. If filters are used, they should be instrumented for a differential pressure
increase across the filter. Filters should be arranged in pairs so a plugged filter
may be cleaned while a clean filter is in service.
Note that API Flush Plan 12 has been deleted as a referenced plan throughout
API 682.
Flush Plan 13 (or ASME Plan 7313) provides circulation from the seal cavity
through an orifice back to pump suction. Plan 13 is used when the pressure in the
seal cavity is the same as the pump’s discharge pressure.
Pump discharge pressure in the seal cavity is determined by the pump’s impeller
design. An impeller in a single stage overhung pump that has no balance ports, no
back wear ring, no back pumpout vanes, or is a reverse vane impeller, will cause the
seal cavity to be at discharge pressure. The seal cavity in most vertical pumps
operates at pump discharge pressure.
Plan 13 reduces pressure in the seal cavity. This flush also removes the heat
generated by the seal faces.
Caution must be taken when sizing the flush line orifice, especially in high head
pumps. If the rate of flush flow is excessive the seal cavity pressure will drop and it
could become too close to the vapor pressure of the liquid. If this occurs, the liquid
film between the seal faces could flash to vapor, resulting in a lack of seal face
lubrication and premature seal failure.
In vertical turbine pumps, API Plans 1, 11, 12, 21, 22, 31,or 41 must be used in
conjunction with API Plan 13.
Flush Plan 14 (or ASME Plan 7314) is a combination of Plan 11 and 13, where the
liquid from the pump discharge flows through an orifice to the seal and back to the
suction nozzle. It is recommended for light hydrocarbon liquids and commonly used
on vertical pumps. This plan allows cooling flow to be supplied to the seal chamber
(Plan 11) while providing complete venting of the seal chamber (Plan 13).
Flush Plan 21 (or ASME Plan 7321) takes fluid from the pump discharge through
an orifice and heat exchanger and then injects the fluid into the seal cavity. This is
similar to Plan 11 with a heat exchanger to cool the liquid going to the seal cavity
below its boiling point. The cooler can use water or air as the cooling medium,
depending on the amount of heat transfer required and the efficiency of the cooler.
Once it has been determined how much heat must be removed from the liquid
flushing the seal cavity, the cooler manufacturer can calculate the cooler and flush
line orifice size.
This flush plan works well for low temperature gradients across the cooler. With
moderate to high temperature gradients, cooler fouling becomes a problem.
Consider Flush Plan 23 for applications with moderate to high temperature
gradients.
If pumping hot water at sea level, cooling the seal flush is recommended at
temperatures above 150F. This temperature should be lowered at higher elevations
to maintain the temperature in the seal chamber approximately 70F below the
atmospheric boiling temperature.
Flush Plan 22 (or ASME Plan 7322) is the same as Flush Plan 21, with the addition
of a Y-strainer in the flush line upstream of the orifice.
The strainer has little or no value for protecting the seal other than keeping the
orifice from plugging. The strainer itself could plug, decrease the flush flow rate and
damage the seal. If a strainer is used, precautions must be taken against plugging.
Flush Plan 23 (or ASME Plan 7323) incorporates a pumping ring mounted in the
seal to provide recirculation of the liquid from the seal cavity through a heat
exchanger and back to the seal.
If the liquid in the seal cavity is too close to its vapor pressure and flashes to a vapor
between the seal faces, then the seal will fail prematurely. Plan 23 attempts to avoid
this by cooling the liquid in the seal cavity below its boiling point.
Flush Plan 23 is identical to Flush Plan 02 with the addition of a flush line, elevated
flush cooler and pumping ring. Fluid circulation can also be assisted by thermo-
siphon effect in a properly designed system.
The liquid head developed by the pumping ring is not always sufficient to pump air
bubbles in the flush line. Therefore all air must first be vented from the flush line.
The vent must be located at the highest point in the flush line.
Flush Plan 23 is used mainly to cool seals in boiler feedwater and flashing
hydrocarbon services. Above a temperature of 180F, water cannot be sealed
successfully with a standard single seal. This is because boiler feedwater is
contaminated with soluble abrasives, namely minerals and boiler compounds. As
the pressure of the film of water between the seal faces drops from seal cavity
pressure to atmospheric pressure, it will flash to a vapor if the temperature between
the faces equals or exceeds 212F. If this happens the abrasives come out of solution
and embed themselves in the softer of the two faces. The soft face then acts like a
grinding wheel, quickly destroying the opposing seal face. To keep the liquid
between the seal faces below 212F, it is necessary to cool the seal cavity to 180F
or less, preferably 150F. Also, the lubricity of water drops substantially as it gets
close to its boiling point.
Once the liquid in the seal cavity has been cooled to 150F, the amount of heat
transfer required to hold it there is minimal, and cooler fouling is also minimized.
Cooler fouling occurs if minerals from the cooling water plate out on the cooling
water side of the cooler’s tube, and is more pronounced if the heat transfer rate is
high. A drain is required on the shell of the cooler for periodic back flushing of the
cooler’s shell side.
Plan 23 has an advantage over Plan 21 because slightly less power is consumed, the
cooled flush stream does not go back into the process stream, and less cooling water
is needed in the heat exchanger because the cooler removes only seal face generated
heat.
The temperature of the flush stream should be monitored with a temperature
indicator installed in the flush line between the seal cavity and the cooler. If the
cooler loses its efficiency, it will require cleaning by chemical or mechanical means.
Be careful not to restrict flow by over-instrumenting the flush line.
If the pumped fluid is a hazard to personnel, all pressure indicators used in the flush
system must have block valves and all temperature indicators must have
thermowells.
Flush Plan 31 (or ASME Plan 7331) is used in dirty, abrasive services. It takes the
dirty, abrasive process liquid from the pump discharge into a cyclone (inertial
separator). Clean liquid comes off the top through an orifice, to the seal cavity, and
then to the low pressure or back side of the impeller.
Abrasives are detrimental to the seal faces.
In Figure 800-14, the abrasive contaminated liquid from the pump discharge enters
the cyclone at A. The action of the cyclone forces the abrasives to exit at C and
return to the pump’s suction. The clean fluid exits into the flush line at B.
Fig. 800-14 Cyclone Separator
Cyclone separators are not effective at removing solids that are nearly the same
density as the fluid. API 682 requires the use of long radius and 45 bends in lieu of
90 short radius elbows. Lines must slope up to the cooler at ½-inch per foot,
minimum. API 682 requires that the density of the solid particles is at least two
times the fluid density. The cyclone separator must have a differential pressure of
2 to 10 atmospheres between connections A and B. Also, it will only work properly
if the pressure at connections B and C are at or near the same pressure. This limits
the length of the piping from connection C back to pump suction.
Flush plans involving cyclone separators are generally not recommended for sealing
applications in Chevron facilities. Although they might successfully remove large
particles that cause damage in the seal cavity, cyclone separators have some of the
following problems:
1. To achieve maximum abrasive removal, detailed engineering is required for
orifices in the cyclone system.
2. Abrasives at high velocities erode cyclone orifices, causing efficiency loss.
Eventually, they will stop working entirely.
3. Orifice plugging causes reduction or loss of the seal flush. If the cyclone plugs,
the seal flush flow reverses, flowing from the seal cavity to the pump’s suction.
If the liquid in the seal cavity is near its vapor pressure, the drop in seal cavity
pressure could cause flashing between the seal faces and failure of the seal.
4. The smaller the abrasive particle, the less efficient the cyclone separator.
Unfortunately it is the small particles that enter between the seal faces.
Some pump manufacturers offer an optional internal inertial separator with internal
passageways that eliminate costly external piping/tubing, and can be used in lieu of
Plan 11.
An external cyclone may be the best choice if contaminants are excessive or
extremely abrasive. In such a case, one can expect frequent cleaning or replacement
of the cyclone and it may be a good idea to install pressure gages on both sides of
the cyclone.
Flush Plan 32 (or ASME Plan 7332) is used to inject a clean cool liquid to the seal
from an external source. The flush flows through the seal cavity to the low pressure
or back side of the impeller.
To allow for flush flow, the pressure of the injection source must be greater than the
pressure in the seal cavity. The injection source must also be compatible with the
liquid being pumped, because it exits via the pump’s discharge. Flush Plan 32
should be considered for the following conditions:
1. If sealing pumped liquids that contain abrasives, injecting a clean cool fluid that
is compatible with the pumped liquid is a straightforward approach to eliminate
abrasives from the seal cavity.
2. If sealing pumped liquids that are hazardous to personnel or violate a pollution
standard if allowed to leak to the atmosphere. An example is a stream with a
high concentration of H2S. Flush Plan 32 isolates the seal cavity from the H2S
with a non-H2S bearing flush stream. If the seal fails catastrophically, the leak
to atmosphere through the seal endplate throttle bushing must be less than the
non-hazardous flush flow to the seal chamber. A close-clearance throttle
bushing in the seal endplate is required to attain this leakage rate.
3. If the liquid in the seal cavity is near its boiling point and will vaporize between
the seal faces, the seal will fail prematurely. This can be overcome by injecting
a liquid with a low vapor pressure, which is compatible with the pumped liquid,
into the seal chamber.
4. If a pump loses suction or has problems picking up suction, the seal fails
prematurely from lack of liquid lubrication. Flushing the seal with Plan 32
provides a constant source of liquid during brief periods of suction loss.
Flush flow velocities of 10 to 15 feet per second through the seal cavity throat
bushing are necessary to keep the pumped liquid from migrating back into the seal
cavity. If the exterior flush flow rate must be held to a minimum, it is necessary to
install a close-clearance throat bushing to maintain this flow velocity. The flush
flow rate can be controlled with an orifice or with visual flow indicators. Orifice
sizes below 1/8-inch are subject to plugging and not recommended. Instead use
multiple 1/8" or larger orifices in series. Flow meters are desirable because the
operator can visually monitor the flow. API 682 requires the piping arrangement to
include provisions to measure both the flush source pressure and the seal chamber
pressure. All plan 32 systems should include a block valve and check valve to
reduce the possibility of back-flow to the flush source in the event that pressure is
lost at the source. If the pumped liquid is a hazard to personnel, all pressure
indicators used in the flush system must have block valves and all temperature
indicators must have thermowells.
Flush Plan 32 often involves the downgrade of higher value fluids to lower value
fluids, process contamination, or creation of additional fluid requiring treatment
(such as sour water). These costs can be quite significant and should be evaluated
before selecting this flush plan. One Chevron refinery determined that their total
downgrade costs for Plan 32 systems exceeded $2MM per year. Eliminating Plan 32
systems became a source of significant operating cost savings for this refinery.
Flush Plan 32 is often used in two-screw pumps in multiphase service. The pumped
fluid in this service can contain as much as 100% hydrocarbon gas that, if allowed
to enter the seal chamber, would cause the seal to fail immediately. An external
source of clean liquid injected into the seal chamber cools the faces of the seal.
Chevron has had good success with this type of flush plan in multiphase service.
Flush Plan 41 (or ASME Plan 7341) is the same as Plan 31, except that a heat
exchanger is used to cool the flush going to the seal cavity. It is designed to remove
abrasives and to cool the flush stream going to the mechanical seal. If the pumped
fluid is a hazard to personnel all pressure indicators used in the flush system must
have block valves and all temperature indicators must have thermowells. Steam is
often used as the quenching fluid.
Flush Plan 51 provides a quench on the low pressure side of the seal. The quench is
supplied from a reservoir. The quench fluid may be water. Also see Flush Plans 61
and 62.
Flush Plan 52 provides for circulation of a buffer fluid between unpressurized dual
seals. This flush plan and its support system is described in more detail in Section
846.
Flush Plans 53 and 54 provide for circulation of a barrier fluid between
pressurized dual seals. These flush plans and their support systems are described in
more detail in Section 846.
Flush Plan 61 (or ASME Plan 7361) provides plugged connections in the gland
plate to allow for future use of an API Plan 62.
Flush Plan 62 (or ASME Plan 7362) allows an external fluid quench (steam, gas,
water, etc.) between the outer seal and the throttle bushing or auxiliary sealing
device. Refer to Section 834, “Quenching” for more details.
Flush Plan 71 is used on Arrangement 2 seals which use a dry containment seal and
in which a dry buffer gas is not supplied initially at seal installation but which may
be supplied in the future.
Flush Plan 72 is used on Arrangement 2 seals that use a dry containment seal and
in which the buffer gas is used to sweep away leakage from the inner seal into the
containment seal chamber (the chamber between the inner and outer seals). It is
often used where emissions from the pumped fluid need to meet some prescribed
limitation and therefore needs to be detected and alarmed prior to a more serious
failure.
Flush Plan 74 is used on Arrangement 3 seals which use a gas barrier fluid. It is a
analogous to a Flush Plan 54 where the barrier fluid is a liquid. The most common
gas barrier fluid is nitrogen and it is operated at pressure of 25 psi above the seal
chamber pressure.
Flush Plan 75 is used on Arrangement 2 seals where a dry containment seal is used
where leakage from the inner seal may condense. The plan can be used with Flush
Plan 71 or 72 but has an external condensate collector vessel is used to collect the
condensing vapors coming from the inner seal.
Flush Plan 76 is used on Arrangement 2 seals which use a dry containment seal
where leakage from the inner seal will not condense. The plan can be used with
Flush Plan 71 or 72 but has an external collection and removal system for the non-
condensing vapors.
Additional Flush Plans. Old pump installations that have been converted from
packing to mechanical seals occasionally have seal problems caused by pump
suction loss. Figure 800-15 shows two flush plans (A and B) that supply liquid to
the seal during brief periods of suction loss. These flush plans should not be
considered when designing a new pump installation. Properly designed pump
installations warn plant operators in advance of conditions that cause pump suction
loss.
Fig. 800-15 Flush Plans
Flush Plan A requires a discharge line with an elevation higher than the throat
bushing in the bottom of the seal cavity. The close-clearance throat bushing restricts
the flow leaving the seal cavity, and allows for a longer retention time of the liquid
gravitating back from the discharge line.
Flush Plan B depicts an external flush that is compatible with the pump’s liquid
stream. Sometimes it is uneconomical to use an external flush continuously, yet
periodic suction loss requires the external flush to be available to keep a flush flow
of liquid going to the seal cavity. Flush Plan B is useful in such cases.
The external flush in Flush Plan B must have less pressure than the pump’s
discharge, but more pressure than the seal cavity. When the pump is operating at
normal rates, check valve No. 2 is open, check valve No. 1 is closed, and the seal is
flushed from the pump’s discharge. If the pump loses suction, the discharge pressure
drops, check valve No. 1 opens, check valve No. 2 closes, and the seal is flushed
from the external source.
834 Quenching
The seal quench shown in Figure 800-16 is used to treat leakage that comes past the
mechanical seal into the atmosphere. The quench port enters the mechanical seal
endplate on the atmospheric side of the seal faces. Fluid injected into the quench
port leaks to the atmosphere through the seal endplate throttle bushing or the seal
endplate drain.
Quenching is used primarily if the pumped fluid will build up coke or salt on the
low pressure side of the seal. This coke or salt can get between the seal faces, open
them, and cause leakage. The quench fluid is used to remove the deposits.
Quenching is also used to cool the minute leakage that can occur across the seal
faces especially in hot oil services.
Fig. 800-16 Seal Quench
Water Quenching. When some liquids evaporate, they leave an abrasive crystal
deposit behind (caustic, for example). Crystals forming on the atmospheric side of
the secondary seal stop the secondary and rotating members from moving forward.
This condition is commonly called seal hangup. The seal faces are also damaged by
abrasives and crystals centrifuged between them.
These problems can be avoided by using API 682 Flush (quench) Plan 62 and
quenching the seal with water. The water dilutes the seal leakage and stops the
formation of the crystals. The quench water exits through the seal endplate throttle
bushing or drain.
Maximum quench flow rates should be low enough to keep the quench water from
entering and contaminating the pump bearing housing. Quenches can be used
continuously or intermittently depending on the situation.
Steam Quenching. When hydrocarbons are sealed at elevated temperatures, coke
forms on the atmospheric side of the seal as a result of normal seal leakage. This
coke formation causes seal hangup and face damage. Steam injected through the
seal endplate quench port per API 682 Flush (quench) Plan 62 aids in stopping the
coke formation by cooling and removing the leakage.
The steam quench is also used to decrease blistering of the carbon seal face.
Blistering occurs when sealing viscous fluids such as asphalt, crude and bunker.
When the liquid is warm it penetrates into the face of the carbon seal. If the liquid
film between the seal faces is allowed to cool when the pump is shut down, it
becomes viscous. When the pump is restarted the shear force of the viscous liquid
film between the seal faces generates heat. The heat expands the liquid that has
penetrated the carbon face. As a result of the liquid expansion small pieces of
carbon raise or break away from the carbon face. These pieces of carbon hold the
faces open and allow excessive leakage. A controlled steam quench keeps the faces
warm, while the pump is out of service.
Quench Rate. The maximum steam quench rate should be low enough to keep the
steam from entering and contaminating the pump bearing housing. The amount of
steam quench should be controlled by a back pressure regulator. Once the back
pressure regulator is adjusted to the desired flow, it remains constant. This
overcomes the problem of plant operators continually readjusting the amount of
steam flow. Any pressure regulator comparable to the Fisher Type 95L with the
following construction features can be used: body size, ¼-inch; orifice size, ¼-inch;
body material, cast steel; diaphragm and inner valve seat material, stainless steel;
maximum inlet pressure and temperature, 300 psi, 450F. Reduced pressure ranges,
2 to 6 psi.
Dry Steam. The problem with steam-quenching high-temperature seals is keeping
the quench dry. When condensate enters the atmosphere side of a seal operating in a
high-temperature service and flashes to steam, the expansion of condensate to steam
opens the seal faces and allows excessive leakage. This produces a popping sound
from the seal endplate.
When installing a steam quench system, insulate the line from the steam source to
the seal end plate. Figure 800-17 is a steam quench piping system that is successful
in high-temperature applications.
Fig. 800-17 Steam Quench Piping System
High-temperature seal leaks can be a fire hazard if the temperature of the liquid
entering the atmosphere is at its flash point. The steam quench isolates and cools the
leakage below its flash point before it enters atmosphere. A close clearance throttle
bushing aids in isolating a hazardous seal leak.
Steam Smothering Ring. Every precaution should be taken against pump fires
when sealing a high-temperature liquid that is near its flash point. The steam
smother ring (Figure 800-18) is an inexpensive tool to prevent igniting a fire when
flammable stock has leaked past the steam quench and the seal endplate throttle
bushing. The valve controlling the steam source to the smothering ring should be
well marked and at a safe distance from the pump’s seal cavity.
Fig. 800-18 Steam Smother Ring
835 Cooling
Seals require very little lubrication. Stocks such as gasoline, propane, and butane
provide sufficient lubrication if they are kept under sufficient pressure and the seal
faces are adequately cooled. Seal face cooling is usually achieved by circulation of
fluid in the mechanical seal chamber. Inadequate cooling causes overheating of the
seal faces, and vaporizing of the stock around the seals, with loss of lubrication and
rapid seal failure. The required flow is normally determined by the seal and pump
manufacturer. Usually, the pumped liquid is the lubricant and coolant. The poorer
the stock’s lubrication qualities and the closer it is to its vapor pressure, the more
important it is to maintain an ample cooling flow.
Stock vaporization at the seal faces is the most common cause of seal failures.
For this reason it is advisable to use a cooler (such as in API 682 Flush Plan 21 or
23) whenever there is a risk of flashing. Keep in mind that many stocks contain a
mixture of compounds; the compound with the highest vapor pressure (lowest
boiling point) must be addressed when designing a cooling system.
Cooling Requirements. Lower temperatures are needed for high vapor pressure
stocks such as light hydrocarbons and ammonia. The amount of cooling needed is
usually specified in terms of vapor pressure rather then temperature.
Light hydrocarbons (0.6 SG and less) and ammonia should be cooled to a
temperature such that the vapor pressure is 50 psi less than the seal chamber
pressure. For other stocks the recommended margin is 25 psi.
836 Pressurization
Cooling is always preferable to pressurization to suppress vaporization at the seal
faces, but cooling is not always feasible. Often one must raise the pressure in the
seal chamber to create the necessary margin between vapor pressure (at seal
chamber temperature) and seal chamber pressure. Follow the same rules outlined
above (i.e., 50 psi for ammonia and light hydrocarbons).
The source of pressure may be from the pump discharge (API Plan 11, 21, etc.) or
from an external source. Usually the throat bushing clearance must be reduced to
create the increased seal chamber pressure. For very close clearances a floating
carbon throat bushing is recommended.
Some locations have used hardened throat bushings, such as Nitronic 60, to reduce
erosion. The hardness must be controlled to prevent galling with the shaft.
840 Dual Mechanical Seals

Although a small percentage of pumps in Chevron facilities have dual mechanical
seals, concerns about poor dual seal reliability are both widespread and justified.
Pressurized dual seals are significantly more troublesome than unpressurized dual
seals, and there are examples of Company experiences where they failed ten to
twenty times during attempted start-ups.
Unpressurized dual seals are being operated with success, but single seals should
always be considered first. A decision to select dual seals should only be made
based on thorough knowledge of their pitfalls.
Readers are particularly cautioned against choosing dual seals just to meet
1000 ppm emission limits or less. Well-designed single seals can easily comply with
1000 ppm limits for Volatile Organic Compounds (VOC) and Reactive Organic
Gases (ROG).
Single seals are also preferable because:
• They are simpler
• They do not require a buffer or barrier fluid system
• Assembly is less complex
• Operation is easier
• Pumping rings are usually not required
• They are unaffected by pressure reversals
• There is no need to properly handle contaminated and/or hazardous
buffer/barrier fluid when performing maintenance.
841 Introduction
Dual Mechanical Seals
Use of the term “dual” in the phrase dual mechanical seals indicates that there is
more than one seal in a seal chamber at a given shaft sealing location in a machine.
The use of more than two seals in a dual seal is fairly unusual, so the word dual
usually refers to seals with two pairs of seal faces. The term “dual” is consistent
with terminology as defined and used in API 682. Two types of dual mechanical
seals are defined in API 682:
• Dual mechanical seal with unpressurized buffer—referred to herein as an
unpressurized dual seal.
The seals are mounted in series in an unpressurized dual seal. Before API 682
was published, this seal configuration was referred to as a tandem seal because
of the way the seals are mounted. Like the riders on a tandem bicycle, both
seals face in the same direction. The buffer fluid is introduced between the two
seals in an unpressurized dual seal at a pressure lower than the seal chamber
pressure.
• Dual mechanical seal with pressurized barrier—referred to herein as a pressur-
ized dual seal.
The seals are generally mounted in series in a standard pressurized dual seal.
As an alternative, the seals can be mounted back-to-back. Before API 682 was
published, this alternate configuration was referred to as a double seal. The
barrier fluid is introduced between the two seals in a pressurized dual seal at a
pressure higher than the seal cavity pressure.
A third type of dual mechanical seal is:
• Dual mechanical seal with dry-running outer seal.
The outer seal (mounted in series) is a dry running seal designed to run with a
gas cushion between the faces much like a compressor dry-gas seal. In an API
Plan 71 version of this seal no buffer or barrier gas is introduced between the
two seals. However, the space may be vented to a closed system such as a vent
or flare header. Also, the seal gland may include a drain connection to drain the
space between the two seals. In an API Plan 72 version of this seal a buffer gas
is introduced between the two seals, and the space is vented to a closed system
such as a vent or flare header. Also, the seal gland may include a drain connec-
tion to drain the space between the two seals.
Unpressurized Dual Seals

An unpressurized dual seal is shown in Figure 800-19. The seal closest to the
pumped fluid is generally sealing that fluid. The other seal is normally not in contact
with the pumped fluid.
Fig. 800-19 Dual Mechanical Seal With Unpressurized Buffer (Unpressurized Dual Seal)
Courtesy of American Petroleum Institute
Both sets of seal faces in an unpressurized dual seal need to be supplied with a
generous quantity of fluid in order to work properly. The fluid supplied to the seal
closest to the pumped fluid (the inner seal) is the pumped fluid itself, either from the
seal cavity or from a positive flush supply as with a single seal. The fluid supplied
to the seal furthest from the pumped fluid (the outer seal) comes from the buffer
fluid support system. The buffer fluid support system is vented to atmosphere or to a
closed low-pressure system, hence the name unpressurized buffer.
Both sets of mechanical seal faces in an unpressurized dual seal generate heat. In the
case of the outer seal, the buffer fluid must remove the heat. This requires the buffer
fluid to be circulated, which is usually accomplished by a pumping ring that rotates
with the pump shaft (similar to the circulation device for a plan 23 flush system on a
single seal).
In addition to a buffer fluid reservoir and the pumping ring, additional hardware is
needed to provide a complete buffer fluid support system. Refer to Section 846 for a
description of this equipment and for further information on dual seal support
systems.
Unpressurized dual seals are generally selected for one of three reasons:
• To provide a back up seal in case the inner seal fails. Some examples include a
back-up seal to prevent a highly toxic fluid from leaking to atmosphere in the
event of an inner seal failure, or a back-up seal to prevent leakage from an inner
seal failure in a pump installed in a remote unattended location.
• To provide a second seal sealing a separate fluid (from the pumped fluid), to
reduce pumped fluid emissions.
• To alter the operating environment on the atmospheric side of the inner seal,
usually to increase the reliability (extend the life) of the inner seal faces.
More information on dual seal selection and application can be found in Sections
842 and 850.
Pressurized Dual Seals

A standard pressurized dual seal is shown in Figure 800-20. In a pressurized dual
seal both seals normally seal the barrier fluid. This is accomplished by operating the
barrier fluid system at a higher pressure than the pumped fluid sealing pressure.
Barrier fluid that passes between the inner seal faces mixes with the pumped fluid
and goes out the pump discharge.
Fig. 800-20 Standard Dual Mechanical Seal With Pressurized Barrier (Pressurized Dual Seal)
Courtesy of American Petroleum Institute
On first inspection, the drawings in Figures 800-19 and 800-20 look the same. The
only noticeable difference between the two drawings is the absence of the
atmospheric side gland connection in the pressurized dual seal drawing. However,
some subtle differences exist in the standard pressurized dual seal that are not easily
shown in an assembly drawing.
In the standard pressurized dual seal, the mating ring in the inner seal, and the outer
seal must be designed to stay in place (not open) in the event that barrier fluid
pressure is lost. This represents a pressure reversal for the inner seal. When this
occurs, the entire seal is designed to operate as an unpressurized dual seal. The
difference between unpressurized and pressurized dual seals is that the mating ring
in the inner seal of the pressurized dual seal is balanced differently to accommodate
pressure reversal.
The barrier fluid supplied to a standard pressurized dual seal must be pressure
regulated to stay 20-60 psi higher than the pumped fluid sealing pressure (the seal
cavity pressure behind the throat bushing). If the pressure is too low, unpressurized
dual seal operation will result. If the pressure is too high, the inner seal will run hot
and may fail prematurely.
The design and operation of the standard pressurized dual seal is aimed at solving
the primary reliability problem with the alternate pressurized dual seal
(back-to-back double seal), pressure reversal. Reversal of pressure across the inner
seal faces in a back-to-back double seal causes the faces to open, and usually to fail.
Also, the barrier fluid reservoir becomes contaminated with the pumped fluid and
overflows through the vent to the collection system (flare). Hence, loss of barrier
fluid pressurization is not acceptable. This serious design flaw makes it very
difficult to operate the double seal through transients, and is the primary reason why
we rarely apply this type of seal in Chevron facilities. An alternate pressurized dual
seal (double seal) is shown in Figure 800-21.
Fig. 800-21 Alternate Dual Seal With Pressurized Barrier (Double Seal)
The barrier fluid in a pressurized dual seal support system must be circulated to
remove face generated heat from the seal. The barrier fluid must also be pressurized
to the appropriate operating pressure. These functions can be handled in two ways:
• By pressurizing the vapor space in the reservoir and by circulating the fluid
with a pumping ring. The reservoir can be pressurized by a process connection
or by connecting it to an inert gas system. The fluid-circulation pumping ring is
similar to the circulation device for a plan 23 flush system on a single seal.
• By a barrier fluid circulation pump, that circulates the fluid at the appropriate
pressure.
Refer to Section 846 for a description of the required hardware and for further
information on barrier fluid support systems.
Buffer and Barrier Fluids

The primary purpose of buffer and barrier fluids is to provide lubrication and
cooling for the seal faces. Intuitively, it might seem like a light lubricating oil would
be an ideal choice for the majority of applications. But testing and experience have
demonstrated that there is more to selecting the right fluid than previously thought.
For example, in seal manufacturer testing, automatic transmission fluid performed
so poorly that it is now not recommended for any application. Similarly, automotive
antifreeze is not recommended because it contains additives that plate-out on the
seal faces and accelerate wear.
Recommended buffer and barrier fluids include water, #2 diesel, glycol/water
mixtures, and some synthetic oils. White mineral oil and single cut paraffinic oil are
not recommended. Some Company locations have reported good results with Royal
Purple BW/IP-22 for extreme temperature conditions. The cost is fairly high
however at about $700 per 55 gallon drum. A more detailed list of recommended
buffer and barrier fluids and their properties follows.
Temperature
Recommended Buffer / Barrier Fluids Range (Deg F)
Kerosene 0 to 275
#2 Diesel -10 to 300
50% Ethylene Glycol / 50% Water -50 to 220
50% Triethylene Glycol / 50% Water -10 to 350
50% Propylene Glycol / 50% Water -20 to 220
Water 32 to 150
Royal Purple BW/IP-22 -60 to 450
Chevron Tegra Synthetic Barrier Fluid -60 to 450
When pressurized dual seals are used, it is important to keep in mind that some of
the barrier fluid will flow into the process fluid, and therefore it may be necessary to
select a fluid that will not create product contamination problems. Normally, the
leakage will be a very small amount ranging from a few drops per day to a few
drops per minute depending on the severity of the sealing conditions. In the event of
an inner seal failure however, the entire contents of the barrier fluid reservoir may
be discharged into the pumped fluid in a matter of minutes.
When selecting a buffer/barrier fluid, the following factors should be kept in mind:
Viscosity: In general, viscous liquids (even a light lube oil) present more sealing
difficulties than liquids with a viscosity like water.
Heat Capacity: Liquids differ in their ability to absorb heat. It takes less heat
energy to raise the temperature of a given quantity of oil one degree than it takes to
raise the same quantity of water by one degree. Water has a greater “specific heat”
than oil.
Vapor Pressure: The fluid must remain cool enough to prevent flashing at the seal
faces.
Mixing of Barrier Fluid and Pumped Fluid: With pressurized dual seals, the
barrier fluid enters the pumped fluid. With unpressurized dual seals, the pumped
fluid enters the buffer fluid.
Leakage to the Atmosphere: With both unpressurized and pressurized dual seals, a
slight amount of the buffer/barrier fluid will leak past the outer seal. Fluids should
be selected so that this leakage does not present a hazard or emission problem.
The Mechanical Seal Data Sheet in API 682 has space for the name, specific
gravity, and vapor pressure of the buffer/barrier fluid. This data should be filled in
by the person specifying the buffer fluid.
Dual Seal With Dry Running Outer Seal

The dry running gas seal has become a viable and popular selection option as the
outer seal for some dual seal applications. The liquid seal and the dry running gas
seals in a dual seal with dry running outer are mounted in series (tandem). The inner
seal requires a positive flush just like a single seal. The vapor space between the
primary and outer seals can be vented and/or drained to atmosphere, or vented
and/or drained to a closed system. A seal support system like a buffer/barrier fluid
system is not required.
Applications for dual seals with dry running outer seals are emerging. As of this
guideline revision date, they have primarily been applied in hydrocarbon services in
locations requiring low emissions. They have also been applied in some services
containing benzene. As of this guideline revision date, the requirements for
purging/venting the vapor space between the primary and outer seals have not been
standardized. These seals may be purchased in accordance with API 682.
842 Typical Applications For Dual Seals

Unpressurized Dual Seals
Figure 800-22 illustrates a typical unpressurized dual seal system. Three common
applications are discussed below:
• A back-up seal is desired
• Environmental regulations require a reduced emissions seal
• The environment on the atmospheric side of the inner seal needs to be altered.
Back-up Seal
The outer seal in an unpressurized dual seal provides a back-up seal in case the
inner seal fails. This is desired for personnel safety against toxic or hazardous fluid
leakage, and for some seals in remote unattended locations.
Sometimes plant operators elect to apply unpressurized dual seals because they want
the security of a back-up seal. The unpressurized dual seal has two seals with the
same pressure rating so that if the inner seal fails, the outer seal can take its place.
Fig. 800-22 Unpressurized Dual Seal and Buffer System
Some pump installations are remote and run unattended. It could take several hours
for an operator to reach the site. The pump could run for an extended period with a
substantial leak, resulting in product loss and environmental or safety hazards. An
unpressurized dual seal is sometimes used in these applications to provide an
indication of an impending release to atmosphere. For flashing services, this is
accomplished by installing a pressure switch in the buffer fluid reservoir. When the
inner seal leaks, the pressure rises and activates the switch, which can trip an alarm
or shut down the unit. For non-flashing services, the reservoir level will rise when
the inner seal leaks. A high level alarm or shutdown can be tripped with appropriate
instruments installed in the reservoir. When the outer seal leaks, the reservoir level
falls, and a low level alarm or shutdown can be tripped.
We must exercise caution when unpressurized dual seals have been installed for
back-up service:
• There is a temptation to continue to run the pump after the inner seal has failed.
• The buffer fluid gradually becomes contaminated with the pumped fluid as the
inner seal leaks. Depending on the pumped fluid, this may pose a hazard, either
through exposure to a toxic or hazardous fluid or through increased hydro-
carbon emissions.
• Seals will sometimes open in the event of a thrust bearing failure. Since both
seals are in series in an unpressurized dual seal, they would both open and
provide no more protection than a single seal.
A common application of unpressurized dual seals is in LPG services. The primary
concern is that a leaking seal may create a vapor cloud, which could then ignite.
Reduced Emissions
The Federal Government (EPA) has proposed a clean air act limiting hydrocarbon
emissions to 1000 ppm per seal. This is consistent with the current law in Southern
California established by the SCAQMD, known as Rule 1173, which requires
“reactive organic gas” (ROG) emissions to be less than 500 ppm per seal. The laws
do not require the use of dual seals.
In addition to the SCAQMD Rule 1173 emission limitation, there is a separate
requirement to install the “Best Available Control Technology” (BACT). BACT
installations are required in Southern California for:
• New installations of pumps in hydrocarbon service
• Pumps being relocated in the plant
• Pumps which are modified to meet new process conditions
• Pumps which fail the Rule 1173 emissions restrictions through periodic inspec-
tion more than 5 times in one year.
BACT is defined by a listing of mechanical options for pump and seal applications.
In order of preference these are:
• Sealless pumps
• Dual seals with a synthetic buffer/barrier fluid
• Dual seals with an organic hydrocarbon buffer/barrier fluid.
BACT does not recognize single mechanical seals as a “Best Available” technology.
At present, BACT is only required in Southern California. However, there is a
Federal Government (EPA) trend to promote the use of “Maximum Achievable
Control Technology” (MACT). This is similar to BACT in application but applies to
all domestic (USA) installations.
Company operations are affected by different environmental agencies, rules and
laws. The majority of Company locations have specific Company employees to
interface with the applicable agencies. Readers are strongly urged to work through
these employees in any matters pertaining to understanding what the laws are, what
they mean, and how we must comply.
Single seals are recommended for use in VOC services, when the Company is given
the choice. It must be emphasized however, that single seals in these services need
to be well engineered in order to work as they should.
Altered Environment
Caustic and crystallizing fluids tend to form solid crystals or deposits on the
atmospheric side of single seals as the fluid leaks past the faces. To keep this from
occurring, it is necessary to alter the environment on the atmospheric side of the seal
faces by keeping the faces wet. One way to accomplish this is by installing an
unpressurized dual seal. With the unpressurized dual seal, the back side (formerly
the atmospheric side of the single seal) is kept wet by the buffer fluid. Deposits
cannot form and the seal does not fail in the manner described.
Pressurized Dual Seals

There are three common applications for pressurized dual seals:
• Nonlubricating fluids (e.g., supercritical CO2).
• Equipment that may operate dry in the seal area.
• Hazardous services when an external flush fluid is not available.
Pressurized dual seals are not usually a good choice for dirty or abrasive services.
Dry Running Applications

Since pressurized dual seals require a pressurized barrier fluid, they do not depend
on the pumped liquid for lubrication.
Pressurized dual (back-to-back double) seals are standard equipment on top entry
mixers (Figure 800-23) because the liquid in the mixer does not reach the seal. The
buffer fluid system is fairly simple because mixers usually do not rotate fast enough
to generate much seal heat. The system usually consists of a pressurized reservoir
without provision for circulation.
Fig. 800-23 Top Entry Vessel Mixer
Pressurized dual seals are sometimes useful for pumps that suffer from periods of
dry running caused by suction loss or cavitation. But because many pressurized dual
seal systems are inherently troublesome, readers are cautioned against rushing into
their use. It may be possible to correct an existing seal problem due to dry running
condition by using an external (API Plan 32) flush.
Nonlubricating Liquids
Some liquids have virtually no ability to lubricate and yet can still be pumped. An
example of such a liquid is supercritical carbon dioxide. Carbon dioxide is usually
handled with compressors but sometimes it is pumped. These pumps cannot be
sealed with single mechanical seals.
Pressurized Dual Seals For Hazardous Services

The term hazardous refers to substances that are dangerous when even small
amounts are released to the atmosphere. Probably the best example in our industry
is hydrogen sulfide (H2S) which can be fatal. Hydrofluoric acid is another example.
Other liquids, such as benzene, may not produce immediate symptoms but also
present a recognized health hazard.
Single seals may be the most reliable choice for hazardous services when an
external flush is available. The source of the flush fluid must be extremely reliable
and non-hazardous. In addition, the use of close clearance throat and throttle
bushings is recommended.
Numerous disadvantages are associated with pressurized dual seals. Before
selecting a pressurized dual seal for sealing a hazardous fluid, consider the
following facts:
1. Pressurized dual seals, because of their inherent complexity, are difficult to
operate and maintain. Repairs are time consuming, expensive, and usually
create greater exposure hazards than occur when the pump is operating.
2. With certain seal designs, pressure reversals can contaminate buffer fluid
A “pressure reversal” is common in actual practice because of unexpected
increases in pump suction pressure or loss of barrier fluid pressure. When this
occurs, the inner seal may open and the barrier fluid becomes contaminated
with the pumped fluid. Exposure to a contaminated barrier fluid can be just as
dangerous as exposure to the pumped fluid.
Many components of a support system for pressurized dual seals would have to
be disassembled, cleaned, and repaired after the barrier fluid is contaminated.
This results in an increased exposure risk.
3. Barrier fluids gradually become contaminated even when there are no seal
failures or pressure reversals.
Even though the barrier fluid pressure exceeds the process pressure, mixing of
the pumped fluid with the barrier fluid occurs. This illustrates the fact that pres-
surized dual seals do not completely contain the pumped fluid.
Some groups will view dual seals as superior technology and argue that it is
irresponsible for a company to use anything but dual seals in hazardous services.
One can easily visualize a scenario in which there is an incident involving loss of
life or creation of a health hazard because of a leak from a pump equipped with a
single mechanical seal. It may be alleged that the company failed to provide the
“best available technology” because they valued profits more than human safety.
Conversely, if dual seals were used, the same (best available technology) argument
might be used to support the Company’s position.
The above is by no means intended to suggest that subjective opinions should be put
above what we know to be the most technically safe practice. Our responsibility is
to provide the safest installation.
Pressurized Dual Seals for Dirt or Abrasives?

Many seal suppliers like to promote pressurized dual seals as a solution for services
that contain dirt and/or abrasives. Certain dual seal components are more vulnerable
to failure than the same components of a single seal. For example, when pusher
seals are used in a back-to-back arrangement, the foreign material can collect under
the secondary seal and cause hang-up or leakage. If abrasives are present, they will
be centrifuged against the inside diameter of the two faces. Centrifugal force tends
to drive the particles between the faces and cause accelerated wear.
When engineering a seal for a service with abrasives, consider the following:
• Use a single seal and API Plan 32 flush.
• Use tungsten carbide versus silicon carbide faces if the seal has to be designed
to handle solid particles.
• Use increased closing force to keep particles out.
• Cyclone separators must be carefully engineered. Often they are not successful
at removing the particles that can cause a problem.
• Stationary seals are preferred over rotating seals.
843 Dual Seal Designs

Seal Types
Dual seals can be made up of pusher or bellows seals. Usually both seals are the
same type, but a dual seal could be made up of one pusher and one bellows seal.
Mounting Configurations
Dual seals can be mounted in a rotating or stationary arrangement. In other words,
the flexible element (the bellows assembly or the part that contains the spring or
springs) can rotate with the shaft or it can be mounted in the gland (stationary).
Dual seals can be cartridge mounted. Cartridge design is especially desirable for
dual seals for two reasons:
1. The cartridge design takes much of the confusion out of positioning the seals at
the proper location on the sleeve to insure the right amount of spring or bellows
compression.
2. A bench pressure test (with air, solvent, oil, or water) can often be done without
a special fixture. This test is recommended before installing the cartridge and is
required for seals supplied in accordance with API 682.
Finally, there is also the option of having pressurized dual seals mounted in series or
back-to-back.
Rotating and Stationary Seals

Whether a pump is using single or dual seals, the same points apply concerning the
use of rotating or stationary mounting for the flexible element. Designers seem to
favor rotating seals but this may be due more to tradition and habit than conscious
choice. Stationary seals have an advantage over rotating seals. They don’t have to
flex with every revolution due to gland fit (perpendicularity) tolerances, and are a
better choice for high speed applications.
Outside Seals
In some pumps, there may not be enough room in the stuffing box to fit two
mechanical seals. Designers might attempt to utilize outside seals. Outside seals are
not recommended for Company applications and are not in conformance with
requirements of API 682. These are some design deficiencies with outside seals:
• An outside seal may be damaged by an external force. This could include
impact or washing it with high velocity water or cleaner, etc.
• Internal pressure tends to open an outside seal. This decreases the tolerance of
the outside seal to increases in pressure in the buffer fluid.
• An outside seal cannot have a throttle bushing and, therefore, there is no way to
restrict leakage when the seal fails. An outside bellows seal would sling or
spray leakage all around if the bellows failed. Both bellows and pusher types
would sling or spray leakage from between the faces.
Back-to-Back (Double) Seals

The advantages of a back-to-back (double) seals are:
• There is minimum exposure of the seal parts to the liquid being pumped. Most
of the seal hardware is in contact with the barrier fluid.
• They can withstand barrier fluid pressures up to the maximum pressure rating
of the seal.
• It is usually easier to circulate barrier fluid (for cooling) in the immediate area
of the seal faces.
The disadvantages of a back-to-back (double) seals are:

• Pressure reversals can open and damage the inner seal.
• The inner seal mating ring is often mounted in the bottom of the seal chamber.
If the seal is not a cartridge type, it could be difficult to remove the mating ring
without disassembling the pump.
Buffer/Barrier Fluid Ports

The buffer/barrier fluid ports are the tapped holes in the gland plate or in the seal
chamber through which the fluid enters and exits.
Ideally, ports should be as large as the inside diameter of the piping or tubing
connecting the seal chamber to the reservoir. This is not always possible because of
the size of the pump and gland plate, but very often there is room to enlarge them
beyond what the vendor would ordinarily supply.
The axial location of the ports is just as important as the size. The problem with
using the pump stuffing box lantern ring connection as a flush port is that it may
result in a circulation path that leaves the seal faces in a dead zone where localized
overheating can occur.
To guard against failures from overheating, require that ports be tangential
(tangential ports greatly increase flow rates), as large as practical, and located close
to the interface between the seal faces. Confirm this during drawing review. Also
verify that the system is adequately sized for cooling by reviewing the calculations
with the supplier.
Pumping Rings
Pumping rings circulate the buffer/barrier fluid for seal cooling. Pumping ring
performance is dependent on a variety of factors such as peripheral speed, design,
clearance, direction of rotation, liquid viscosity, and system resistance. The actual
performance of the various designs is discussed in Section 846.
The important points to remember about pumping rings are:
• Pumping rings come in a wide variety of types and some are better than others.
• Designs that rely on maintaining a close clearance with the seal chamber bore
have a history of poor performance.
• If the pumping ring is one that will not pump if it is installed backwards, insist
on a feature that only permits installation in the correct direction.
• Insist on a feature that only permits installation in the correct axial location.
Examples of Dual Seals on the Market

There is a considerable difference between the dual seals available for API pumps
and those for ASME or ANSI pumps. Seals for ASME pumps are available in
several “off-the-shelf” designs which are nicely illustrated and described in vendors
bulletins.
ANSI dual seal designs include:

John Crane Type 88
Sealol Type 622
Flowserve (BW/IP) Uniseal II
Flowserve (Durametallic) X 200
There is a significant benefit to using predesigned seals whenever possible because

the vendor has had more opportunities to refine the design based on feedback from
the field.
Unfortunately, Company experience with multiple seals by manufacturer and model
is not well documented. Users are encouraged to develop experience histories at
Company locations and other facilities where similar seals are installed.
Specifying Dual Seals

Dual seals should be of proven design. Proven design is defined as having:
• At least two years successful operating experience under these service
conditions:
– A liquid of similar viscosity, specific gravity, vapor pressure, specific heat,
and lubricating ability.
– Equal or greater temperature, pressure, running speed, and seal size.
• The same design features, such as:
– Face mounting method
– Sleeve design
– Gland plate configuration
– Port sizes and locations
– Drive collar design
– Pumping ring type
– Gasket design
The combination of different pump types, different seal vendors, and different
service conditions makes it difficult to consistently get a dual seal that has a proven
track record for that exact combination. As a result, many of the dual seals we buy
will be to some degree unproven. We can, however, substantially improve our
chances of getting a reliable seal by identifying those design features which work
well and incorporating them into the seals we buy.
844 Reliability and Experience

General
Mechanical equipment reliability is often expressed in mean time between failures
(MTBF). For single seals, the MTBF is considered acceptable at about four years or
more. For dual seals however, there is a wide variation in reports and opinions
concerning “normal” life. The successful application should take into account the
details of the pump design such as:
The successful application should take into account the details of the pump design
such as:
• seal chamber size and pressure
• shaft diameter, speed, and deflection
• construction of the pump casing
• overall heat transfer characteristics—liquid temperature, pressure, viscosity,
cleanliness, and vapor pressure all affect seal design.
How Can the User Maximize Reliability?

The collective Company experience with seals is vast. Individuals with extensive
machinery/seal design and field experience should be used as consultants for new
applications.
Every operating facility has a financial incentive to reduce inventory and maximize
interchangeability by standardizing on seals which are proven winners.
The steps that can be taken at the specification, purchasing and commissioning
stages are the same as those given in Company guidelines on quality assurance.
They are:
• Vendor Prequalification and Seal Selection
• Comprehensive Specifications
• Design Review
• Quality Assurance Checks
• Testing
• Pre Start-Up Field Check Out
Vendor Prequalification and Seal Selection

Prior to selecting a seal type or vendor, these questions should be answered:
• Are there other pumps with seals installed which are in identical or similar
service and have a good repair history?
• Are there pump seals of identical size already in stock in-house?
• Is the proposed vendor capable of furnishing the hardware quality, engineering
support, and follow-up necessary for a successful installation?
• Does the facility have maintenance familiarity with the seal being considered.
Company experience with major seal vendors listed in order of seal population is as
follows: John Crane, Flowserve. The use of the Chevron Preferred Supplier
Agreements can help bring seal inventory and consistent seal designs to a location.
Comprehensive Specifications
The Company has Specification PMP-SC-4662 “Shaft Sealing Systems for
Centrifugal and Rotary Pumps - Exception to API 682, 2004 (ISO 21049:2004),”
which is a companion specification to API 682. Users are urged to include these
documents with all dual seal purchases for API 610 and 676 pumps.
Design Review
A design review affords an opportunity to confirm that the vendor understands the
specifications and has performed the necessary calculations required for the
application.
Quality Assurance Checks

This step allows the Company to verify that the hardware is being manufactured and
assembled according to the Company’s and vendor’s requirements.
Testing
It is highly recommended that seals be qualification tested, and that each seal is
given a seal manufacturer (air) test in accordance with PMP-SC-4662 and API 682.
The best place to find and correct deficiencies is in the supplier’s facility.
Pre Start-Up Field Check Out

A complete visual inspection for correct location of components and proper piping
hookup should be performed. This is an opportunity to check for leaks, verify
operation of instruments, and, possibly, even circulate buffer fluid and confirm that
the necessary head and flow rate are developed.
Once a system is ready to be put into service, operators should be thoroughly
familiarized with the system operation including venting (if required), reservoir
pressurization, reservoir filling, and frequency of checking vital signs such as
temperature, pressure, and flow rate. Operators should be taught how the seals and
support system work and what factors cause premature failure.
In addition to formal training, special attention should be given to providing clear,
easy to read, easy to understand operating manuals located where they will be read
and used. In addition, signs should be posted at various places on the pump and
support system with labels, reminders, procedures, and instructions. This is a step
that is easy to overlook and yet it has proven its value over and over again.
845 Costs
Initial cost
Dual seals will cost more than single seals for a given pump and service. There is a
higher initial cost and higher repair costs.
A common rule of thumb is “dual seals cost two to three times as much as single
seals.” This rule only applies to seal hardware.
There are extra costs associated with the installation of dual seals, too. Most systems
need additional piping, electrical, and instrument work. Pressurized dual seal
systems with a barrier fluid pump may require a foundation as well.
Maintenance and Repair Costs

Independent of reliability, dual seals and buffer/barrier fluid systems will also be
more costly to repair just because of their complexity.
Leaking inner seals often allow the buffer/barrier fluid to become contaminated with
the pumped fluid making it necessary to cleanup the system. Draining, dismantling,
cleaning, and flushing the reservoir and piping can easily take four or five mandays
with additional time and expense required to properly dispose of the old fluid. A
repair to a component such as a barrier fluid pump will take two to four mandays,
instrument repairs probably two mandays, system check out and start up another
two mandays.
846 Buffer/Barrier Fluid Support Systems

General
The buffer/barrier fluid support system is an extremely important part of a dual seal
arrangement. It is analogous to a seal life support system since it provides the vital
fluid for lubrication, cooling, and sometimes pressure maintenance. If a dual seal
arrangement is to be reliable, the same degree of care must be exercised in the
design, operation, and maintenance of the support system as with the seals
themselves.
In the following discussion, a dual seal refers to any two seals on the same shaft,
inside the same seal chamber. There are two types of dual seal systems:
• API 682 Arrangement 2 seal—a dual seal where the containment seal chamber
pressure, the chamber between the dual seals, is at a pressure less than the seal
chamber pressure.
• API 682 Arrangement 3 seal—a dual seal where the containment seal chamber
pressure, the chamber between the two seals, is at a pressure greater than the
seal chamber pressure
Basic System Descriptions

Unpressurized Systems (API Plan 52, ASME Plan 7352). An unpressurized
reservoir is used if the fluid between the two seals in a dual seal does not need to be
maintained at an elevated pressure, as with an API Arrangement 2 dual seal. API
682 Flush Plan 52 is defined simply as having an external fluid reservoir that is
vented. Figure 800-24 shows a typical Flush Plan 52 flush system.
The fluid in the reservoir is called a buffer fluid.
The seal reservoir (or seal pot) typically has a capacity of 2-5 gallons, where the
working volume is around 1 gallon. The piping or tubing to and from the pot should
be arranged to allow the buffer fluid to thermosiphon and the seal design should
provide pumping action (centrifugal or axial flow pumping ring) to cause forced
circulation of the buffer fluid in order to remove heat generated by the seals.
Fig. 800-24 Unpressurized Dual Seal Support System, API Plan 52. Courtesy of the American Petroleum Institute
Some applications require a heat exchanger or cooling coil in the seal reservoir (seal
pot) to cool the buffer fluid and enhance thermosiphon.
The seal reservoir should be high enough above the seal to provide a minimum level
of buffer fluid. The recommended level is 1 foot, as shown in Figure 800-24.
Additionally, the supply and return piping/tubing from the seal reservoir to the seal
should be adequately sized and arranged to minimize head losses. The piping/tubing
should be arranged to avoid a vapor trap in the return line from the seal to the seal
reservoir.
The buffer fluid level in the seal reservoir must be maintained above the return line
entry position in the seal reservoir to achieve and maintain the thermosiphon effect.
If the fluid falls below this level, and if circulation is dependent only on the
thermosiphon effect, flow will discontinue, and cause seal damage. A means of
monitoring this critical low level must be provided in the seal reservoir and is
usually a local sight glass and a remote low-level alarm. Sometimes a shutdown is
included for the pump itself if the seal level gets too low.
The seal reservoir must be vented before depressurizing the pump. If not, pressure
could open the outer seal’s faces (face-to-back or the old tandem seal arrangement)
and dump buffer fluid to the atmosphere.
Detecting Unpressurized Dual Seal Failures. Detecting the failure of the outer
seal in an unpressurized dual seal is no different than failure detection in a single
seal because the leakage is visible. Detecting failure of the inner seal, however,
depends on the type of liquid being sealed.
Leakage from an inner seal enters the buffer fluid. If the leakage does not vaporize
in the buffer fluid reservoir, then the liquid level will rise. In this case, a high level
alarm in the buffer fluid reservoir is all that is needed to indicate an inner seal leak.
In a flashing service (such as propane), an inner seal leak will not create an
increased level in the buffer fluid reservoir. In fact, a severe leak will actually cause
the level in the reservoir to decrease and may empty the reservoir completely. A
common method for detecting an inner seal failure for flashing services is the use of
a high-pressure alarm in the reservoir and a restriction orifice in the reservoir vent
line. The idea is that as inner seal leakage increases, the back pressure against the
orifice will increase and set off the high pressure alarm. The majority of
unpressurized dual seal alarms for flashing service are configured this way, but there
are two problems with this arrangement. First, if the check valve in the vent line
sticks closed, there will be a gradual normal build-up of pressure in the reservoir
that will set off the alarm. Second, if the check valve does not seal properly, a
pressure increase in the relief system may set off the alarm.
The correct setting for the high pressure alarm must be determined in the field based
on the frequency and magnitude of relief system pressure excursions and the
reliability of the check valves used in the relief line. The lowest practical setting is
most desirable because higher back pressures will tend to blow the buffer fluid out
the relief line. If this happens, the outer seal may be starved for buffer fluid.
Company field tests with 1/8" diameter orifices in the relief line (1/16" diameter
orifices had plugging problems) showed significant loss of buffer fluid at back
pressures as low as 15 psi with water as the buffer fluid. If the high pressure alarm
and orifice system proves to be impractical for a particular location, consider one of
the following two options:
1. Install a differential pressure alarm and gage across the orifice in the relief line.
This arrangement will be unaffected by normal pressure build-up due to a stuck
check valve and unaffected by increases in pressure in the relief system. A
differential pressure of 3 to 5 psi across a 1/8" orifice represents enough
leakage to warrant a seal repair in a flashing service.
2. Install a flow switch in the reservoir relief line. The Company has begun to
evaluate a thermal dispersion design in reciprocating compressor packing and
distance piece vent lines with good results so far. One example is the model
12-64 flow switch / monitor by Fluid Components Inc. of San Marcos,
California. It appears to be well-suited to unpressurized dual seal (flashing
service) applications.
Pumping Ring Performance. The pumping ring must be regarded as a critical
component of the system. Seals work best when the liquid circulates vigorously
around the seal faces to provide a high degree of cooling. Pumping rings have
trouble doing this for several reasons:
1. None are positive displacement devices so they develop very low head.
2. Most are highly dependent on maintaining close tolerances (the importance of
this is frequently overlooked).
3. Some designs are unproven and are installed without design engineering review
or performance testing.
4. The system head curve may not match the pumping ring capability.
5. Pumping rings are very sensitive to the buffer fluid viscosity.
API 682 allows only pumping rings with published performance curves based on
test results.
Buffer Fluid. The buffer fluid in unpressurized dual seals provides lubrication and
cooling for the outer seal only. The process stream provides lubrication and cooling
for the inner seal.
The purchaser needs to select a buffer fluid that is compatible with the process
liquid. The fluid needs to have good lubricating properties for the outer seal in an
Arrangement 2 unpressurized dual seal. It should be non-corrosive and not too
viscous. In cold weather, it should not freeze.
Buffer fluid systems that operate at atmospheric pressure will usually need a
reservoir and a pumping ring. It is important to note, however, that these systems
can still suffer short seal life due to local overheating at the faces if the flush flow is
inadequate or if the flush ports are poorly located.
Pressurized Reservoirs (API Plan 53, ASME Plan 7253). A pressurized reservoir
is used if the fluid between the two seals in a dual seal needs to be maintained at an
elevated pressure. API 682 Flush Plan 53 is defined simply as having an external
fluid reservoir that is pressurized by an external source. In Flush Plan 53A, the
reservoir is connected to a supply of pressurizing liquid or gas. If gas is used, it is
usually nitrogen. In Flush Plan 53B, the reservoir is a pre-pressurized bladder
accumulator, and in Flush Plan 53C, the reservoir is a piston accumulator.
Figure 800-25 shows a typical Flush Plan 53 system. This support system is
employed with pressurized dual seals. The fluid in the reservoir is called a barrier
fluid.
Fig. 800-25 Pressurized Dual Seal Support System with Pressurized Reservoir, API Plan 53 Courtesy of the American
Petroleum Institute
Plan 53 is typically used with Arrangement 3 pressurized dual seal and is similar to
Plan 52, except that Plan 53 uses a pressurized system with a blanket pressure,
20 psi or higher than the seal chamber. If the pressure is too great, the secondary
(outer) seal will see too great a pressure differential and experience poor reliability.
Many of the tips, mentioned under Plan 52, above, apply.
The barrier fluid in Plan 53 provides lubrication and cooling for both the inner and
outer seals of the dual seal. Therefore the cooling requirements of this plan are
greater than for an Arrangement 2 seal. Plan 53 does not perform well in a pure
thermosiphon system. It requires a pumping ring or external pump to get adequate
circulation for cooling
Pressurizer/Circulator (API Plan 54, ASME Plan 7354). Plan 54 is also used
with Arrangement 3 seals. The system uses a motor driven pump to circulate barrier
fluid and to create the required barrier fluid pressure between the pressurized dual
seals. Figure 800-26 shows a typical Flush Plan 54 system. A backpressure
regulator, filter, alarms for pressure, flow, and level are typically included.
The advantage of this system is that a high barrier fluid flow rate can be achieved,
which greatly improves the heat removal process and makes this system suitable for
high pressure-velocity (PV) applications.
Pressurized dual seals require a system which maintains a pressure between the
seals. In this case there is a choice between the pressurized reservoir and the
pressurizer/circulator. Each system has its advantages and disadvantages which are
listed here.
Advantages of pressurized reservoir:
• Fewer components
• Cost may be lower
Disadvantages of pressurized reservoir:
• Usually uses pumping ring (low flush flow rate)
• Filling may be tricky
• Dependent on reliable gas supply
• Gas entrainment in barrier fluid above about 150 psig (if a gas blanket system
is used)
• Connections must be kept gas-tight to avoid additional emissions or loss of
pressurizing gas
• Relies on thermosiphon effect for fluid circulation while the pump is not
running
Advantages of pressurizer/circulator
• Excellent cooling
• Allows use of low flow alarm
• Easy to fill reservoir
• Barrier fluid can be filtered
• Large reservoir capacity
• Can circulate barrier fluid while sealed pump is down.
• Can serve several pumps
Fig. 800-26 Typical Double Seal Support System with Non-Pressurized Reservoir, API Plan 54. Courtesy of the
American Petroleum Institute
Disadvantages of pressurizer/circulator
• Takes up more space
• More components
• May cost more to purchase, maintain, and operate.
• Lower reliability
• Seal system is dependent upon a motor driven pump
Operating Pressure
Pressurized dual seals require a pressure between them that is higher than the
pressure in the seal chamber. Barrier fluid pressure is usually set at about 25 psi
higher than seal chamber pressure. The seal chamber pressure in many pumps will
vary which may require a higher barrier fluid pressure setting. A low pressure alarm
may be needed in order to sense the differential pressure between the seal chamber
and the barrier fluid in order to avoid spurious trips.
There are several variables that designers must consider to insure adequate
lubrication of the seal faces, including:
• Spring or bellows stiffness
• Seal set length
• Fluid port size and location
• Face design parameters, which include face width, dynamic stability, and face
response to changes in temperature and pressure.
Users can prevent many of these problems from occurring by following a few rules:
• Require proven designs and verify experience whenever possible.
• Require the vendor to show the engineering calculations used in designing the
seal.
• Make comprehensive seal testing part of the pump test. Require a separate seal
performance test for any seal in a critical service. (See API 682 for testing.)
• Do not wait until plant start-up to discover problems. Check out the system in
the field as soon as it can be run.
Instruments
Instrument requirements for buffer/barrier fluid systems vary depending on the type
of system and, to some extent, user preferences. For maximum reliability, the
proven equipment approach should be applied to individual instruments as well as
to packaged systems.
Packaged Systems versus Custom Engineered Systems

Not all mechanical seal suppliers offer predesigned support systems for dual seals.
However, the advantage of a predesigned system is that it is a proven design.
Support systems designed after receipt of order may lack attention to design detail,
quality assurance, and testing, resulting in reliability problems which persist long
after start-up.
John Crane and Flowserve build three types of predesigned seal support systems:
pressurizer/circulators, pressurized reservoirs, and unpressurized reservoirs.
Company experience with any of these units is limited, but reports from users are
favorable.

Before the right mechanical seal, flush plan, and auxiliary flush plan can be
selected, the seal service conditions must be thoroughly understood. API 682
provides guidance on seal selection in the appendices of the Standard. The
Company specification, PMP-SC-4662 also provides some guidance. To do a
complete job of selecting and specifying a seal system requires the user and supplier
to complete the API 682 Mechanical Seal Data Sheet in good faith.
Many larger Chevron facilities have developed local seal service application charts
which give guidance for many routine applications. In the absence of local
guidelines, a standardized list of mechanical seal selections for the more routine
applications appears in both the John Crane and Flowserve mechanical seal guides
(PMP-DG-4662-CRN and PMP-DG-4664-FS, respectively), as associated with the
respective Preferred Supplier Alliances (PSAs).
860 Troubleshooting and Failure Analysis
861 Background
Chevron Product Company’s Integrated Machinery Inspection (IMI) Organization
found that in the Refineries, mechanical-seal failure was a frequent and costly
centrifugal pump problem. The cost of these failures averaged approximately
$5,000 (1987) per repair.
Mechanical seals can run for several years without failure. Although an average seal
life of 6 to 12 months is commonly considered acceptable by some users, a properly
applied and installed seal, operated correctly, should last 5 to 10 years in normal
service, and 1 to 4 years in severe service. In simple terms, seals seldom wear out.
Failures are caused by improper application, maintenance, or operation.
862 Elements of Successful Seal Performance

A mechanical seal has several basic requirements for successful operation. Although
specific causes vary, all seal failures fall into one of these five categories:
1. The seal must be properly constructed and installed.
2. The faces must be continuously lubricated by a liquid film. If this liquid
film is lost due to loss of liquid in the pump or vaporization of the liquid
between the faces, extreme heat is generated at the faces (temperatures above
2000F are not uncommon) and the seal fails.
3. The seal must be protected from abrasives that enter between the seal
faces. Special face materials or flush systems are available to overcome this
problem.
4. The seal materials must be compatible (chemically inert) with the pumped fluid
or external flush fluid.
5. The seal must be protected from leakage deposits on the atmosphere side,
such as coke and crystals that build up under the seal faces. Special quench
systems are available to stop or wash away these deposits.
See Figure 800-27, below, for start-up procedure.
Fig. 800-27 Reliable Seal Operation
An informal analysis of 69 pumps (large CUSA refinery, 1980-1986) with excessive

seal failures found the following deficiencies.
• Loss of face lubrication 33%
• Bellows cracking caused by chloride or sulfide stress corrosion 21%
(See Appendix F for typical bellows-failure summary)
• Hostile liquid environment requiring clean exterior flushes 17%
• Corrosion 13%
• Corrosion fretting (wear) of the sleeve under the secondary seal 10%
• Coke or crystal build up on the atmosphere side of the seal under 6%
the faces
The analysis above does not include failures due to improper construction and
installation, although these no doubt also contribute to many failures.
863 Mechanical Seal Failure Analysis

The following process is an organized approach to finding the specific cause for a
mechanical seal failure.
Figure 800-28 provides a checklist that may be helpful in investigating seal failures.
Fig. 800-28 Checklist for Seal Failure Analysis
1. Pump/seal Maintenance History
Satisfactory seal life? ______ Yes ______ No
Trend Established? Describe
Equally spaced over time? ______

Random ______
History of cavitiation ______ Yes ______ No
2. Mechanical Condition of Pump
Thrust bearing failed? ______ Yes ______ No
Condition of radial bearing?
Impeller nut. ______Loose ______Tight

Sleeve gasket tight? ______ Yes ______ No
Gasket face condition:
Rubs on impeller wear rings?______, throat bushing?______, throttle bushing?______, other_________

Impeller/damage?
Throat bushing clearance exceeds design specification?
Checklist for Seal Failure Analysis (Cont’d)

Yes No
3. Condition of Mechanical Seal
A. Pusher
Springs. Condition: ____________________________________________
Sleeve under secondary seal. Condition: ___________________________
Secondary seal. Condition: _____________________________________
Stationary face anti-rotation pin damaged? _______ _______
B. Metal Bellows:
Pressure test bellows. OK? _______ _______
Bellows material: _____________________________________________
Failure at first convolution? _______ _______
Leaks between face and nose piece fit? _______ _______
C. All Seals
Set Screws:
Loose on shaft? _______ _______
Signs of sliding? _______ _______
Dull drive points? _______ _______
Metal Parts:
Corrosion? _______ _______
Rubs? _______ _______
Gasket Condition:
Heat? _______ _______
Swelling? _______ _______
Chemical Damage? _______ _______
Mechanical Damage? _______ _______
Seal Faces:
Surface cracks? _______ _______
Chipping? Inside diameter _______ _______
Chipping? Outside diameter _______ _______
Blisters? _______ _______
“Phonographic” scoring? _______ _______
“Drive lug” wear? _______ _______
Evidence of generalized heat? _______ _______
Shiny spots? _______ _______
Erosion? _______ _______
Wide or narrow wear track? ____________________________________
Checklist for Seal Failure Analysis (Cont’d)

Yes No
4. Flush System
Flush blocked in? _______ _______
Plugged orifices? _______ _______
Eroded orifices? _______ _______
Property piped? _______ _______
5. Cooling System
Fouled seal chamber water jacket? _______ _______
Fouled flush cooler? _______ _______
Fouled cooling piping? _______ _______
Cooling water manifolded correctly? _______ _______
Step 1. Review Operating Data

Distributed control system trends and operating logs often give an indication of why
a failure occurred or provide data before and during impending failure. These
resources should be reviewed prior to and up to the time of failure. It is also
beneficial to talk to plant operators. They may be able to give an eyewitness account
of the failure and may be aware of process/operating changes that may have
occurred immediately prior to failure.
Step 2. Review the Pump/Seal Maintenance History

Random failures are more likely to be plant upsets or people (operator,
maintenance) related. If there is no history file available, establish one. This is a
necessary tool for the person that follows up on your observations and
recommendations.
Quite often a mechanical failure of the pump will result in a seal leak. The most
obvious visual evidence in the field is often incorrectly diagnosed by the operator as
seal failure. Such an observation recorded on the work order may lead to confusion
and an incorrect maintenance history.
You may find that you are satisfied with the reliability of the seal and no further
investigation is required. However, if you have determined that the seal has an
unacceptable failure frequency, review the repair history for trends. Patterns such as
plant shutdown and startup failures, and equally spaced repetitive failures are
usually design-related.
Step 3. Inspect the Pump’s Mechanical Condition

If the thrust bearing failed, axial shaft movement would allow the seal to open. Do
not spend any more time looking at the seal. Concentrate on the cause of the thrust
bearing failure.
If the radial bearing failed, did the bearing failure cause a seal leak or did the seal
failure allow product to enter the radial bearing, causing a bearing failure? To
determine which failure occurred first, you need to broaden your inspection to
include the seal parts. See Step 3.
Inspect the impeller nut for looseness. A loose impeller nut will allow leakage
through the seal faces or under the sleeve.
Inspect the sleeve gasket and gasket faces for damage that is causing leakage.
Inspect for rubs on the impeller wear rings, throat bushing and throttle bushing. If
there is evidence of a rub look for the reason. Did a process upset cause excessive
shaft deflection? Does the pump have a history of excessive cavitation?
Check impeller balance. A foreign object in the impeller may have caused excessive
shaft deflection. Also check wear rings for evidence of rubbing.
Step 4. Inspect the Mechanical Seal

Pusher Seal Design (springs and secondary seal). Inspect the springs and the
sleeve under the secondary seal. If there is excessive wear to either the secondary
seal, the sleeve or the springs, suspect vertical misalignment of the stationary face or
excessive axial movement of the shaft.
Metal Bellows Seal. Pressure test the bellows for cracks. If the bellows material is a
300 series stainless steel suspect sulfide or chloride stress corrosion cracking.
Marginal face lubrication or excessive seal chamber pressure will also cause
bellows failure. This failure will be at the first convolution on either the nose or the
drive collar end of the bellows.
All Seals. Inspect all the metal parts of the seal for corrosion and upgrade materials
as necessary. Gasket deterioration can be from either corrosion or heat. The effect of
heat on an elastomer will be deforming, hardening, cracking or charring.
Inspect the condition of the seal faces. Evenly spaced surface cracks on the metallic
face—called heat checking—are caused by the loss of lubricating film between the
rotating and stationary face. When the liquid film is lost, the extreme heat generated
causes immediate surface cracking. This film loss is a result of boiling or flashing
and is most common in liquids with high vapor pressure. The liquid boils because of
either a drop in seal chamber pressure or a rise in temperature at the seal faces.
Causes for a reduction in seal chamber pressure include:
• Pinched suction valve
• Low NPSHA
• Partially plugged suction screen
• Flow rates beyond the pump’s best efficiency point
• Excessive throat bushing clearance
Causes for a temperature increase at the seal faces include:

• Temperature increase at the suction source
• Flow reduced below the pumps best efficiency point
• Loss of seal flush
• Excessive seal tension
The remedy for these conditions is to maintain a combination of temperature and
pressure in the seal chamber that suppresses boiling.
“Phonographic” scoring of the hard face is usually caused by abrasives. The
abrasives embed in the soft face and act as a grinding wheel on the hard face.
Abrasives can be assigned to two categories:
1. Suspended abrasives, such as sand or catalyst fines;
2. Soluble abrasives, such as boiler compounds that form between the seal faces if
the liquid between the faces is allowed to boil.
The accepted approach to sealing suspended abrasives is a combination of hard-face
materials, such as tungsten carbide versus silicon carbide, that won’t be scored by
the abrasives that enter between them. API 682 Flush Plan 32 is also used. Plan 32
is the injection of cool clean fluid into the seal chamber. To avoid the formation of
soluble abrasives between the seal faces, control the liquid temperature or pressure
in the seal cavity to suppress boiling.
Step 5. People Involvement

Be consistent during your investigation. Establish a list of questions prior to
contacting the process engineer, plant operator, or maintenance person. Make
them a part of the solution. Build their confidence and discourage blaming and
finger pointing. Take notes of their comments and compare them for events that
support each other. Evaluate their experience level. Are repair or operating
procedures and checklists available? A well-designed system will fail if those who
work with it do not know how to operate or repair it.
Step 6. Documents
The documents needed to investigate a seal failure are the same documents required
to design the seal. Keep these documents updated to reflect problem investigations
and solutions in the following areas:
• Liquid characteristics
• Operating conditions
• Plant flow diagrams
• Flush system
• Seal manufacturer
• Seal drawings
• Material list
• Seal part numbers, type and model
• Pump drawings and pump curves
864 Seal Failure Troubleshooting

After investigating the problem and collecting relevant information, this section
may be helpful in identifying the cause(s) and in suggesting possible solutions.
Figure 800-29 and Figure 800-30 are based on checklists found in Mechanical Seals
(1981, Chevron Research Company).
Note No attempt has been made to update these figures to make them consistent
with API 682. These figures still refer to API 610, and the terminology is inconsis-
tent with dual seal terminology in API 682. Though dated, this information is still
very useful in problem, cause, and solution identification.
The information is organized as follows:
• Figure 800-29 — Mechanical Seal Troubleshooting (Symptoms/Causes)
• Figure 800-30 — Causes of Seal Failure (Causes/Solutions)
Section 865, which follows, contains the solutions to the symptoms and causes
listed in these figures.
Fig. 800-29 Mechanical Seal Troubleshooting (1 of 2)
Causes
Symptoms (Each numbered cause is defined in Checklist 2)
Intermittent leak (may be accompanied by a popping 1-2-3-4-5-6-9-12-29-51-52-53-57-58-59
sound):
Exceeds local environmental emission limits: 1-2-3-4-5-6-9-12-29-51-52-53-57-59
Holds static pressure. Leaks when running: 1-2-3-4-5-6-7-8-9-10-12-29-50-51-52-53-57-59
Leaking under the sleeve: 13-14-15-16-17
Leaking between the gland and seal chamber face: 14-16-17
Sleeve fretting (under secondary seal): 8-18-19
Corrosion of metal parts 18
Rubbing: I.D. of throat, throttle bushing or 20-21-22-23-24-25-26-56
stationary face:
Rubbing rotating member 20-21-254-25-26-56
Spring wear 8-19
Spring breakage 8-18-19-28
Broken bellows: 1-2-3-4-5-6
Elastomer vulcanized to metal part: 18-29-30-31
Extruded O-ring 18-32-33-34-35
Elastomer (O-ring) deformed: 18-33-34-35-36
Worn drive lugs or pins 8-19-37
Excessive face wear: 1-2-3-4-5-6-9-12-18-29-38-44-49-50-51-52-53-57-59
Wide wear track: 24-25-56
Shiny spots on hard face: 41-42
Fig. 800-29 Mechanical Seal Troubleshooting (2 of 2)
Causes
Symptoms (Each numbered cause is defined in Checklist 2)
Chipped carbon. Outside diameter: 1-2-3-4-5-6-9-12-29-31-37-49-50-51-52-53-59
Chipped carbon. Inside diameter: 1-2-3-4-5-6-9-12-29-31-37-40-43-49-50-51-52-53-59
Blisters on the carbon face: 54
Broken carbon: 32-45-55
Eroded carbon: 11
Stationary or rotating face heat check: 1-2-3-4-5-6-9-12-18-29-31-37-49-50-51-52-53-59
Galling of rotating or stationary face: 18
Rotating assembly spun or slipped on shaft or sleeve: 29-32-37-46-47-48
Cartridge seal slid on shaft: 32-46-47-48
Coke or crystal deposits at atmosphere side of seal: 43
Fig. 800-30 Causes of Seal Failure (1 of 3)
Solutions
Causes (Each solution is described in Section 865)
1. Pump operating too close to shutoff. Increased 1-2-3-4-5-6-9-12-35
temperature is causing liquid to flash at the seal
faces.
2. Product temperature is above design temperature. 1-7-9-12-35
Liquid is flashing at the seal faces.
3. Suction pressure is below normal. Lack of vapor 1-9-11-12-35
suppression is allowing the liquid to flash at the seal
faces.
4. NPSHA below normal. Lack of vapor suppression is 1-9-10-12-14-15-35
allowing the liquid to flash at the seal faces.
5. Light ends are present and are flashing at the seal 1-9-10-12-35-59
faces.
6. Excessive seal tension. Increased face heat is 16-17-18-19-20
causing the liquid to flash at the seal faces.
7. Secondary seal is hanging up causing face separa- 21-22-23-24-25-37-54
tion.
8. Stationary face not aligned vertically to the shaft 25-53
center line.
9. Excessive seal cavity pressure. Increased face heat 1-9-20-26-35
is causing the liquid to flash at the seal faces.
10. Flush stream is improperly located. Velocity of the 27
flush is opening the faces.
11. Flush stream is improperly located. Velocity of the 27
flush is eroding the seal.
Solutions
12. Flush stream is improperly located. Lack of heat 12-28
removal is causing the liquid to flash at the seal
faces.
13. Inadequate lap joint between the sleeve shoulder 32
and end of the shaft.
14. Damaged gasket. 42
15. Sleeve not tight against the shaft shoulder. 34
16. Improperly seated gasket. 42
17. Damaged gasket surface. 42
18. Incorrect material selection. 33
19. Excessive axial shaft movement. 38
20. Loose or misaligned backup plate to bearing bracket 39
registered fit.
21. Loose or misaligned gland to seal chamber regis- 39
tered fit.
22. Misaligned throttle or throat bushing fit. 39
23. Misaligned mating ring fit. 39
24. Excessive sleeve runout. 39
25. Excessive shaft runout. 39
26. Insufficient clearance. 39
27. Inadequate bellows welds. 58
28. Bellows or springs are failing from sulfide stress 33
corrosion or chloride stress corrosion cracking.
29. Temperature excursion above normal. 43
30. Incorrect lubrication used during assembly. 40
31. Liquid is flashing at the seal faces and face 1-28-29-30-31-35-41-43-49-50-52-55
temperatures have exceeded design maximum.
32. Seal cavity pressure exceeded design pressure. 41
33. Teflon anti-extrusion backup ring is required. 44
34. O-ring groove measurements exceed design stan- 57
dards.
35. Incorrect O-ring size. 57
36. Excessive temperature. 33
37. Torque has increased due to marginal face lubrica- 1-8-9-16-17-18-20-26-29-30-31-43
tion or loss of liquid at the seal faces.
38. Faces are opening allowing abrasives to enter 45
between the faces.
Solutions
39. Seal chamber pressure is too high and seal face is 46
bowing.
40. Thermal distortion. 1-9-16-17-18-35
41. Mechanical distortion when assembled. 47
42. Improperly lapped face. 48
43. Leakage is coking or crystallizing on the I.D. 24-54
(atmospheric side) of the carbon.
44. Liquid is flashing. Soluble abrasives are coming out 1-9-16-17-18-20-30-31-35
of solution between the faces.
45. Secondary elastomer swelling. 33
46. Inadequate number of set screws. 41
47. Incorrect set screw material. 56
48. Inadequate set screw torque. 41
49. Pump not vented prior to startup. 52
50. Exterior flush is blocked in. 55
51. Fouled flush cooler. 50
52. Fouled seal chamber cooling water jackets. 49
53. Plugged flush line orifice. 51
54. Fluid that has soaked into the carbon expands on 54
startup causing face damage.
55. Oversized sleeve. 22
56. Excessive shaft deflection. 13
57. Incorrect seal balance ratio. 9
58. Quench steam is wet and is flashing at atmosphere 54
side of seal causing the seal to open.
59. Seal cavity pressure is too close to liquid’s vapor 12-29-30-31-35
pressure.
865 Solutions to Causes of Failure

Following are detailed descriptions of the solutions to the problems and causes
listed in Figure 800-29 and Figure 800-30.
Solution 1
Provide cooling with one of the following flush plans. Also refer to page 42 of the
API 610 in this manual for details on flush plans.
API 610 Flush Plan 2. When the liquid in the seal cavity is too close to its vapor
pressure, the liquid may be cooled below its boiling point using Plan 2 (see
Figure 800-31).
Fig. 800-31 Flush Plan 2 Courtesy of the American Petroleum Institute
Water circulating through the seal cavity water jackets removes enough heat from
the liquid in the seal cavity to keep the liquid below its boiling point. Contact the
pump manufacturer to establish the efficiency of the pump’s water jacket.
Over a period of time, minerals will plate out and foul the water jacket walls. This
fouling will reduce heat transfer from the seal cavity liquid to the cooling water. The
water jackets should be inspected for fouling during routine plant shutdowns, and
chemically cleaned, if necessary.
Plan 2 is not very efficient with pumps that are designed with internal seal cavities
due to the lack of the throat bushing. The throat bushing does act as a barrier
between the cooled liquid in the seal cavity and the hot liquid behind the impeller.
When using Plan 2, the seal cavity must be full of liquid which requires installing a
vent at the top of the seal cavity.
API 610 Flush Plan 21. If the liquid in the seal cavity is too close to its vapor
pressure and flashes to a vapor between the seal faces, the seal will fail prematurely.
One method of vapor suppression is to cool the liquid going to the seal cavity to
below the liquid’s boiling point (see Figure 800-32).
Cooling with Flush Plan 21 is accomplished by installing a heat exchanger (cooler)
in the flush line between the pump’s discharge and the seal cavity. The cooler can
use water or air as the cooling medium, depending on the amount of heat transfer
required, and the efficiency of the cooler. When it has been determined how much
heat must be removed, the cooler manufacturer can calculate the cooler and flush
line orifice size.
A maintenance problem that develops with water-cooled flush coolers is fouling of

the cooler’s shell side. Minerals plating out on the tube O.D. reduce the efficiency
of the cooler. The return cooling water lines also foul and reduce the flow of cooling
water through the cooler. The return cooling water piping and the shell side of the
cooler should be inspected for fouling during routine plant shutdowns and
chemically cleaned if necessary.
Furthermore, debris commonly becomes trapped in the small passages of these
coolers. A regular program of backflushing these coolers (cooling-water side) has
been helpful in reducing overall failure rates in one large refinery.
API 610 Flush Plan 23. If the liquid in the seal cavity is too close to its vapor
pressure, and flashes to a vapor between the seal faces, the seal will fail
prematurely. (See Figure 800-33).
One approach to vapor suppression is to cool the liquid in the seal cavity below the
liquids boiling point with Flush Plan 23. Flush Plan 23 is Flush Plan 2 with the
addition of a flush line, flush cooler and pumping ring. The pumping ring
(Figure 800-34) is mounted in the seal cavity and rotates with the pump shaft,
circulating the liquid from the seal cavity through a cooler and back to the seal
cavity.
Fig. 800-34 Pumping Ring
The liquid head developed by the pumping ring is not always sufficient to pump
through an air bubble in the flush line. This requires venting all air from the flush
line. The vent must be located at the highest point.
Flush Plan 23 is used predominantly to cool boiler feed water seals. Water cannot be
sealed successfully above a temperature of 180F with a single seal. Boiler feed
water is contaminated with soluble abrasives in the form of minerals or boiler
compounds. As the pressure of the liquid film of water between the seal faces drops
from the seal cavity pressure to atmospheric pressure, it will flash to a vapor if the
temperature between the faces is at or above 212F. When the liquid film between
the faces vaporizes, the soluble abrasives come out of solution and embed
themselves in the softer of the two faces. The soft face then acts like a grinding
wheel and will quickly destroy the opposing seal face. To keep the liquid between
the seal faces below 212F, it becomes necessary to cool the seal cavity to 180F or
less.
When the liquid in the seal cavity has been cooled to 180°F by Flush Plan 23, the
amount of the heat transfer required to hold at 180F is minimal. This reduces
cooler fouling. Cooling fouling occurs when minerals from the cooling water plate
out on the cooler’s tube O.D. and is more pronounced when the heat transfer is high.
A drain is required on the shell of the cooler. Periodic back flushing of the cooler’s
shell side will reduce fouling. The temperature of the flush stream should be
monitored with a temperature indicator in the flushing line. The T.I. should be
installed upstream of the cooler.
If the cooler loses its efficiency, it will require cleaning by chemical or mechanical
means. (See Figure 800-34.)
API 610 Flush Plan 32. This is the injection of a clean cool fluid to the seal from
an external source. (See Figure 800-35.)
To allow for flush flow, the pressure of the injection source must be greater than the
pressure in the seal cavity. The injection liquid must also be compatible with the
liquid being pumped, because the two get mixed together into the pump product.
Flush Plan 32 is considered for the following conditions:
1. Sealing liquids that contain abrasives. Injecting a clean cool fluid that is
compatible with the pumped liquid is a straightforward approach to eliminating
the abrasives from the seal cavity.
2. Sealing liquids that are hazardous to personnel or violate a pollution
standard if allowed to leak to atmosphere. An example of this is a stream
that has a high concentration of H2S. Flush Plan 32 isolates the seal cavity from
the H2S, with a non-H2S bearing flush stream. If the seal fails catastrophically,
the leak to atmosphere through the seal end plate throttle bushing must be less
than the non-hazardous flush flow to the seal cavity. A close clearance throttle
bushing in the seal end plate is required to restrict this leakage rate.
3. The liquid in the seal cavity is near its boiling point. If the liquid in the seal
cavity is near its boiling point, and vaporizes between the seal faces, the seal
will fail prematurely. This can be overcome by injecting a liquid with a low
vapor pressure, which is compatible with the pumped liquid, into the seal
cavity.
4. Pump Suction Loss. When a pump loses suction or has problems picking up
suction, the seal will fail prematurely from lack of liquid lubrication. Flushing
the seal with Plan 32 will ensure the seal of a constant source of liquid during
these brief periods of suction loss.
Flush flow velocities of 10 to 15 ft. per second through the seal cavity throat
bushing are desired to keep the pumped liquid from migrating back into the seal
cavity. When the exterior flush flow rate must be held to a minimum, it becomes
necessary to install a close clearance throat bushing in the seal cavity to maintain
the flow velocity through the throat bushing. The flush flow rate can be controlled
with an orifice or with visual flow indicators. Orifice sizes below 1/8 inch are
subject to plugging.
Solution 2
Work with operations to achieve operating conditions that allow the pump to operate
at or close to its best efficiency point. This may or may not be possible, depending
on the operational requirements. If not, the pump may need to be resized.
Solution 3
Resize the impeller.
Solution 4
Install a recirculation (bypass) line on flow control.
Solution 5
Install a low flow alarm.
Solution 6
Install a low flow shutdown device.
Solution 7
Work with operations to achieve the original design operating temperatures. Again,
this may, or may not be possible.
Solution 8
Measure the clearance between the sleeve balance shoulder and the back side of the
seal ring nose. If these two shoulders touch when the seal is assembled, the resulting
heavy face load and temperature increase will cause the liquid between the faces to
flash and the seal will fail.
Solution 9
Contact the seal manufacturer and investigate the possibility of redesigning the seal
face to reduce the face heat. This can often be done by changing face area or the
balance ratio.
Solution 10
Increase NPSHA by raising suction pressure, raising liquid level, or reducing line
losses. Reduced NPSHA will lower the pressure in the seal cavity. This reduction in
vapor suppression can cause the liquid between the seal faces to boil. The seal will
fail for lack of face lubrication.
Solution 11
Causes for low suction pressure include operational changes, a plugged strainer or
filter, a partially plugged suction line, or a partially closed suction valve. If any of
these conditions reduce the seal cavity pressure too close to the liquid’s vapor
pressure, the liquid between the faces will boil and the seal will fail.
Solution 12
When investigating a mechanical seal failure, where liquid boiling between the seal
faces is suspected as the cause of failure, 20°F should be added to the normal design
temperature. This increase in temperature is caused by liquid shearing and face heat
generated by the rotation of the mechanical seal. It is critical in liquid streams where
the vapor pressure is close to the pump suction pressure.
Solution 13
The major causes of shaft deflection are:
1. Imbalance. This may be an improperly balanced rotating element, a foreign
object stuck in an impeller, or a piece of the impeller broke off.
2. Cavitation. This is usually caused by the operating conditions. Check the
NPSHA and flow to see if they have deviated from the original design
requirements.
3. Design. Some packed pumps use the packing as a support for the rotating
element. If these pumps are converted to seals, the support is lost and shaft
deflection can become a problem.
4. Low Flow. Some pumps (those with Suction Specific Speed >10,000) are
especially vulnerable to operation at low flow rates.
Solution 14
Inspect the low level alarm if one exists. A faulty low level alarm will often lead to
continued operation with low levels.
Solution 15
Calibrate the level control instrumentation. Low levels contribute to a loss of vapor
suppression in the seal cavity. This will allow the liquid between the faces to boil
and damage the seal faces.
Solution 16
The seal may have been overcompressed on assembly. Obtain correct operating
length from seal drawing. Increased seal tension causes increased face heat. The
increased face heat may cause the liquid between the faces to boil and damage the
seal faces.
Solution 17
On impeller hung-between-bearings pumps, temperature-induced shaft growth can
increase the seal tension on the seal opposite the thrust bearing. This increased
tension causes increased face heat. The increased face heat may cause the liquid
between the seal faces to boil and damage the seal faces. If the pump case growth
does not compensate for the shaft growth, the cold setting of the seal opposite the
thrust end must be set to compensate for the shaft growth.
Solution 18
If incorrect springs of excessive length were installed during a repair, the increased
seal tension will cause increased face heat. The increased face heat may boil the
liquid between the seal faces. Loss of the liquid film will damage the faces. Always
reference the seal drawing for the correct part number, or contact the seal supplier
for the correct dimensions taken from the fabrication drawings.
Solution 19
During a repair, incorrect seal faces can be installed in the seal. Either an incorrect
face of excessive length or a reconditioned face that is too short. Both of these
errors will contribute to poor seal reliability. Always reference the seal drawing for
the correct part number, or contact the seal supplier for the correct dimensions taken
from the fabrication drawings.
Solution 20
Seal cavity pressures above the pressure limits of the seal will increase face
temperatures and boil the liquid between the faces. Loss of this liquid film will
damage the faces. Torque will also increase resulting in damage to the drive
mechanism of a pusher seal. A bellows seal operating in these conditions will often
fail the first bellows convolution on either the nose or drive collar end of the
bellows.
Solution 21
An undersized anti-extrusion ring will contribute to seal hang-up. The seal will
usually hold static pressure but will leak on startup prior to reaching full speed.
Solution 22
An oversized sleeve contributes to seal hang-up. The seal may hold static pressure
but leaks on startup prior to reaching operating speed.
The excessive radial pressure the oversized sleeve exerts on the inside of the carbon
seal ring can also cause carbon breakage.
Solution 23
An oversized secondary seal o-ring contributes to seal hang-up. The seal may hold
static pressure but leaks on startup prior to reaching operating speed. O-ring and
O-ring groove dimensions can be found in the Parker O-ring Handbook published
by the Parker Seal Company located in Lexington, Kentucky.
Solution 24
When some liquids evaporate, they leave an abrasive crystal deposit behind.
(Caustic is an example of a liquid which has this characteristic.) When these
deposits build up under the atmospheric side of the seal faces, they contribute to
premature seal failures. Crystals forming on the atmospheric side of the secondary
seal stop the secondary and rotating members from moving forward. This condition
is commonly called seal hang-up. The seal faces are also damaged by abrasives and
crystals that are centrifuged between them.
These problems can be avoided using API 610 Quench Plan 62 and quenching the
seal with water (see Figure 800-36). The water dilutes the seal leakage and stops the
formation of the crystals. The quench water exits through the seal end plate throttle
bushing or drain as shown in Figure 800-37.
Fig. 800-36 Flush Plan 62 Courtesy of the American Fig. 800-37 Seal End Plate Gland
Petroleum Institute
Maximum quench flow rates should be low enough to keep the quench water from
entering and contaminating the pump bearing housing.
Solution 25
Vertical misalignment of a rotating seal’s stationary (mating ring) face can be
caused by the following: The seal cavity face misaligned vertically to the shaft
center line; or the mating ring face misaligned vertically to the gland gasket face.
The mating ring anti-rotation pin is too long and won’t allow the mating ring to seat
properly in gland fit. Any one of these conditions will cause the rotating seal to
move forward and backward twice per revolution. This backward and forward
movement causes excessive face leakage, wear between the secondary seal and the
sleeve, spring wear (flat spots on the side of the spring), and wear to the seal’s drive
mechanisms.
Solution 26
Tandem seals are used when the pressure differential across a single seal is above
the seal’s maximum pressure capability. If the impeller-end seal (Figure 800-38)
has a differential capability of 1,000 psi and is required to seal 1,500 psi, the
differential pressure across the seal can be reduced by installing another seal and
applying 750 psi between the tandem arrangement. Both seals will now operate with
a pressure differential of 750 psi, which is well within their pressure range.
Fig. 800-38 Tandem Seal
Solution 27
The velocity of a seal flush stream impinging on the seal can cause erosion and
premature seal failure. This same jetting action can upset the seal balance and
contribute to excessive face leakage. One correction is to redirect the flush steam so
it does not make contact with the seal. Another and usually an easier method is to
move the flush line orifice further away from the seal. Take the pressure drop and
velocity increase at a point in the flush line where it does not come in contact with
the seal.
Solution 28
One reason for the seal flush is to remove the heat that builds up at the seal faces
and in the seal cavity from the rotation of the seal. A flush that enters the center of
the seal cavity is inefficient at removing heat at the seal faces. The flush should
enter the seal cavity through the gland plate as close to the face intersection as
possible. In high vapor pressure liquids where face flashing is causing seal failure,
the flush exit from the gland can be designed so that it surrounds the seal face to
increase the efficiency of the heat removal.
Solution 29
Flush coolers are used to cool the flush stream to the seal and suppress boiling of
the liquid between the seal faces. As the cooler fouls, it loses its heat transfer ability.
When repairing a seal that utilizes a cooler in its flush stream, always remove the
cooler shell and inspect the cooler for fouling. Chemical clean, hydroblast, sandblast
or replace the cooler as necessary. Back- flushing is usually inadequate once the
cooler has become fouled.
Solution 30
Boiler feed water pumps are an example of where the seal cavity water jacket is
used to remove heat and suppress boiling of the liquid at the seal faces. If the walls
of the water jacket foul, it loses its heat transfer ability. Always inspect and clean
the seal cavity water jacket for fouling.
Solution 31
Cooling efficiency will decrease as the inlet and outlet cooling water pressure loses
its differential and the flow decreases. There are two reasons for this: either the
return piping is fouling, or additional equipment has been added and the system is
overloaded. If the system is fouled, then clean or replace the piping. If the system is
overloaded, a larger cooler or a system upgrade may be required.
Solution 32
A sleeve-gasket leak will exist between the inside diameter of the sleeve and the
shaft. This type of leak is often mistaken for a seal leak. The shoulder of the sleeve
should be lapped against the shaft shoulder to ensure a good gasket surface. Apply
Prussian blue to one shoulder and rotate it against the other shoulder to ensure full
contact.
Solution 33
Improper material selection is a major cause of mechanical seal failures. Refer to
the Company’s material specification GF-G99876-7 and GD-G99877-7 when
selecting materials. Manufacturers of mechanical seals and elastomers are another
good source because their material manuals cover a wide range of liquids.
Physical changes in the seal parts such as loss of metal, O-ring swelling, and galling
of the seal or mating ring faces are signs of improper material selection.
Swelling of the secondary O-ring will cause seal hang-up and possible carbon seal
ring breakage.
Galling may occur when using a Tungsten-carbide seal ring against a Tungsten
carbide mating ring. (If the seal cavity pressure or the speed of the rotating face is
too great, the face will become damaged.)
Stress corrosion cracking of springs and metal bellows is also a determining factor
in material selection.
AM-350 is a material that is commonly used in manufacturing metal bellows.
AM-350 is susceptible to hydrogen embrittlement or sulfide stress cracking when
the following conditions exist:
1. The liquid being sealed contains 1 to 2 parts per million or more of H2S.
2. The temperature is below 300F.
3. Moisture (H2O) is present.
Bellows cracking from the atmospheric side of the seal occurs when normal leakage
crosses the seal face and contaminates the atmospheric side of the bellows with
liquid containing H2S. The moisture (H2O) can come from either a steam quench or
from the atmosphere. When the pump is shutdown and allowed to cool below
300F, the bellows is susceptible to sulfide cracking.
Bellows cracking from the liquid side may occur during a routine column repair.
Columns are usually steam purged and water washed when shutdown and steam
purged prior to starting up. When moisture (H2O) contacts an AM-350 bellows that
is contaminated with any H2S, and the temperature is below 300F, the bellows is
susceptible to sulfide cracking.
Type 316 stainless steel is another material that is commonly used in the
manufacturing of metal bellows. When a 316 stainless steel bellows operating in
temperatures between 150F and 300F comes in contact with moist chlorides, it is
susceptible to stress corrosion cracking. A typical example of this is a bellows seal
operating in the upper side cuts of a crude unit fractionation column. The chlorides
going overhead will contact the upper side-cut pump seals. The result is chloride
cracking of the 316 stainless steel seal bellows, or pusher seal springs.
Corrosion cracks in bellows cannot always be found by visual inspection. The cause
of the seal failure will often remain unknown or will be classified as an operating or
maintenance error. When the crack is observed visually or by pressure testing, the
failure is often attributed to fatigue cracking, poor bellows welds, or lack of bellows
strength.
The problem of hydrogen embrittlement or sulfide stress cracking in the AM-350
and chloride stress corrosion cracking in the 316 series stainless bellows can be
overcome with a material change. Neither Inconel X-750, Hastalloy-C, or
Inconel 625 is susceptible to cracking in chlorides or H2S. The type chosen is
determined by cost, availability, or the manufacturer’s preference.
Stress corrosion cracking is quite hard if not impossible to observe with the naked
eye. The Company’s materials lab has established a file on types and causes of
bellows failure throughout Chevron. This history, combined with their expertise and
special equipment, will help establish the cause of the failure.
Solution 34
Under-torquing of the impeller will often allow a leak between the sleeve and shaft
shoulder. The answer to this is to establish torque values and use the necessary tool
that will adapt the torque wrench to the impeller nut.
Solution 35
There is a molecular film of lubricant between the stationary and rotating faces of a
mechanical seal. Vaporization of this lubricant film can lead to an explosive
expansion and separation of the faces. The loss of lubricant between the faces will
also cause a rapid rise in face temperature. Face temperatures in excess of 2,000F
have been observed, the result being heat check and excessive face wear. Either the
temperature rise or the face separation can distress the seal faces to the point of
premature seal failure and possible pump fires.
In order to maintain this film of liquid lubricant between the seal faces, the pressure
differential between the liquid’s vapor pressure at pumping temperature and the seal
cavity pressure should be kept at a minimum of 25 psi for specific gravity heavier
than 0.6 and a minimum of 50 psi for a specific gravity lighter than 0.6. There can
be as little as ½ to 1 pint of liquid surrounding the mechanical seal in the seal cavity.
Liquid shear at the seal faces and around the mechanical seal will cause an increase
in temperature of approximately 20F plus at the seal faces. The 20F plus must be
added to the operating temperature when comparing the pressure differential
between the liquid’s vapor point and the seal cavity pressure.
Example:
Assume a mechanical seal is to be designed for a pump in reflux service with the
following conditions:
Liquid Propane
Specific Gravity Less than 0.6
Pumping Temperature 110F
V.P. at Pumping Temperature 220 psia
Suction Pressure 230 psia
Discharge Pressure 485 psia
Seal Cavity Pressure 255 psia
The Seal Cavity Pressure was found by adding 10% of the pump differential to
suction pressure. (This is a common rule-of-thumb for estimation – a pressure gage
is the only reliable way to know for sure.) To find the approximate temperature at
the seal faces, add 20F to the pumping temperature of 110F. A Hydrocarbon
Vapor Pressure Chart will show that the vapor pressure of propane at 130F is
260 psia. The seal cavity pressure of the reflux pump is 255 psia. This lack of vapor
suppression will allow the lubricant film between the seal faces to vaporize. The
seal will produce a high leak rate and an expected seal life of approximately
2 months.
To suppress the vaporization of the liquid film between the seal faces, there must be
a minimum differential of 50 psi between the liquid’s vapor point at normal
operating temperature and the seal cavity pressure. This can be achieved by one of
the following methods.
A Hydrocarbon Vapor Pressure Chart shows that the propane in the seal cavity will
be 50 psi below its boiling point if cooled to 85F. This temperature can be obtained
by installing a seal flush cooler in the pump’s seal flush line. Cooling of the liquid
in the seal cavity can also be obtained by dead-ending the seal cavity per API 610
Flush Plan 2. If cooling is not possible, the seal cavity can be pressured to 50 psi
above the liquid vapor point by using the pump’s discharge and a close clearance
fixed or floating throat bushing installed in the bottom of the seal cavity. (See
Solution 36.)
Solution 36
The normal throat clearance for most centrifugal pumps is .032 to .062 diametrical
between the seal cavity throat and the shaft sleeve. To maintain cavity pressure,
pumps using mechanical seals sometimes require that this clearance be reduced with
a fixed or floating throat bushing. Using a fixed throat bushing, the standard
clearance between the seal cavity throat and the shaft sleeve can be reduced to a
clearance of 010 diametrically. Clearances lower than this will result in excessive
bushing wear, caused by shaft deflection or normal misalignment of the pump shaft
through the throat bushing. If it becomes necessary to reduce the throat clearance
below .010, the floating throat bushing will compensate for shaft-to-throat
misalignment and, for brief periods, some shaft deflection. (See Figures 800-39A
and 800-39B.)
An expander retainer that grips the I.D. of the seal cavity retains multiple springs.
The springs hold the floating throat bushing against the bottom of the seal cavity.
The floating bushing can align itself around the shaft sleeve because of the large
clearance between the O.D. of the floating bushing and the I.D. of the seal cavity.
This type of bushing allows clearance as low as 0.004 diametrically between the
bushing I.D. and the shaft sleeve. The restriction of the throat clearance for pumps
using mechanical seals becomes necessary under the following conditions:
1. Incompatibility of the pumped liquid with the mechanical seal.
2. Seal cavities operating close to liquid vapor pressure.
When mechanical seal life is reduced because of the incompatibility of the pumped
liquid with the mechanical seal, it is often necessary to inject clean fluid to the seal
cavity. To keep the pumped liquid from migrating back into the seal cavity, while
keeping flush flow rates to a minimum, seal manufacturers recommend flush flow
velocities of 10–15 feet per second at the throat of the seal cavity. This can be
achieved with the reduced clearances of the fixed or floating throat bushing. Contact
the seal manufacturer for recommended clearances.
Fig. 800-39 Fixed and Floating Bushing
Most seal cavities operate close to the pump’s suction pressure. If the pressure in the
seal cavity is too close to the vapor pressure of the liquid, the heat added by the
rotation of the mechanical seal will cause the liquid to vaporize between the seal
faces and the seal will fail. Vapor suppression can be accomplished by pressurizing
the seal cavity liquid above its boiling point. This is done by installing a close
clearance throat bushing in the bottom of the seal cavity and flushing the seal from
the pump’s discharge. Caution must be exercised when dealing with vapor
suppression in this fashion. If the flow of liquid through the seal cavity is severely
restricted by the close clearance throat bushing and does not remove the heat
generated by the mechanical seal, the temperature will rise and the liquid between
the seal faces will again reach its vapor point.
Various styles of close clearance floating throat bushings and throttle bushings are
available from seal manufacturers. If long deliveries make them unavailable, they
can be fabricated from the floating bushing drawings (Figures 800-40 and 800-41)
which have proven to be quite successful. Close clearance bushing must be made of
a non-galling material such as carbon.
Fig. 800-40 Typical Floating Throat Bushings
Solution 37
If the secondary seal o-ring is of the wrong material, it may swell and cause the seal
to hang-up. This hang-up will keep the seal from compensating for axial movement.
The result is the faces open and excessive leakage occurs. Reference the
Corporations Material Recommendations for the correct material selection. The
Parker O-Ring Handbook is another source for O-ring material selection. This
handbook is published by the Parker Seal Company, located in Lexington,
Kentucky.
Solution 38
The outer face of the thrust bearing should not be able to move axially in its housing
more than .002 to .004 inches. If this axial movement is excessive, wear between
the secondary seal and the sleeve will occur. Flat spots will form on the side of the
springs from rubbing against the retainer.
Fig. 800-41 Typical Floating Throat Bushings
Solution 39
If rubbing between the rotating element and the stationary members occurs, suspect
internal misalignment, a bent rotating element, or failed bearings. The corrective
action to failed bearings is to replace them and investigate the cause of the failure if
reliability is a problem. To check the rotating element for straightness, rotate it with
its bearing journals setting on precision V-blocks or rollers. Some shops rotate the
element on precision centers. If the total indicator runout exceeds 0.002 inches,
repair or replace the shaft.
All registered fits must be measured and if the male to female clearance exceed the
manufacturers specification, they must be restored. Excessive register fit clearances
contribute to misalignment. Diametric clearance should not exceed 0.002. There are
exceptions that will require less than 0.002.
Not enough clearance at the throttle bushing, throat bushing or in the seal cavity can
also contribute to rubbing. Safe clearances for fixed throttle bushings can be
established by increasing the clearance to 0.005 greater than the pump’s wear ring
clearance.
If the normal throttle bushing clearance must be reduced to minimize the leakage of
hazardous liquids to the atmosphere, the floating throttle bushing (Figure 800-39B)
should be used.
The center line of the rotating element must be in the center of all the registered fits.
If misalignment exceeds a total indicator runout of 0.002, it must be corrected.
Solution 40
If the secondary seal o-ring is vulcanized to the sleeve, either the wrong material or
the wrong lubricant was used to assemble the seal. If the o-ring is compatible with
the temperature, change the lubricant that is applied between the o-ring and the
sleeve during assembly. Molybdenum Disulfide (Molycoat) has been successfully
used in this area.
Solution 41
Axial slippage of a cartridge seal sleeve away from the pressure in the seal cavity
will increase the seal tension. The increased tension will cause face temperatures to
go up. The liquid between the faces will boil and the seal will fail.
The reason the sleeve slides away from the seal cavity pressure is the sleeve has
more area exposed to the pressure than it does to atmosphere. There have been
various solutions to this problem. (1) Add set screws, (2) apply proper set screw
torque, (3) reduce the area of the sleeve exposed to the pressure, and (4) fabricate a
clamp that can be installed on the shaft which will add axial support to the sleeve.
Solution 42
There are various types of gasket designs used in a mechanical seal: o-rings,
wedges, v-rings, u-cups, and flat gaskets. The composition of the gasket can vary
over a wide range of material. Gasket damage is caused by excessive heat, excessive
pressure, corrosion, or mechanically. Heat, pressure, and corrosion problems, will
cause a gasket failure after the seal has been running for a while and can be
corrected by upgrading the material. Mechanical damage will usually cause a leak at
startup. Mechanical damage is caused by sharp shoulder corners, improper assembly
techniques, or improperly designed gasket containment area. Solutions to solving
the problem are quite straightforward once the cause of failure is determined.
Solution 43
Review the operating conditions for temperature excursions that exceed the normal
design temperatures. The increase in temperature can cause the liquid between the
seal faces to boil causing face damage and premature seal failure.
Drive torque loads will also increase when the liquid between the faces is lost,
causing the seal or the mating ring to spin or slip in the seal gland fit.
If the temperature increase is above the temperature limit of the elastomer, it can
cause the elastomer to vulcanize to the shaft.
If the temperature excursions cannot be controlled, the liquid around the seal will
have to be cooled. Use one of the flush plans in Solution 1 of this section.
Solution 44
If excessive pressure in the seal cavity extrudes the o-ring between the sleeve and
the inside diameter of the seal ring, it will lock the seal ring to the sleeve. This is
called seal hang-up. A teflon anti-extrusion ring, sometimes called a back-up ring is
installed between the o-ring and the nose of the seal ring. Its close clearance over
the sleeve will stop the o-ring from extruding under the seal ring.
Solution 45
Abrasives can enter between the seal faces during periods of face separation. Face
separation can occur during the pump startup, shutdown, or from pressure
fluctuations.
Assume one of the seal faces is carbon. During the periods of face separation, the
abrasives that enter between the seal faces will imbed in the carbon. The carbon in
effect will become a grinding wheel against the mating seal face. A straightforward
solution to this problem is injecting a flow of non-abrasive fluid into the seal cavity.
If this method cannot be used, an alternative is to design a mechanical seal with a
face combination of hard materials. These faces are less affected by abrasive wear.
In severe cases, face combinations of Tungsten carbide versus Tungsten carbide or
Tungsten carbide versus Silicon carbide can be used. This hard face combination is
also used when the abrasives are dissolved in the liquid. As the liquid pressure drops
across the seal faces toward atmosphere, it may flash to a vapor somewhere near the
inside diameter of the seal face. If this happens, the dissolved abrasive will solidify
between the seal faces. These hard face combinations have a tendency to gall in
liquids with low specific gravity, poor lubricating ability and liquids that are near
their boiling points. Given the correct operating conditions and liquid
characteristics, the seal manufacturer will recommend the correct face combination.
Another approach is to attempt to remove the abrasives from the flush stream. This
is done by installing a cyclone separator in the seal flush line.
Cyclones are very successful in removing the large particles that tend to cause
erosion damage in the seal cavity.
Some of the problems with cyclones are as follows:
1. Detailed engineering is required when orificing the cyclone system to ensure
maximum abrasive removal.
2. Abrasives at high velocities tend to reorifice the cyclone, causing efficiency
loss.
3. Orifice plugging will cause reduction or loss of the seal flush.
4. The smaller the particle size, the less efficient the cyclone separator is. This is
somewhat defeating, for it is the small particles that enter between the seal
faces, and these are the particles not being removed by the cyclone.
Solution 46
Measure the outside diameter of the narrow wear track on the mating ring face. If it
is smaller than the outside diameter of the seal ring, suspect bowing of the seal ring.
When the pressure in the seal chamber bows the seal ring inwards toward
atmosphere, the outer edge of the seal ring will prevent the liquid from entering
between the faces and the seal will fail. This is most predominant with carbon faces
that have thin cross sections, operating in high seal chamber pressures. The solution
is to install a seal with a stronger carbon cross section.
Solution 47
Shiny spots on the seal ring can be caused by distortion of the seal gland during
assembly. Seal gland distortion is caused by uneven torquing of the gland bolts, seal
gland or seal chamber gasket faces not flat, and narrow gland gaskets that allow the
gland to bend when the gland bolts are overtightened. Designs that permit full
contact between the gland plate and the seal chamber face are best.
Solution 48
The mating surfaces of the seal ring and the mating ring must be flat. A transparent
quartz optical flat and monochromatic light are required to check face flatness. A
lapping plate is used to produce the flatness required. If the surface inspected with
the optical flat is not flat, it is usually because the lapping plate was not flat or the
hard face was being distorted during lapping. Mechanical Seal Manufacturers
recommend face flatness to be held to three helium light bands. The face flatness is
measured using a helium light shining through an optical flat resting on the piece to
be measured.
Solution 49
Refer to Solution 1, Flush Plan 2, of this section for an explanation of seal chamber
water jacket fouling.
Solution 50
Refer to Solution 1, Flush Plan 21, of this section for an explanation of seal flush
cooler fouling.
Solution 51
The seal flush is used in various configurations to carry heat away from the seal. If
the flush orifice plugs, the temperature increase at the seal faces may cause the
liquid between the faces to boil and the seal will fail.
Solution 52
If the pump is not liquid-packed prior to startup, the seal faces may be damaged
from lack of liquid between the faces.
Solution 53
A mechanical seal is often required to operate in liquids that have corrosion rates
which require teflon gaskets and a stainless steel sleeve. If there is movement
between the teflon gasket and the stainless sleeve or shaft, corrosion fretting will
occur. When using a pusher seal, this occurs where the teflon gasket (Figure 800-42)
contacts the sleeve or shaft. Mechanical problems such as the stationary face not
being square with the shaft, or excessive axial movement of the shaft, cause axial
movement between the rotating seal member gasket and the sleeve or shaft. The
result of the corrosion fretting is damage to the sleeve and a premature leak between
the gasket and the sleeve or shaft.
Fig. 800-42 Pusher Seal
Corrosion fretting can be avoided by using a bellows (non-pusher) seal (Figure

800-43). The rotating seal member gasket “D” is static in a bellows seal. Corrosion
fretting can also be avoided by applying overlays that are corrosion resistant, to the
sleeve area under the rotating seal member gasket. The chrome oxide overlays are
porous; therefore, caution must be used that the bond between the overlays and the
parent metal be resistant to the corrosive liquid.
The desired overlay finish is 6 to 8 R.M.S. If these finishes are not achieved, the
rotating seal member gasket will “hang up” and tend to hold the seal faces apart.
Fig. 800-43 Bellows Seal
Solution 54
When hydrocarbons are sealed at elevated temperatures, a coke formation on the
atmospheric side of the seal will often result from normal seal leakage. This coke
formation will cause seal hang-up and face damage, and will distress the seal to a
premature failure. Steam injected through the seal end plate quench port aids in
stopping the coke formation by cooling the leakage.
The steam quench is also used to decrease blistering of the carbon seal face.
Blistering occurs when sealing viscous fluids such as asphalt, crude, and bunker.
When the liquid is warm, it will penetrate into the face of the carbon seal. If the
liquid film between the seal faces is allowed to cool when the pump is shut down, it
becomes viscous. When the pump is restarted, the shear force of the viscous liquid
film between the seal faces generates heat. The heat expands the liquid that has
penetrated the carbon face. The result of the liquid expansion is small pieces of
carbon raising or breaking away from the carbon face. These pieces of carbon hold
the faces open and allow excessive leakage. A controlled steam quench keeps the
faces warm and the liquid between the faces fluid.
Blister-resistant carbons are available and should be installed if blistering is causing
premature failures.
The maximum steam quench rate should be low enough to keep the steam from
entering and contaminating the pump bearing housing. The amount of steam quench
should be controlled by a back pressure regulator. Once the back pressure regulator
is adjusted to the desired flow, it remains constant. This overcomes the problem of
plant operators continually readjusting the amount of steam flow. Any pressure
regulator that is comparable to the Fisher Type 95L and has the following
construction features can be used. Body size, ¼ inch. Orifice size, ¼ inch. Body
material cast steel. Diaphragm and inner valve seat material, stainless steel.
Maximum inlet pressure and temperature, 300 psi, 450F. Reduced pressure ranges,
2 to 6 psi.
The problem with steam quenching high-temperature seal applications is keeping
the quench steam dry. When condensate enters the atmospheric side of a seal
operating in a high-temperature service and flashes to steam, the expansion of
condensate to steam will open the seal faces and allow for excessive leakage. This
will be heard as a popping sound from the seal end plate.
When installing a steam quench system, insulate the line from the steam source to
the seal end plate. Spend the extra amount necessary to obtain a hot, dry source of
steam. Figure 800-44 is a steam quench piping system that is used successfully in
high-temperature applications.
Fig. 800-44 Steam Quench Piping Systems
Solution 55
Refer to Solution 1, Flush Plan 32, of this section. If the flush is lost while
operating or remained blocked in during a startup that followed a repair, the seal
will fail.
Solution 56
The seal set screw material is usually 300 series stainless. This soft material allows
the tip of the set screw to dull and lose its holding torque. If this is allowing the seal
to slip or rotate, you may want to consider hardened steel set screws. There are
some advantages and disadvantages to steel set screws. The advantage is the
hardened tip cuts into the shaft or sleeve and will transmit higher torque loads
without slipping. The disadvantage is the high corrosion rate of steel. If the set
screw is operating in the liquid, its material must be compatible with the liquid. If
the set screw is used in the drive collar of a cartridge seal, which is located in the
atmosphere, and leakage or the atmosphere corrodes the hexagon socket the allen
wrench fits into, you won’t be able to loosen the set screws. Since they are
hardened, they cannot be drilled out. A simple cartridge seal repair now becomes a
major repair because the pump will have to be sent to the shop for seal removal.
If steel set screws are used in the cartridge seal drive collar, fill the hexagon socket
area with grease. This may give it some corrosion protection.
Solution 57
If the o-ring is extruding out of its containment area, compare the o-ring and the
o-ring groove dimensions with the recommended design data. O-ring and O-ring
groove dimensions can be found in the Parker O-ring Handbook published by the
Parker Seal Company located in Lexington, Kentucky.
Solution 58
The majority of bellows failures can be attributed to corrosion, stress corrosion
cracking, and fatigue. (See Appendix F.)
Inadequate bellows welds are often erroneously blamed for the bellows failure.
Bellows fabrication techniques developed by the major seal manufacturers have
greatly decreased the problem of poor weld quality. The best approach to
establishing the cause of a bellows failure is to have Chevron’s Materials
Laboratory examine the failed area of the bellows.
Solution 59
Light ends such as butanes, propanes, or ethane that have carried over or are
entrained in the liquid being pumped will cause seal failure if allowed to boil at the
seal faces. An example of this is a gas oil pump in a solvent deasphalting unit. The
gas oil has a low vapor pressure. The propane that is being used as a solvent in the
gas oil has a high-vapor pressure. If allowed to boil at the seal faces, it will cause
premature seal failure.
Revision History
Date Description Author Sponsor
1999 Initial release — —
May 2007 General revision JDEM JDEM
February 2011 Errata — Replaced references to PMP-EG-4662 with PMP-SC-4662 JDEM JDEM
(E) and updated its title. Added references to PMP-DG-4662-CRN and
PMP-DG-4664-FS in Section 850.
900 Inspection and Testing
Abstract
This section contains information on the purpose of and general practices for
inspecting and testing pumps. It covers several quality-control tests, giving general
guidance on when the respective tests may be cost-effective and appropriate. This
section emphasizes that inspection and testing can significantly increase the cost of
purchased equipment.
Contents Page
910 General Comments 900-2
920 Inspection Visits and Their Purposes 900-2
930 Non-Witnessed, Witnessed, and Observed Tests 900-6
July 1998 © 1998 Chevron USA Inc. All rights reserved. 900-1
900 Inspection and Testing Pump Manual
910 General Comments

Inspection (which includes witnessed testing) at the point of manufacture is
normally done by the ETC Quality Assurance (QA) team or by a qualified inspector
who contracts with the QA team to perform the work. The purpose of shop inspec-
tion is to obtain reasonable assurance that the equipment meets specification and
order requirements, displays good workmanship, and is free of significant defects or
damage before it ships.
The degree of inspection varies widely. Small, low-head, single-stage pumps in non-
flammable/non-toxic service usually do not warrant the expense of any shop inspec-
tion at all. A large, multi-stage horizontal centrifugal pump in critical service, on the
other hand, may warrant 12 to 16 inspection visits. Additional visits may be
warranted for inspection at the point of manufacture of the driver, lube oil console,
and/or gearbox.
On large, critical pumps, inspection coverage may also involve the project engineer,
a mechanical equipment specialist, and/or a specialist from the customer’s mainte-
nance organization. Coverage should be agreed upon as early as possible and
communicated to the Vendor. Typically, a project engineer or mechanical equip-
ment specialist (when involved) would cover only the following witness points:
• Pump starts and stops
• Thrust Test
• Performance Test
• NPSH Test
• Mechanical Seal Leakage Test
• String Test (of entire machinery train)
• Dismantle and Inspect After Performance Test
Occasionally, a full-time quality surveillance person is justified to observe the
assembly of very large critical-service pumps. This person would normally be a
specialist from the customer’s maintenance organization. This is only justified when
there is a clearly defined need, or when a representative is available full time on a
large project to follow all the machinery purchases.
The degree of inspection might be reduced with an alliance supplier.
920 Inspection Visits and Their Purposes

The most frequently made inspection visits for pumps, in descending order of occur-
rence, are:
• Final Inspection Prior to Shipment
• Pre-Inspection Meeting
• Witness Hydrostatic Test
• Witness Performance Test
• Other Inspections/Tests
Inspection is usually done by the ETC QA team or its designated contract inspector.
900-2 © 1998 Chevron USA Inc. All rights reserved. July 1998
Pump Manual 900 Inspection and Testing
Final inspection, or final inspection plus one of the other tests above, will cover
about 80% of all pumps. There are, however, about 17 principal inspection points
that can be considered for a pump. They are listed below in the normal manufac-
turing sequence. The cost of testing is usually added to the purchase cost in a pump
quotation.
1. Pre-Inspection Meeting (All types). Held to review specifications and other
requirements at the point of manufacture to verify there will be compliance.
This visit should always be made unless the value and criticality of the pump
will only support a final inspection visit. The pre-inspection meeting will help
to bring to light and resolve ambiguities that may delay final shipment, or result
in a pump which is not what the user specified.
2. Review of Pump-Casing Fabrication Drawings (All types). (By purchaser’s
inspector prior to start of fabrication.) This visit should always be made for
pumps with fabricated casings (including horizontal barrel pumps and fabri-
cated discharge heads and suction barrels of vertical turbine pumps). The
casings are pressure containing parts, but pump manufacturers frequently use
joint designs and materials that do not meet normal minimum requirements for
pressure vessels or pressure piping (as now required by API 610).
3. Visual Inspection of Fabricated or Cast Casings Before Machining (All
types). This visit should always be made for fabricated casings, even when
sound joint designs are on the fabrication drawings. Welds frequently have
flaws that can be found visually; weld repairs must be made before machining
since some distortion from welding is inevitable. Large cast casings should be
visually inspected prior to machining (when nozzle size exceeds about 18
inches).
4. Non-Destructive Examination (NDE) of Fabricated or Cast Casings (liquid
penetrant, magnetic particle, ultrasonic, radiographic) (All types). Pump
casings are normally accepted on the basis of visual examination and a
successful hydrostatic test. NDE should not be added unless it is clearly
justified by the service conditions, material characteristics, or established
specification requirements. A materials engineer, QA engineer, or both should
be consulted if supplementary NDE is being considered for other reasons. NDE
in itself is frequently inexpensive; it is the resulting repairs to upgrade cast-
ings which have been subjected to NDE that can be very expensive. When
some form of supplementary NDE is specified, an acceptance standard must
always be specified as well. When some form of supplementary NDE is speci-
fied, it should always be witnessed (radiographs interpreted when radiography
is specified) and must be identified as a witness point on the pump data sheet.
(See Section 930 for definitions.)
5. Visual Inspection of Fabricated Baseplates Before Machining (All types).
This visit should be made to ensure proper fabrication of large baseplates (when
driver is over about 250 HP). Weld quality and inadequate weld size have been
problems on large baseplates.
6. Hydrostatic Test (All types). A casing hydrostatic test is virtually always

performed to ensure integrity of the case. Witnessing the test is warranted for
pumps in toxic, hazardous, or flammable service or for alloy casing materials.
7. Parts Inspection Prior to Assembly (All types). Parts are usually not
inspected prior to pump assembly. The purpose, when performed, is to ensure
the parts are dimensionally correct, are of the correct materials, and to catch
any visible defects. Parts inspection might be warranted for some multi-stage
pumps or for very large pumps (nozzle size over about 18 inches). Consult a
mechanical equipment specialist, QA engineer, or both if parts inspection prior
to assembly is being considered. If it is to be done, it should be shown on the
pump data sheet as a witness point.
8. Incremental (Stack) Balancing of Rotor (Centrifugal). This test is to assure a
balanced multistage rotor by correcting imbalance on each component as it is
assembled on the rotor. This test is usually not witnessed. It can require days of
inspector time since the rotor is balanced several times during the course of
assembly. It may be warranted for critical horizontal multi-stage pumps.
Consult a mechanical equipment specialist if incremental balancing is being
considered. If it is to be done, it should be shown on the pump data sheet as a
witness point.
9. Final Balancing of Rotor (Centrifugal and Rotary). This test is usually not
witnessed. It may be warranted for critical horizontal multi-stage pumps in
order to ensure minimum vibration. Consult a machinery specialist if this is
being considered. If it is to be done, it should be shown on the pump data sheet
as a witness point.
10. Thrust Test (Centrifugal). This test is always witnessed. The purpose of this
test is to see if the axial force transmitted by the pump rotor exceeds the design
load capability of the thrust bearing. This is accomplished by installing a load-
cell or equivalent device to measure the thrust while operating the pump at
several points on its curve. The maximum design thrust based on the specified
bearing life must be known before beginning the test. A thrust test is rarely
done but may be justified on some large, horizontal, multi-stage pumps, and on
pumps that have thrust loads that extend the design beyond the Vendor's normal
experience, and in similar cases.
11. Pump Starts and Stops (Centrifugal). Pump starts and stops are sometimes
specified for pumps that have bushings lubricated by the pumped fluid, to
determine if bushing wear is excessive. This would include horizontal multi-
stage, vertical line-shaft, and some vertical in-line pumps. The justification is
greater for pumps with relatively hard bushing materials that have galling
tendencies. The number of starts and stops (usually 10) must be specified.
Dismantling and inspection after the performance test must be specified to
allow inspection of the bushings for galling or metal transfer.
12. Performance Test (All types). This visit is required if performance of the
pump is critical and significant losses will occur if the pump fails to perform as
required. The visit is recommended for any multi-stage horizontal pump to
Pump Manual 900 Inspection and Testing
verify the pump is satisfactory mechanically and meets performance require-

ments. If it is to be done, it should be shown on the pump data sheet as a
witness point.
All pumps purchased in accordance with API Standard 610 are given a perfor-
mance test, but witness of the performance test, if required, must be specified .
13. NPSH Test (All types). An NPSH test is normally recommended for a centrif-
ugal pump when the margin between NPSHA and NPSHR is within four feet,
for reciprocating pumps when the difference is 3 psi (not feet), or less, and for
rotary pumps when the difference is 4 psi or less. This test verifies that the
quoted NPSHR is met. When an NPSH test is required, it should always be
witnessed. This should be shown on the pump data sheet as a witness point.
14. Dismantle Pump and Inspect Parts After Performance Test (All types).
This is an expensive option, only done routinely for barrel-type horizontal
pumps. It is also done if some form of distress is noticed during the perfor-
mance test. It can be considered for all horizontal multi-stage pumps, but there
is always the possibility of reassembly errors that will not show up until job site
startup. Consult a mechanical equipment specialist if this is being considered. If
it is to be done, it should be shown on the pump data sheet as a witness point.
15. String Test of All Job Equipment (All types). A string test is a mechanical
(and sometimes performance) test of all the ordered equipment assembled
together on the job baseplates (driver, gearbox, pump, etc.). The purpose of the
test is to assure all components will operate together without excessive vibra-
tion, or other distress. Note that performance and NPSH tests are done with
calibrated shop drivers and not on the job baseplate. String tests are usually
limited to multi-stage horizontal pumps in critical services or for locations
where re-work/repair is very difficult. Consult a mechanical equipment
specialist if a string test is being considered. A string test should always be
witnessed. If it is to be done, it should be shown on the pump data sheet as a
witness point.
16. Final Inspection (preferably before finish painting, all pump types). This is
the most basic inspection visit. When “shop inspection” is checked on the pump
data sheet but no other witnessed tests are specified, inspection is limited to
final inspection (and a Pre-Inspection Meeting if the value and criticality of the
pump will support its cost). Final inspection includes, but is not limited to:
– Review of equipment against specifications and data sheets line by line.
– Verification of correct pump materials.
– Dimensional check against company-reviewed vendor outline drawings.
– Review of performance curve and test log against performance require-
ments (if either an unwitnessed or witnessed performance test is required).
– Review of hydrostatic test certification and any other data required by the
specification.
– Visual inspection for defects or damage.
17. Final Inspection After Finish Painting (all pump types). This visit (or
several visits) is warranted when sophisticated painting systems are specified
(principally for offshore or coastal applications).
930 Non-Witnessed, Witnessed, and Observed Tests

Inspection and testing of a pump or parts of a pump by the manufacturer may be
non-witnessed, witnessed, or observed:
Non-Witnessed means that the manufacturer does the required test and certifies the
results; the test results are reviewed by the Purchaser’s inspector if the pump is
inspected. Production is not stopped as it is for “witnessed” tests.
Witnessed means that a hold is applied to the production schedule and the test
carried out with the Purchaser’s inspector present. This may result in a double test.
The Vendor will include in his bid an extra cost for witnessing a test.
Observed means that the Purchaser requires advance notification of the test’s
timing. However, the test is performed as scheduled, and if the Purchaser’s inspector
is not present, the Vendor may proceed with the test. A hold is not placed in the
production schedule. Since only one test is scheduled, the Purchaser’s inspector can
expect to be in the factory longer than for a witnessed test, while debugging of the
pump or test stand equipment is done. The Vendor will include an extra cost for
observing a test that is about half the cost of witnessing a test. The differential cost
between observed and witnessed tests can be more than offset by increased
inspector’s time for observed tests; observed tests may also be missed because of
insufficient advance notification by the Vendor. Specifying an observed test is there-
fore not recommended. If a test is important enough to warrant the presence of an
inspector or engineer, a witnessed test should be specified.
1000 Pump Installation
Abstract
This section contains general guidelines and checklists for installation of new
pumps and reinstallation of existing ones.
Contents Page
1020 Installation Checklist 1000-2
January 1991 © 1991 Chevron USA Inc. All rights reserved. 1000-1
1000 Pump Installation Pump Manual
1010 Introduction
Although these guidelines focus on new installations, many aspects also apply to
existing installations where equipment has been repaired, rerated, retro-fitted, etc.
Proper pump installation helps ensure trouble-free startup and long term reliability.
If installed improperly, even the best pump can be a source of costly maintenance.
For example, a pump running perfectly in a well-designed installation may experi-
ence many problems if moved to a poorly designed installation. The opposite is also
true. Additionally, an unreliable pump installation can affect plant availability, even
if fully spared. Two fully spared but poor installations never equal a single, reliable
installation.
This section discusses general items that should be considered to ensure reliable
pump installations.
Although the items on the checklist in Section 1020 are simple, it is surprising how
often they are not checked or not properly followed. Typical examples include:
Baseplates improperly grouted. This leads to excessive deflection and vibration,
shortening pump and driver life.
Pump and driver shafts improperly aligned. This shortens bearing, seal, and
coupling life.
Excessive pipe cold-springing. This distorts the pump case and baseplate, short-
ening bearing, seal, and coupling life.
These problems can be expensive. For example, the cost of three or four seal fail-
ures may equal the original pump purchase price. There is a significant economic
incentive to achieve good, reliable pump installations from the start.
1020 Installation Checklist

The attached checklist covers most common onshore and offshore pump installa-
tions. Where appropriate, installation items specific to certain types of pumps are
listed separately. This checklist does not include specific acceptance criteria (align-
ment tolerance, pipe forces, etc.). These criteria should be developed for the specific
application according to the manufacturer's service manual, applicable industry
standards, and local practices.
The following checklist assumes that there are no problems inherent in the equip-
ment. This means that the pump design matches the intended service, materials are
correct, and the pump has been properly manufactured. It also assumes that the
associated process and auxiliary piping system, foundations, and baseplate are prop-
erly designed and built.
1000-2 © 1991 Chevron USA Inc. All rights reserved. January 1991
Pump Manual 1000 Pump Installation
Pump Installation Checklist

Initial Date
General - All Pump Types
1. Data Sheets (Pump & Driver)
Drawings:
Properly Completed Data Sheets ________ ________
Pump Cross Section/Performance Curves ________ ________
Shaft Seals/Packing ________ ________
Coupling ________ ________
Auxiliary Systems (lubrication, sealing, cooling) ________ ________
Shop Performance/Mechanical Test Results ________ ________
Service Manual (covering all major pump train components) ________ ________
Safety Instruction Sheets ________ ________
Maintenance History (existing pumps) ________ ________
Performance/Vibration History (existing pumps) ________ ________
Spare Parts List ________ ________
2. Foundations and Grouting

Concrete Foundations:
Concrete properly cured before grouting ________ ________
Concrete surface properly roughened and cleaned ________ ________
Mounting plate metal surface clean/properly painted ________ ________
Mounting plate corners properly rounded ________ ________
Proper grout material/thickness ________ ________
Coupling disconnected before levelling ________ ________
Mounting plate levelled ________ ________
Pump mounting pads horizontal and at same elevation ________ ________
Grout properly cured ________ ________
No hollow areas in grouting ________ ________
Anchor bolts properly tightened ________ ________
Other Foundations:
Mounting plates located at load bearing support beams ________ ________
Shims of non-corrosive material ________ ________
Baseplate/skid properly levelled ________ ________
Bolting properly tightened ________ ________
Pump Installation Checklist (continued)

Initial Date
3. Equipment Condition (Before Installation)
Storage protection requirements met ________ ________
Internal/external cleanliness checked ________ ________
Rust preventative/lubricant compatibility acceptable ________ ________
Baseplate/skid straight and flat ________ ________
Pump feet free of burrs ________ ________
Pump and driver not bolt-bound ________ ________
Nozzle gasket surface finish acceptable ________ ________
Pump and driver rotates freely by hand (no excessive radial or lateral play) ________ ________
Alignment shims (number, material, condition) acceptable ________ ________
Coupling condition acceptable ________ ________
Seal leak free (static or running test) ________ ________
Rotation arrow permanently marked, checked ________ ________
Oil level marked, checked ________ ________
Alignment jack screws ________ ________
4. Process Piping Systems

P&ID Check ________ ________
Proper blinding & tagging suction/discharge lines ________ ________
Suction screen location and orientation correct ________ ________
Check/block/bypass valves location and orientation correct per P&ID ________ ________
Correct gasket design/material installed ________ ________
Relief valves properly set ________ ________
Piping/pump flanges parallel ________ ________
Check for cold spring with indicators ________ ________
Correct piping supports (hot and cold) installed ________ ________
Expansion joints properly adjusted ________ ________
Piping strain acceptable (hot service check) ________ ________
Check ANSI Class on flanges for correct rating ________ ________
Eccentric reducers - flat side up, angled side down ________ ________
Suction piping sloped to prevent vapor traps ________ ________
Piping clean as dictated by installation ________ ________

Initial Date
5. Auxiliary Piping Systems
General:
Piping properly braced ________ ________
Piping cleaned ________ ________
Piping proper size and material ________ ________
Piping properly sealed/socket welded ________ ________
Piping properly traced and insulated ________ ________
Proper thread lubricant used ________ ________
Check instrumentation per any available drawing ________ ________
Seals:
Flush properly connected ________ ________
Flush system orifice/strainer/cyclone separator correctly located ________ ________
Quench properly connected/plugged ________ ________
Vent properly connected/plugged ________ ________
Drain properly connected/plugged ________ ________
Packing:
Lantern ring location/flush connection correct ________ ________
Cooling:
Stuffing box properly connected ________ ________
Lube oil cooler connected ________ ________
Pump pedestals connected (hot services) ________ ________
Lubrication:
Piping properly cleaned/water-free ________ ________
Proper orifices installed (pressure systems) and tagged ________ ________
Housings flushed with correct oil ________ ________
Screwed piping properly assembled (without tape) ________ ________
Oil levels correct ________ ________
No bends/kinks between housing & oiler (purge mist) ________ ________
Oilers properly vented (at bearing cap) ________ ________
Oil sump bottle non-vented (purge mist) ________ ________
Oil sump bottle vented (pure mist) ________ ________
Oil mist fittings - proper sizes ________ ________
Oil mist - check flow ________ ________

Initial Date
Case Vents/Drains:
Vent properly piped ________ ________
Drain properly piped/plugged ________ ________
Warm-up connection properly piped for hot pumps ________ ________
Ports properly connected for Sundyne pumps ________ ________
6. Couplings and Alignments

Alignment shims entirely under each pump foot ________ ________
“Soft foot” checked ________ ________
Avoid spring effect due to an excessive number of shims ________ ________
Correct axial spacing between pump & driver shafts (hot and cold) ________ ________
Axial float of motor driver checked ________ ________
Motor driver properly positioned (limited end float coupling) ________ ________
Proper coupling interference fit ________ ________
Coupling lube holes properly positioned ________ ________
Gear coupling properly greased and gaskets installed ________ ________
Flexible coupling shim packs properly installed ________ ________
Pump/driver cold aligned & bolting properly tightened ________ ________
Coupling hub runout checked ________ ________
Cold alignment to manufacturer's recommended tolerances ________ ________
Hot alignment checked ________ ________
Pump/driver properly dowelled ________ ________
Coupling guard installed ________ ________
7. Miscellaneous
Proper packing installed, gland not cocked ________ ________
Special tools available ________ ________
Driver rotation checked (prior to coupling) ________ ________
Instrumentation, alarms, and shutdown in proper condition ________ ________
Pump/driver properly protected during idle period ________ ________

Initial Date
Other Specific Pump Types
1. Vertical Centrifugal
Curb ring properly leveled and grouted ________ ________
Pump assembly within acceptable vertical deviation (proper lift with axial ________ ________
setting)
2. Rotary Pumps
Internal cleanliness checked ________ ________
Belts and sheaves properly aligned ________ ________
Matched belt sets used ________ ________
Belts properly tensioned ________ ________
Suction strainer cleanliness checked ________ ________
Pressure gage installed before/after strainer ________ ________
Jacket piping properly connected ________ ________
Belt guards installed ________ ________
3. Reciprocating Pumps
Baseplate/foundation bolts retightened ________ ________
Belt and sheaves properly aligned ________ ________
Matched belt sets used ________ ________
Belts properly tensioned ________ ________
Belt guards installed ________ ________
Plunger/piston/packing correct size ________ ________
Rod straightness/runout/surface condition acceptable ________ ________
Packing properly installed and lubricated ________ ________
Packing gland properly adjusted (initial) ________ ________
Correct bladder pressure/fluid level in pulsation dampers ________ ________
Lubricators operating correctly ________ ________
1100 Startup and Troubleshooting
Abstract
This section contains a pump startup checklist and troubleshooting guides. Infor-
mation on troubleshooting mechanical seal problems is included in Section 800.
Although some references are made to vibration problems, the reader is referred to
the CUSA IMI Candidate’s Manual or to the “Other References” section in this
manual for more information on troubleshooting those problems.
Contents Page
1110 Startup Checklist 1100-2
1120 Introduction to Troubleshooting 1100-10
1130 Troubleshooting Insufficient Flow/Pressure
from Centrifugal Pumps 1100-12
1140 Procedure for Performance Monitoring Centrifugal Pumps
(“Curving the Pump”) 1100-13
1150 Centrifugal Pump Troubleshooting Checklist 1100-20
1160 Vertical Turbine Pump Troubleshooting Checklist 1100-22
1170 Metering Pump Troubleshooting Checklist 1100-24
1180 Reciprocating Pump Troubleshooting Checklist 1100-27
1190 Rotary Pump Troubleshooting Checklist 1100-30
1100 Startup and Troubleshooting Pump Manual
1110 Startup Checklist

Startup Items for All Pumps
Initials Date
Piping
1. Inspect check valves for correct flow direction. ________ ________
2. Verify that all process and auxiliary piping is free of cold spring at the pump flanges. ________ ________
3. Flush and, where appropriate, chemically clean all process and auxiliary piping which ________ ________
has been opened or repaired. Verify cleanliness with filters or flange screens.
4. Perform thorough leak checks of all welds, flanges, and fittings during any hydrotests. ________ ________
Do not permit hydrotests against pump mechanical seals unless they have been rated
for the higher pressures. Test with a fluid compatible with the process whenever
possible. Otherwise, thoroughly drain and flush the test fluid.
5. Verify that all operational personnel are familiar with the suction and discharge piping ________ ________
systems as well as control valves. Do the same with all auxiliary piping and turbine
driver steam piping. Personnel should also be familiar with suction-to-discharge
bypass systems and their purposes for startup, low flow, or high pressure control.
6. Prepare a list of piping blinds to be removed prior to startup. Include a signoff column ________ ________
where operating personnel can certify and document the status of the blinds.
7. Open seal flush valve (if equipped). ________ ________
8. Ensure that all drain valves are properly plugged. ________ ________
9. Verify that all auxiliary piping for sealing and flushing is connected to the right ports at ________ ________
the pump gland.
10. Verify all safety tags are removed from system valves and switchgear. ________ ________
11. Inspect suction screen shortly after startup. ________ ________
12. Prime the pump by filling suction line and pump with process fluid and bleeding air to ________ ________
atmosphere or relief as appropriate.
13. Inspect for static (pre-startup) leaks. ________ ________
Instrumentation
1. Verify operating personnel are familiar with all associated processes and auxiliary ________ ________
instrument systems. Ensure the systems have been sufficiently calibrated, loop
checked, and functionally tested.
2. Prepare a list of instrument systems which must be tested on-line. Ensure that test ________ ________
provisions do not impair personnel safety or machinery reliability.
3. Verify that all systems pressure safety relief valves have been tested and set per local ________ ________
policy. Ensure that relief block valves have been locked open with a locally accepted
method.
4. Check vibration systems for proper installation, calibration, and alarm/shutdown ________ ________
settings.
5. Verify Automatic Pump Start (APS) systems are installed and operationally tested. ________ ________
Pump Manual 1100 Startup and Troubleshooting
Initials Date
Shaft Mechanical Seals
1. Review seal-flush plan, function, and operation. ________ ________
2. Leak test the mechanical seal(s) in place at a pressure corresponding to the maximum ________ ________
design stuffing box pressure. Repair leaks before starting.
3. Verify that all auxiliary flushing, quenching, and cooling systems provide flow at the ________ ________
design pressures, temperatures, and rates. When critical to operational reliability or
safety, insure these auxiliary systems are alarmed and/or have redundant features.
4. Bleed and fill the stuffing box with flush fluid prior to starting. (This is particularly impor- ________ ________
tant for vertical pumps and high fluid vapor pressure pumps.)
5. Check the following for any external seal-flush cooler: ________ ________
– Open cooling water valves to flush cooler. ________ ________
– Vent-tube side (flush) of cooler at high point for pumping-ring systems, if ________ ________
non-hazardous.
6. Check the following for double seals with external pressurizer/circulator: ________ ________
– Fill reservoir with proper buffer (barrier) fluid. ________ ________
– Start buffer circulating pump. ________ ________
– Set buffer fluid backpressure to a minimum of 25 psi above impeller-side ________ ________
pressure at inner seal.
– Test buffer fluid low pressure, low flow, and low level alarm settings. ________ ________
7. Check the following for double seals with pumping rings: ________ ________
– Fill buffer fluid reservoir with proper fluid. ________ ________
– Vent all air out of buffer system. ________ ________
– Pressurize buffer system to 25 psi above impeller-side pressure at inner seal. ________ ________
– Inspect all buffer pressure connections for leaks. ________ ________
– Test low pressure, low level alarm, and any other alarms. ________ ________
8. Check the following for tandem seals with pumping rings: ________ ________
– Fill buffer fluid reservoir with proper fluid. ________ ________
– Vent all air out of buffer system. ________ ________
– Test buffer fluid high pressure, high level, and low level alarms. ________ ________
9. Verify that all real vent and drain parts are properly connected and/or plugged. ________ ________
10. During startup, note and document all seal leaks, their location, and whether they ________ ________
appear to be increasing. Generally, any leak requires a repair. Seals rarely “run in” and
stop leaking.
11. When required by local regulations, check the running seal for excessive fugitive ________ ________
hydrocarbon emissions.
Initials Date
Shaft Packing
1. Verify the flush supply to the lantern ring, if included, is at adequate pressure. ________ ________
2. Verify that the stuffing box cooling jacket water, if supplied, is flowing. ________ ________
3. Verify that the shaft turns freely to check for over-tightened packing. Once the pump is ________ ________
running, tighten adjustable packing to a slow drip to ensure adequate lubrication.
Tighten the packing by evenly turning the gland boltnut a quarter turn at a time. (Avoid
initial overtightening of packing materials such as Graphoil. Consult packing manufac-
turer recommended tightening procedures for initial run-in.)
4. If a pump with self-adjusting packing has a leak greater than a drip, shut down and ________ ________
replace the packing.
Bearings
1. Ensure the bearing housings are clean of grit, sand, metal shavings, or other debris. ________ ________
Verify that bearing housings and lube oil systems are filled with the proper oil and filled
to the correct level.
2. Monitor bearing vibration throughout startup. Do not exceed prescribed danger levels ________ ________
at any time. Watch for increasing vibration levels as an indicator of deteriorating
mechanical conditions.
3. Observe bearing housing or oil temperatures throughout startup. Do not exceed the ________ ________
prescribed danger levels any longer than necessary to shut the pump down. On ball
bearing-type pumps, this is best done by measuring the bearing housing temperature.
On pressure lubricated sleeve bearings, use thermocouples to sense the bearing
metal temperature or the exit (outlet) oil temperature.
4. Drain an oil sample from the bearing housings to look for signs of dark oil, metallic ________ ________
debris, or other contamination. Shut down and investigate if contamination is found
after a short period of time.
5. Check that oiling rings or slingers provided with ball bearings are rotating and deliv- ________ ________
ering oil to the bearings.
6. Bearing temperature should not exceed 180°F. Do not run water over hot bearings. ________ ________
Such action is more likely to contaminate the oil with water than it is to cool the
bearing. A hot bearing is a sign of an overload or impending failure.
Lubrication Systems
1. Check the following when starting an oil-mist system: ________ ________
– Verify the reservoir is filled with the correct oil and not over filled. ________ ________
– Verify air pressure regulator setting. ________ ________
– Test generator alarm lights. ________ ________
– Test the low pressure, high pressure, low temperature, high temperature, and low ________ ________
oil level switches.
– Verify pressure at the end of the main header is the same as the generator ________ ________
pressure.
– Check for visible signs of mist at the last out-of-service pump on the header. ________ ________
Initials Date
Lubrication Systems (continued)
2. Check the following when starting equipment with pressure lube systems: ________ ________
– Verify reservoir is filled to the proper level with the correct lube-oil. ________ ________
– Check for installed breather, and plug all reservoir drain valves prior to system ________ ________
run-in.
– Check for presence of flow restriction orifices at individual bearings (if required by ________ ________
Vendor’s design). Check for correct sizes at each location.
– Run-in lube system prior to initial main machinery operation. Check for leakage. ________ ________
Watch filter differential pressure, and check for cleanliness as measured by debris
caught in filters or temporary in-line screens. (Refer to API Standard 614 for guid-
ance on Cleanliner Standards.)
– Sample lube-oil prior to on-line operation and change if necessary. Replace filter ________ ________
elements.
– Establish cooling water flow to oil coolers. ________ ________
– Check settings of lube heater pressure regulators and relief valves. ________ ________
– Test alarm and shut down switches. ________ ________
– Verify operation of the auxiliary lube pump during main machinery startup, shut- ________ ________
down, and on low lube pressure. (Shaft-driven main lube pumps that are situated
above the reservoir are usually primed by the auxiliary pump, which requires the
auxiliary pump to be running prior to starting the equipment.)
– Check for a minimum lube-oil temperature of 100-110°F prior to main machinery ________ ________
startup.
– During initial on-line operation, check for oil flow at each of the bearing sight ________ ________
glasses.
– Oil temperature rise through bearings should not exceed 50°F when inlet oil ________ ________
temperature is at or below 110°F.
Motor Drivers
1. Verify that all coupling guards are installed and bolted down. ________ ________
2. Ensure motor heaters, if installed, are working when motor is off. ________ ________
3. “Bump” motor to check for correct rotation. Verify that rotation arrow on pump matches ________ ________
pump drawing or data sheet.
4. Determine how many restarts the motor is permitted in one hour’s time and follow ________ ________
those limits.
5. Do not frequently push the “Stop” button before motor has reached full speed, particu- ________ ________
larly on larger motors. Do not push the “Start” button until motor has coasted to a
complete stop. Such actions may trip or even damage the electrical equipment.
6. If motor repeatedly trips on start attempts, check: ________ ________
– Process for excessively high flow or pressure demands. (Some centrifugal pumps ________ ________
can only be started against a nearly closed discharge valve. Many positive
displacement pumps must be started on a suction-to-discharge bypass.)
Initials Date
Motor Drivers (continued)
– Improper electrical switchgear relay or circuit breaker settings. ________ ________
– Low system voltage during starting. ________ ________
– Higher-than-design fluid viscosity in the pump. ________ ________
– Mechanical problem such as bad motor or pump bearings or internal rubbing. ________ ________
Steam Turbine Drivers

1. Open drains to remove condensate from turbine casing and steam lines. ________ ________
2. Overspeed test the turbine. ________ ________
3. Run-in carbon rings. ________ ________
4. Open the steam exhaust valve slowly. ________ ________
5. Warm the turbine by cracking open the supply steam block valve or by opening its ________ ________
warmup bypass.
6. Close drain valves when condensate has been purged. ________ ________
7. Draw a sample of bearing oil to visually check for water content. Drain and refill oil if ________ ________
water is found.
8. Open cooling water valves to turbine when provided. ________ ________
9. Open the sealing steam lines, when provided, on both the low and high pressure ________ ________
ends.
10. Manually begin turbine rotation up to idle speed by further cracking open the main ________ ________
steam supply valve.
11. Check bearings for oiling ring rotation and listen for rubbing sounds. ________ ________
12. Hand trip the overspeed mechanism to check for proper action. Reset to continue ________ ________
startup.
13. Open the steam block valve completely to allow the turbine to accelerate to governor ________ ________
control. Check that it is controlling to the desired speed.
Startup Checklist For Centrifugal Pumps

1. When required for process safety, drain the pump casing. (One example would be a ________ ________
hot pump which has recently been steamed out.)
2. Verify any external mechanical seal flush supply line is open and flowing. Check for ________ ________
the prescribed flush pressure or flow. Refer to “shaft mechanical seals” checklist, this
section, for checks on other sealflush plans.
3. When supplied, check that cooling water is flowing to bearings, stuffingbox, seal ________ ________
coolers, gearbox coolers (on Sundynes), and pump supports.
4. Ensure that steam is open to pump jackets and piping tracing. Verify that all tracing ________ ________
steam traps are functioning correctly.
5. When provided, open pump warm-up lines to allow pump to reach prescribed starting ________ ________
temperature. Note that some pumps are warmed through a small hole in the discharge
check or block valves. (Warm-up any pump with operating temperatures above
350°F.)
Initials Date
Startup Checklist For Centrifugal Pumps (continued)

6. Open the pump suction line and vent casing to atmosphere, closed drain system or ________ ________
relief, as appropriate.
7. Verify that the pump minimum flow bypass is open and that any associated control ________ ________
valves and instrumentation are functioning. (This is particularly important on high pres-
sure, high horsepower, parallel operation, or high-speed pumps.) When a minimum
flow bypass is not provided, crack open the discharge block valve prior to starting.
8. Vent pump case. ________ ________
9. Start pump with discharge valve cracked, then fully open the discharge valve once the ________ ________
pump is at speed.
10. Do not operate pump more than 15 seconds without discharge pressure. Do not ________ ________
operate below minimum stable flow. Avoid parallel operation.
11. Check for and correct any of the following problems: ________ ________
– High vibration. ________ ________
– Normal discharge pressure and flow. ________ ________
– Cavitation or “pumping marbles” type sounds. (If heard, immediately look for ________ ________
suction or discharge blockages, excessive flow, low flow, excessively hot fluid, or
low suction vessel level.)
– Mechanical seal leaks. Excessive package leakage. ________ ________
– Excessive power required. ________ ________
– Hot bearings. ________ ________
– Grinding or growling type noises typical of metal-to-metal contact. ________ ________
– Lack of oiling or slinger rotation in bearing housings. ________ ________
– Improper lubrication system operation. ________ ________
– Malfunctioning instrumentation. ________ ________
12. High temperature or turbine driven pumps may require hot alignment. If so, pump ________ ________
should be shut down after a few hours operation to check the alignment and make
adjustments.
13. The performance of new, modified, or significantly overhauled pumps should be ________ ________
checked once the process is stabilized. This will require accurate pressure gages and
flow meters as well as knowledge of the fluid’s specific gravity. Motor ammeter read-
ings are also needed. Refer to Section 1140 of the Pump Manual for test procedures
and calculations.
Startup Checklist For Vertical Centrifugal Sump Pumps

1. Verify that the sump level float controls are properly set and functional. ________ ________
2. When supplied, check that the bearing flush supply is open and flowing. (Note that ________ ________
many flush systems are opened with some type of solenoid device when the pump
motor is energized. If so, check for proper operation of the flush controls.)
3. When shaft bearings are grease lubricated, inject grease on a prescribed basis ________ ________
through provided fittings.
Initials Date
Startup Checklist For Vertical Centrifugal Sump Pumps (continued)
4. Start pump and check for high vibration. ________ ________
5. If pump cycles on and off frequently, consider repositioning float switches to optimum ________ ________
levels.
6. Evaluate pump performance by observing discharge pressure and time to pump out ________ ________
the sump. Compare to Vendor’s curve.
Startup Checklist For Steam-Driven Reciprocating Pumps

1. Check that all linkages have been recently lubricated. ________ ________
2. Verify that the lubricator reservoir is full of the specified oil. Refer to “Shaft Packing” ________ ________
checklist in this section of the manual for other packing items to be checked.
3. Open drains on the steam casing and piping to purge condensate. ________ ________
4. Crack open steam inlet valve to warm casing, leaving the drains slightly open. ________ ________
5. Open suction and discharge valve on pumped fluid side. ________ ________
6. Slowly open the steam inlet valve to desired pumping rate. If pump is on external ________ ________
control, open block valve completely.
7. If pump stalls or relief valves open, check for blockage in the discharge line. ________ ________
8. If the pump chatters, check for restrictions or loss of fluid in the suction system. ________ ________
9. If piping vibration is excessive, adjust speed slightly to avoid resonance. Brace piping ________ ________
as needed.
10. Verify that the lubricator is functioning properly. ________ ________
11. Readjust packing to obtain a slow drip. ________ ________
Startup Checklist For Motor-driven Reciprocating Pumps
1. Verify that the discharge-to-suction relief valve is unblocked. Check tags that the valve ________ ________
has been tested and set per local policy. Ensure any block valves are locked open.
2. Check that the crankcase oil level is full. ________ ________
3. Ensure that the discharge-to-suction startup bypass line is open. ________ ________
4. Verify that suction/discharge pulsation bottles are fully operational (N2 filled, steam ________ ________
dome on, etc.).
5. When packing lantern rings are provided, ensure the flushlines are open and pressur- ________ ________
ized.
6. When packing lubricators are provided, check that the supply reservoir is filled with the ________ ________
prescribed oil.
7. If supplied with auxiliary lube-oil pump, verify its proper operation and control settings. ________ ________
8. If variable speed control is provided, slowly bring the pump up to full speed. ________ ________
9. If belt driven, check for proper belt tension and sheave alignment. ________ ________
Initials Date
Startup Checklist For Motor-driven Reciprocating Pumps (continued)
10. Check for and correct any of the following problems: ________ ________
– High vibration of the pump. ________ ________
– Loud chattering or pounding noises indicative of insufficient suction pressure. (This ________ ________
can be very destructive. Stop the pump immediately.)
– Low lube-oil supply pressure or high differential pressure at the filter. ________ ________
– Improper conditions of the motor and the speed changers (gearbox, belt, hydraulic ________ ________
drive, etc.).
– Excessive packing leaks. ________ ________
11. Check for excessive piping vibration. If any, verify that any gas-filled dampers are ________ ________
properly charged. Brace piping as needed.
12. Readjust packing to obtain a slow drip. ________ ________
Startup Checklist For Proportioning Pumps

1. Verify that the crankcase is full of the proper oil. For pumps with the hydraulic section ________ ________
separate from the crankcase, verify the hydraulic section is also full of the proper oil.
2. Bleed air from hydraulic section between double diaphragms when pump is new or ________ ________
has been repaired.
3. Set the flow adjustment to zero. ________ ________
4. Open the suction valve and bleed the line where possible. ________ ________
5. Check that the discharge-to-suction relief valve is not blocked and has been tested ________ ________
and set per local policy. For pumps with internal relief in the crankcase or hydraulic
section, verify the relief has been set correctly.
6. Open the discharge valve. ________ ________
7. Start pump motor and check for unusual vibration or noises. ________ ________
8. Slowly increase flow setting to the desired level. If pump does not immediately take ________ ________
suction, allow it running time to purge any residual gas in the system. If it continues
not to pump, shut down and check for valve problems or problems in the inlet system.
9. If pump has packing, adjust as required to obtain a slow drip. ________ ________
10. Set pump to automatic flow control, where provided. ________ ________
1120 Introduction to Troubleshooting

General Comments
The following comments may be helpful background to the detailed checklists in
this section, plus those in Section 800:
1. Although machinery (pump) problems can be exceptionally complex, practical
solutions are usually quite simple. Look for the simple cause/solution first. For
example, there are many reasons why a bearing will fail prematurely; some
reasons are quite complicated. Nevertheless, the most likely cause, and the first
you should check, is inadequate lubrication.
2. No machine operates perfectly, or in a perfect environment. Numerous devia-
tions exist in every machine, yet do not surface as operational or maintenance
problems. For example, every machine operates with some imbalance, some
misalignment, some imperfections in construction, etc.
Therefore, when called on to solve a problem that has surfaced, expect to find
several “problems” during the investigation. The job is not to find a deviation
(problem), but to find the deviation, or combination of deviations which are
causing the problem that needs to be corrected. A common error is to stop at
the first deviation from correct operation, assuming it to be the cause. This
leads to wasted time and further efforts in the future.
3. As in any problem-solving effort, one of the most important steps is to define
the problem. For example, problems often surface as “pump won’t put out -
repair as necessary.” The pump is overhauled at considerable expense even
though the real problem may simply be a plugged line restricting flow.
4. Many problems with pumps fall into one of the following categories:
a. Broken or worn components in the pump.
b. Broken or worn components in the driver.
c. Control mechanisms out of adjustment.
d. Mechanical components out of adjustment.
e. Problems external to the pump (for example, insufficient flow due to
plugged piping, faulty flow meter, excessive pressure loss, etc.).
Defining the problem in terms of the above categories can often reveal that
there is no problem with the pump at all, or the corrective action will be less
expensive than expected. It’s much easier to adjust a turbine governor than to
overhaul a pump. When investigating, look to the less expensive, simpler
causes, first.
5. As implied above, the causes of many pumping problems are often related to
factors outside the pump. This is especially true for the following “problems.”
a. Insufficient flow rate or discharge pressure.
b. Insufficient power or driver “kicking out.”
Problem Solving in General

The following steps are a common problem-solving technique:
1. Define the desired performance.
2. Define the deviation from that performance. This is the real problem.
3. Analyze the cause of the deviation, based on a combination of practical and
technical knowledge.
4. Take corrective action to eliminate the cause.
5. Monitor performance following corrective action.
6. Document the important points and communicate to those who will benefit
from the knowledge gained. Retain in files.
Although the checklists that follow are generally aimed at helping in Step 3, it is
important to remember all the steps. Repetitive problems are usually caused by
failing to complete one of the steps above.
Problem Solving Checklists

The figures and charts in the following sections help identify the cause of common
pump problems:
1130 Troubleshooting Flow Chart for Insufficient Flow or Pressure from
Centrifugal Pumps.
1140 Procedure for Monitoring the Performance of Centrifugal Pumps
(“Curving” the Pump).
Centrifugal Pump Troubleshooting Checklist.
1160 Vertical Turbine Pump Troubleshooting Checklist.
1170 Metering Pump Troubleshooting Checklist.
1180 Reciprocating Pump Troubleshooting Checklist.
1190 Rotary Pump Troubleshooting Checklist.
1130 Troubleshooting Insufficient Flow/Pressure from Centrifugal Pumps

Fig. 1100-1 Troubleshooting Flowchart for Centrifugal Pumps
1140 Procedure for Performance Monitoring Centrifugal Pumps

(“Curving the Pump”)
Background
This technique determines if pump internals such as wear rings and/or impellers are
worn. This information is useful to: (1) monitor the trend of a pump’s performance
over time, or (2) decide if the pump needs to be disassembled and repaired.
Appendix H of this manual is a guideline for testing centrifugal pumps in the
factory. In order to check a pump against its curve, the following fundamentals
should be understood.
1. Density is the weight per unit volume of any liquid. Water weighs 62.4 lbs. per
cubic foot at 60°F. Most petroleum stocks weigh less. A few liquids (e.g.,
caustic) weigh more.
2. Specific Gravity is a comparison of the weight of the liquid to that of water at
60°F. Petroleum has a specific gravity less than one, generally between 0.5
and 0.9.
Another way to express specific gravity is in “Degrees API.” (Water is 10
“Degrees API.”) Note that Degrees API increase while specific gravity
decreases. A conversion formula is shown in the attached “Pump Curve Work-
sheet,” or refer to the specific gravity chart in the Appendix.
3. Effect of Temperature. Calculations use specific gravity at the flow tempera-
ture. Therefore, specific gravity will have to be converted to the value at
pumping temperature. The specific gravity of almost all liquids decreases as the
temperature is raised. For example, a stock with a specific gravity of 0.9 at
60°F may be 0.8 at 340°F. The Appendices include information relating
specific gravities to temperature.
4. Gage and Absolute Pressure. The pressure shown on most pressure gages is
called “gage pressure,” expressed as psig. Gage pressure is the pressure above
atmospheric pressure. Absolute pressure is pressure measured above a perfect
vacuum. Pressure below 0 psig is a vacuum. Note that many pumps may have a
vacuum on the suction side. Atmospheric pressure is 14.7 psi (sea level) above
a perfect vacuum, or 14.7 psia.
To convert from psig to psia at sea level, add 14.7 (At other elevations, add the
local atmospheric pressure.)
psia = 14.7 + psig
Figure 1100-2 may help you keep these concepts in mind (assuming standard
atmospheric conditions at sea level):
Fig. 1100-2 Absolute and Gage Pressure
Absolute Pressure Gage Pressure
24.7 psia 10.0 psig
14.7 psia 0.0 psig
0.0 psia -14.7 psig
5. Centrifugal Pump Characteristics. For practical purposes, a centrifugal pump

will put up the same differential head for any liquid. Differential head is
defined as the pressure increase across the pump expressed as feet of liquid
(sometimes referred to as total head). Consider a pump that puts up 100 differ-
ential feet of water. It will also put up 100 feet of any other liquid.
One can think of this as a column of water 100 feet high at the pump’s
discharge (assuming the suction is zero). If we change the liquid to propane and
the suction stays the same, the discharge pressure will be a column of propane
100 feet high, the same as water.
Because water and propane have different specific gravities (propane is lighter),
the pressure at the bottom of each column will be different. The pressure at the
bottom of the water column will be greater than at the bottom of the propane
column. Therefore, the discharge pressure in psig will be different for water and
propane. Figure 1100-3 shows that a pressure gage would read 43 psig for
water, and 21.5 for propane.In an actual pump, the differential head can be
obtained by subtracting the suction pressure from the discharge pressure. (For
example, a discharge pressure of 500 psig and a suction pressure of 100 psig
will yield a differential pressure of 400 psi or, 924 feet of head if the fluid is
water with SG=1.0).
Fig. 1100-3 Centrifugal Pump Characteristics
6. Flow is volume of liquid going through a pipe per unit of time. It can be
expressed in gallons per minute, barrels per hour, or barrels per day. However,
almost every pump curve expresses it in gallons per minute.
Flow is usually measured on an instrument calibrated to read the flow at 60°F,
even when the actual stock temperature is higher. Thus, the recorded flow
must be corrected to actual flow for the pump curve.
Recorded Flow × Specific Gravity @ 60°F-

ActualFlow = ------------------------------------------------------------------------------------------------------
Specific Gravity @ Flow Temp.
7. Viscosity is a measure of “thickness” (a fluids resistance to shear), and is

usually measured in SSU or centistokes. The higher the number, the “thicker”
the stock and the more it resists movement. High viscosity stocks require more
pumping horsepower. The flow and head of a pump are less than they would be
at lower viscosities.
Most pump curves are based on the viscosity of cold water. However, the
vendor who supplies the pump usually draws a new curve that represents the
proper viscosity. If not, you may have to adjust the head-capacity curve down
and the horsepower curve up for significant changes in viscosity (see
Section 200).
Performance Monitoring Procedure

The following procedure can be used with the performance monitoring worksheet to
determine the condition of a centrifugal pump:
1. Run a field test. Copy the pump suction and discharge pressure. If possible, use
the same pressure gage to eliminate instrument error or use an accurate differ-
ential pressure gage. Suction pressure is very important to check but often diffi-
cult to obtain.
Ideally, the pump should be tested at 4 points:
a. Best efficiency point (BEP)
b. Some point about 10–25% above BEP
c. Some point about 10–25% below BEP
d. At or near shutoff. If possible, momentarily close the discharge pressure
valve, then record suction and discharge pressure at zero flow. This point is
called “shutoff” head. Do not keep the discharge valve closed more than a
few seconds or you may damage a pump component, especially a mechan-
ical seal.
2. Record the discharge pressure, suction pressure and flow rate at each test point.
Also record the stock temperature if available. Suction pressure is often diffi-
cult to obtain. Do your best to obtain or estimate an accurate number.
3. Use the worksheet to convert to the units on the pump curve (usually pressure
to feet of head and flow to GPM). Record the other information on the work-
sheet.
4. Plot the worksheet data on the pump’s individual performance curve. If it is
within 10% of the curve, the pump is healthy. If the pump is operating more
than 10% below the curve and the conditions from the flow chart in Section
1130 are satisfied, the pump is probably worn internally and needs an over-
haul. Any uncertainty about the accuracy of flow rate should be noted in the
horizontal direction of the chart, and uncertainties about the accuracy of
specific gravity and pressure measurements should be noted in the vertical
direction (S.G. is proportional to pressure).
Notes
1. It is a good idea to run the “shutoff” head test in addition to the other tests.
Pumps with a plugged suction line or plugged inlet will usually put up the
design shutoff head, but will plot below the curve at increased flow. Doing both
tests could detect a plugged inlet or suction line and avoid a needless repair.
2. If possible, record the amps on motor drivers. Determine if the amps are greater
than normal. This could indicate an unmetered flow, a faulty flow meter, or
internal recirculation.
3. If no suction pressure gage exists, the suction pressure will have to be calcu-
lated. Add the pressure in the suction vessel to the static head (the vertical
distance between the liquid level in the suction vessel and the centerline of the
pump suction flange) and subtract the friction loss in the suction piping.
Suction lines are usually designed with low friction loss.
4. Each pump is designed for one specific application. It has its own individual
performance curve. The most difficult job throughout this entire process may be
to find the correct performance curve. Many pumps are old and their curves
have been lost. To locate the right curve, refer to the Engineering files, or
perhaps the vendor catalog. You may have to request the correct performance
curve from the vendor. Make sure the curve is corrected for viscosity. The time
spent depends, of course, on potential costs involved.
5. Once the correct performance curve is found, make sure the speed and impeller
diameter of the pump are the same as shown on the pump’s performance curve.
If not, the curve will have to be adjusted. This can be done from the old curve
once the new speed or new impeller diameter is known. Use the “Affinity” laws
(see Section 200).
Sources of Error
1. An underperforming driver will cause the pump to underperform. While elec-
tric motors normally run at-speed, or not at all, it is not uncommon for a turbine
driver to run at underspeed because of a mechanical problem (in the turbine or
the pump) or because of steam supply problems. A portable tachometer can be
used to quickly check driver speed. You should also note motor rotation, a
pump running backward may put up as much as 60% of normal head.
2. An erroneous test or test analysis will produce false conclusions about perfor-
mance. Some common errors are:
a. Mis-estimating suction pressure when no gage is available.
b. Failing to correct the measured flow rate to flow rate at flowing tempera-
ture.
c. Using the wrong specific gravity. (If the tester assumed a higher specific
gravity than actual, he would conclude there was a significant perfor-
mance problem.)
d. A different size impeller from the original one can cause profound varia-
tion from the original curve because pump head capacity varies as the
square of the impeller size. (Check local records to verify the running
impeller size).
e. A highly viscous fluid will impair pump performance and increase power
required. Viscosities of some hydrocarbons change substantially over wide
temperature swings. This is particularly important if viscosity is above
100 centistokes (water @ 70°F has a viscosity of 1 centistoke).
f. Incorrect flow, pressure, and/or temperature readings will result in an inac-
curate curve. Flow meters may not be correctly calibrated, especially if
recent process changes have been made. PI’s and TI’s should be replaced if
there is any doubt about their accuracy.
Example
Is the following pump (Figure 1100-4) operating on its curve?
Fig. 1100-4 Pump Curve Diagram
The pump curve in Figure 1100-5 shows one operating point obtained from a field
test and calculated on the example worksheet (Figure 1100-6). Calculations show
the pump at 181 GPM flow and 1204 ft differential head. Plotting this on the curve
shows the pump operating on its performance curve. The pump is “healthy.”
(Figure 1100-7 illustrates a blank Pump Curve Worksheet.)
Fig. 1100-5 Field Test Pump Curve
Fig. 1100-6 Pump Curve Worksheet (Sample)
Fig. 1100-7 Pump Curve Worksheet
1150 Centrifugal Pump Troubleshooting Checklist

From Pump Handbook, (1976) Edited by Karassik, Krutzch, Fraser & Messina.
Used with permission from McGraw Hill.
Symptoms Possible Causes

(Each number is defined in the list below)
Insufficient capacity delivered 1-2-3-4-5-6-7-8-9-10-11-13-14-17-19-20-21-23-24-25-26-32-33-34
Insufficient pressure developed 5-11-13-14-17-19-21-22-24-32-34
Pump requires excessive power 12-13-14-15-16-17-20-21-24-27-30-31-32-33-34-35-36-39
Pump overheats and/or seizes 2-4-18-19-30-36-37-38-40
Pump vibrates, cavitates, or is noisy 1-2-3-4-5-9-10-18-20-21-27-28-29-30-31-34-37-38-40-41-42-43-
44-45-46
Pump loses prime 3-5-7-8
Excessive package leakage 27-29-35-36-37-38-39-53
Packing life short 27-29-31-35-36-37-38-39-53
Bearings overheat or wear rapidly 27-40-42-43-44-45-46-47-48-49-50-51-52-54
Mechanical seal problems Refer to Section 800
Possible Causes
Suction Causes System Causes
1. Suction manifolded improperly 11. Speed too low

2. Pump or suction pipe not completely filled with liquid 12. Speed too high
3. Suction lift too high 13. Wrong direction of rotation
4. Insufficient margin between suction pressure and 14. Total head of system higher than pump design head
vapor pressure 15. Total head of system lower than pump design head
5. Excess air or gas in liquid 16. Specific gravity of liquid different from design
6. Air pocket in suction line 17. Viscosity of liquid differs from design
7. Air leaks into suction line 18. Operation at very low capacity
8. Air leaks into pump through stuffing boxes 19. Parallel operation of pumps unsuitable for such
9. Inlet of suction pipe insufficiently submerged operation
10. Suction screens/piping plugged 20. Spare pump’s check valve stuck open
21. Malfunctioning minimum flow by-pass
22. Lower-than-design specific gravity
23. Plugged discharge piping/valves
24. Malfunctioning pressure or flow indication
25. Fouled heat exchanger downstream
Mechanical Causes
26. Foreign matter in impeller 42. Lack of lubrication

27. Misalignment 43. Improper installation of antifriction bearings (damage
28. Foundations not rigid during assembly, incorrect assembly of stacked bear-
ings, use of unmatched bearings as a pair, etc.)
29. Shaft bent
44. Dirt in bearings
30. Rotating part rubbing on stationary part
45. Rusting of bearings due to water in housing (bearing
31. Bearings worn
seals or improper protection while idle for long
32. Wear rings worn periods)
33. Impeller damaged 46. Excessive cooling of water-cooled bearing resulting in
34. Defective casing gasket permitting interstate leakage moisture condensation in the bearing housing
35. Packing improperly installed 47. Incorrect oil level (too high/too low)
36. Incorrect packing for operating conditions 48. Insufficient bearing cooling
37. Shaft running off center because of worn bearing or 49. Bearings too tight, or excess preload
misalignment 50. Oil ring not functioning
38. Rotor out of balance results in vibration 51. Oil mist problems
39. Gland too tight resulting in no flow of liquid to lubri- 52. Improper lubricant
cate packing
53. Stuffing box, neck ring, shaft, or packing sleeve worn
40. Excessive thrust caused by mechanical failure inside
54. Insufficient oil flow (insufficient pressure, wrong orifice
the pump or by failure of the hydraulic balancing
size, etc.)
device, if any
41. Excessive grease or oil in antifriction bearings
housing or lack of cooling, causing excessive bearing
temperature
1160 Vertical Turbine Pump Troubleshooting Checklist

From Practical Machinery Management for Process Plants, Vol. 2, by Heinz P.
Bloch and F.K. Geitner. Copyright 1994 by Gulf Publishing Co., Houston, TX. Used
with permission. All rights reserved.

Insufficient Capacity 1-2-8-10-11-13-17-22-23-26
Insufficient Discharge Pressure 1-7-9-11-13-15-22-23-26
Vibration 4-5-6-12-13-14-16-18-19-20-24-25-29-30-31-32
Abnormal Noise 3-6-12-16-19-21-28-30
Power Demand Excessive 11-12-14-16-24-25-27
Mechanical Seal Problems See Section 800
Possible Causes and Solutions

Cause Solution
1. Pump Suction Interrupted (Water Level Below Inlet) Check Sump Level
2. Low Water Level Check Water Level
3. Cavitation Due to Low Submergence Check Submergence
4. Vortex Problem Install Vortex Breaker Shroud
5. Suction or Discharge Recirculation Establish Design Flow
6. Operation Beyond Maximum Capacity Rating Establish Proper Flow Rate
7. Entrained Air Install Separation Chamber
8. Strainer Clogged Inspect and Clean
9. Impeller Plugged Pull Pump and Clean
10. Impeller or Bowl Partially Plugged Pull Pump and Clean
11. Impellers Trimmed Incorrectly Check for Proper Impeller Size
12. Improper Impeller Adjustment Check Installation/Repair Records
13. Impeller Loose Pull Pump and Analyze
14. Impeller Rubbing on Bowl Case Check Lift
15. Wear Rings Worn Inspect During Overhaul
16. Shaft Bent Pull Pump and Analyze
17. Shaft Broken or Unscrewed Pull Pump and Analyze
18. Enclosing Tube Broken Pull Pump and Analyze
19. Bearings Running Dry Provide Lubrication
20. Worn Bearings Pull Pump and Repair
21. Column Bearing Restrainers Broken Pull Pump and Analyze
Cause Solution
22. Wrong Rotation Check Rotation
23. Speed Too Slow Check RPM
24. Speed too High Check RPM
25. Misalignment of Pump Assembly Inspect for Excessive Pipe Strain
26. Leaking Joints Inspect
27. Pumping Sand, Silt, or Foreign Material Check Liquid Pumped
28. Motor Noise Check Sound Level
29. Motor Electrical Imbalance Perform Phase Check
30. Motor Bearing Problems Consult Driver Manual
31. Motor Drive Coupling Out of Balance Inspect
32. Resonance: System Natural Frequency Near Perform Vibration Analysis and Modify and
Pump Speed Brace as Needed
1170 Metering Pump Troubleshooting Checklist

From The Metering Pump Handbook by McCabe, Lanckton and Dwyer. Industrial
Press, Inc., Copyright 1984. Reprinted with permission.

Pump does not start 1-2-3-4-7-8-13-14-33-34-54
Pump starts, no delivery 9-11-12-13-14-35-36-37-60
Pump does not deliver to capacity 4-5-7-8-15-16-18-19-22-25-26-27-34-38-40-41-42-43-55-56-
58-60
Capacity falls off 4-15-18-28-29-30-32-34-40-41-43-53-55-56-57-60
Erratic delivery 4-7-16-27-29-31-32-42-53-56-58-60
Loss of prime 9-10-11-12-13-14-15-18-19-25-26-34-35-36-40-41-43-57-60
Noisy liquid end 21-23-24-25-26-27-32-37-55-56-58-60
Noisy mechanical end 17-19-24-33-39-40-44-46-51-55
Noisy piping 10-23-24-25-26-32
Rapid packing, plunger, or piston wear 14-18-19-20-24-28-37-40-41-43-53-57-58-59
Pitted seals and valves 27-29-32-53
Overheated motor 3-6-8-24-31-39-45-54
Pump mechanism running hot 28-31-38-39-45-46-47-50-51-52-54-56-57-58-59
Pump leaking oil 24-46-47-48-50-52
Contamination in crankcase power source 20-38-46-50-58-59

Cause Solution
Power Source
1. Circuit Breaker Open or Fuse Blown Locate Reason for Overload Before Replacing or
Resetting
2. Broken or Disconnected Wire Repair or Replace
3. Improper Wiring Check Wiring Diagrams
4. Incorrect Signal Check Signal Source
5. Motor Speed Too Slow Check Voltage, Frequency, Wiring, and Specifications
6. Motor is Overloaded Check Job Conditions
7. Insufficient Air Supply Repair Compressor; Clean Filters
8. Low Voltage Check and Correct
9. Supply Vessel Empty Fill With Process Liquid
10. Suction Piping Too Long or Too Small Shorten, Increase Size, or Install Accumulator Near
Suction Point
11. Strainer Clogged Clean or Replace

Cause Solution
Supply and Discharge System
12. Suction Pipe Not Submerged Fill Tank or Realign
13. Valves Closed Open Full; Check All Valves in Suction and Discharge
Piping
14. Clogged Piping Clean; Check Product Conditions
15. Insufficient Suction Pressure ‘Raise Supply Tank; Increase Pipe Diameter; Pressurize
Tank
16. Insufficient Discharge Pressure Install a Loop or Back-Pressure Valve
17. Too Much Suction Pressure Lower Supply Tank; Install Back-Pressure Valve
18. Leak in Suction Piping Locate and Repair, If Foot Valve Required, Test for Seal
19. Pump Not Level or Rigid Realign
20. Dirty Environment Clean Immediate Area of Pump; Change Lubricant
21. Extreme Pipe Vibration Shorten Discharge Line; Increase Size; Install Dampener
22. Calibration Incorrect Evaluate Method of Measurement and Measuring Instru-
ments for Possible Error
23. Discharge Line Too Long or Too Small Shorten; Replace; or Install Dampener
24. Discharge Pressure Too High Install Dampener; Remove Restriction; Increase Size of
Piping
25. Surge Chambers or Dampeners Full of liquid Recharge With Gas or Air Liquid
26. Surge Chambers or Dampeners Improperly Charged Check Specifications and Recharge Properly
27. NPSHA Not Sufficient Check and Correct
28. Pump Vapor bound Allow Pump to Operate at Low Pressure Through Bypass
to Eliminate Vapor, or Put in Vent Before Suction
29. Change in Liquid Properties Compare Original Job Conditions to Present Conditions
30. Viscosity Changed or Too High Improve Suction Head; Lower Viscosity; Change
Metering Head and Valve Design
31. Change in Ambient Temperature Relocate Pump; Control Ambient Temperature
32. Process Liquid Vaporizing Check NPSHA; Change Head, Location, Piping, Temper-
ature as Necessary
Pump Mechanism – Mechanical

33. Coupling Loose or Disconnected Tighten or Replace
34. Check Valve Dirty or Worn Clean or Replace
35. Check Valves Lodged Open by Solids Clean; Inspect; Add Strainer or Slurry Valves
36. Check Valves Stuck Open by Corrosion Clean Guides; Inspect for Wear
37. Prime Lost Check and Reprime
38. Seals Leaking Inspect Seals and Seal Surfaces; Replace Defective
Seals
39. Pressure-Relief Valve Set Too High Adjust to Required Pressure Within Pump Design Limits
40. Plunger Loose or Scored Tighten; Repack; If Cylinder Worn, Also Change
41. Packing Worn or Hardened Adjust or Change

Cause Solution
42. Pressure-Relief Valve Set Too Low Adjust to Required Pressure Within Pump Design Limits
43. Piston Seals Worn or Hardened Change
44. Loose Bearings Adjust or Replace
45. Tight Bearings Adjust or Replace
46. Worn Cams, Rods, or Gears Inspect; Check for Overload; Inspect Bearings
47. Pump Overfilled with Lubricant Check and Adjust Level
48. Defective Seals Replace
49. Plugged Breather or Vent Inspect and Clean
50. Low Lubricant Level Inspect for Leaks and Refill
51. Lubricant Contaminated Inspect for Cause and Change
52. Erosion/Corrosion Check for Dirty Product; Incompatible Materials; Replace
53. Mechanism Frozen Up Check for Corrosion on Moving Parts
Pump Mechanism – Hydraulic

54. Hydraulic System Underprimed Check for Cause and Reprime
55. Hydraulic Relief and Make-up Valves Functioning Check for Restricted Supply or Discharge, Broken
Diaphragm, Worn Packing, Defective Valves. Do Not
Tighten Beyond Limits of Pump
56. Hydraulic Fluid Level Low Fill to Correct Level
57. Hydraulic Fluid Breaking Down Check Temperature, Pump Load, Restricted Supply
58. Hydraulic Fluid Contaminated Check Filters, Breathers, Seals, Diaphragm
59. Hydraulic System Vapor Locked Check Bleeder, Make-Up Valve, Restricted Supply, and
NPSHA
1180 Reciprocating Pump Troubleshooting Checklist

Courtesy of Wheatley Gaso, Inc.

Failure to deliver rated capacity 1-2-3-4-5-6-7-8-9-10-11-12-13-14-15-16-17-18-19-20-21
Liquid not delivered 22-23-24-25-26-27-28
Cavitation 29-30-31-32-33
Leak at cylinder head or valve cover 34-35-36
Water in crankcase 37-38-39-40-41
Oil leak from crankcase 42-43-44-45
Excessive heat in power end 46-47-48-49-50-51-52-53-54-55-56-57-58
Pump overloads driver 59-60-61-62-63-64-65-66-67
Stuffing box leakage 68-69-70-71-72
Stud failure 73-74-75
Excessive valve noise 76-77-78-79
Inlet or discharge line vibration 80-81-82-83-84-85-86-87-88-89
Noisy operation or knocks 90-91-92-93-94-95-96-97-98-99-100-101-102-103-104-
105-106-107-108
Broken shafts, bent, stripped threads, and 109-110-111-112-113-114-115-116-117
catastrophic failures
Packing failure 118-119-120-121-122-123-124-125
Valve failure 126-127-128-129-130-131
Plunger failure 132-133-134-135-136-137-138-139
Causes
1. Air or vapor pocket in inlet line 11. Loose valve covers or cylinder heads
2. Capacity of charge pump less than capacity of power 12. Worn valves and seats
pump 13. Safety relief valve partially open, or not holding pres-
3. Air or vapor trapped in or above inlet manifold sure
4. Air leak in liquid supply piping system 14. Worn liners, piston rings or plungers
5. Loose bolts in pump inlet manifold 15. Bypass valve open, or not holding pressure
6. Air or gases entrained in liquid 16. Blown liner gasket
7. Foreign object holding pump inlet or discharge 17. NPSHA not sufficient
valve(s) open 18. Liquid bypassing internally
8. Incorrect drive ratio 19. Foreign object blocking liquid passage
9. Loose belts 20. Vortex in supply tank
10. Incorrect motor or engine speed
21. Insufficient power delivered by motor 51. Liquid end packing adjusted too tight (adjustable
22. Pump not primed packing only)
23. Air or vapor pocket in inlet line 52. Pump speed too high
24. Clogged inlet line 53. Low voltage or other electrical trouble
25. All inlet valves stuck open 54. Trouble with engine, turbine, gear or other related
equipment
26. All discharge valves stuck open
55. Excessive discharge line pressure
27. Loose bolts in pump inlet manifold
56. Clogged discharge line
28. Valve velocities too high
57. Closed or throttled valve in discharge line
29. NPSHA too low
58. Incorrect liner size for application
30. Liquid not delivered to pump inlet connection
59. Improper bypass conditions
31. Excessive stuffing box leakage
60. Overtightened stuffing box glands on adjustable
32. NPSHR too high
packing
33. Acceleration head too high
61. Worn packing
34. Operating over recommended pressure
62. Worn rods or plunger
35. Loose cylinder head, valve cover
63. Worn stuffing boxes
36. Damaged gasket.
64. Wrong size or type packing
37. Water condensation
65. Worn O-ring seal (replaceable boxes)
38. Worn seals
66. Excessive discharge pressure
39. Clogged air breather(s)
67. Improper torquing of nuts
40. Worn crankcase packing
68. Shock overload caused by pump pulsations
41. Loose covers
69. Broken or weak valve spring
42. Oil level too high
70. Pump cavitation
43. Work seals
71. Air leak in inlet piping or loose bolts in pump inlet
44. Worn crankcase packing manifold
45. Loose crankcase cover 72. Air trapped above inlet valve
46. Pump operating backward at too low a speed 73. Piping inadequately supported
47. Insufficient oil in power end 74. Inlet line too long or too small in diameter
48. Excessive oil in power end 75. Too many bends in inlet line
49. Incorrect oil viscosity 76. Multiple pump installations operating in parallel
50. Operating in excess of recommended pressure 77. Obstruction under valve(s)
78. Main bearings too tight 117.Improper material, wrong packing

79. Drive misaligned 118.Improper lubrication
80. Belts too tight 119.Adjustable packing - gland too tight
81. Discharge valve of one or more cylinders stuck open 120.Dirty liquid
82. Insufficient cooling 121.Plunger or piston rod misalignment
83. Pump speed too low 122.Dirty environment
84. Inadequate ventilation 123.Damaged rod or rod coating
85. Valve noise amplified 124.Normal wear
86. Pump cavitation 125.Pump cavitation
87. Liquid knock 126.Abrasives of foreign matter in fluid
88. Air leak in inlet piping 127.Incompatibility of valve components to corrosive
89. Loose bolts in pump inlet manifold liquid
90. Hydraulic noise in liquid end 128.Galvanic corrosion
91. Loose or worn crosshead pins and bushings 129.Incorrect installation – driving on the valve stem,
improper torque on jam nut, valve seat and valve
92. Loose connecting rod cap bolt(s)
deck not thoroughly clean and dry when seat
93. Worn connecting rod bearings installed
94. Worn crosshead 130.Startup against closed valve in discharge line. If
95. Main bearing end play excessive valve seats are discovered driven too deeply after
96. Worn gears or chains operation of the pump, look for the following pattern
of driven seats, indicative of startup or run against a
97. Gears or chains out of line closed discharge line valve:
98. Pump running backward
Triple single acting plunger pump:
99. Partial loss of prime 2 inlet and 1 discharge valve seat, or
100.Pulsations in piping system 1 inlet and 2 discharge valve seats
101.Water in power end crankcase
102.Poorly supported piping, abrupt turns in piping, piping Quintuplex single acting plunger pump:
misaligned, pipe size too small 3 inlet and 2 discharge valve seats, or
103.Packing worn 2 inlet and 3 discharge valve seats
104.Operating in excess of recommended pressure or
speed Duplex double acting piston pump:
2 inlet and 2 discharge valve seats
105.Low NPSHA
106.Surge chambers or dampers need recharging
131.Low oil level
107.Surge chambers or dampers missing
132.Packing too tight
108.Piston or plunger loose
133.Thermal shock (cold water hitting hot ceramic
109.Contaminated oil plunger)
110.Main bearing failure 134.Inlet valve becomes disassembled while pump is
111.Piston or plunger striking cylinder head operating
112.Disintegration of worn valves 135.Stuffing box gland rubbing on plunger due to
113.Frozen liquid in liquid end improper tightening procedure
114.Air leak in liquid supply system 136.Dirty liquid
115.Loose bolts in pump inlet manifold 137.Dirty environment
116.Normal wear 138.Wrong packing/packing material
139.Inadequate flush to lantern ring
1190 Rotary Pump Troubleshooting Checklist

From Pump Handbook, (1976) Edited by Karassik, Krutzch, Fraser & Messina.
Used with permission from McGraw Hill.

No Liquid Delivered 1-2-4-15
Insufficient Capacity 1-3-5-6-9-13-1417
Starts, but Loses Prime 3-4-5-6
Excessive Wear 2-3-7-9-10-11-12-18
Excessive Heat 2-10-11-12
Vibration and Noise 1-2-3-5-6-9-10-11-12-16-19
Excessive Power Demand 8-9-10-11-12-19

Cause Solution
1. Suction Filter or Strainer Clogged Clean Strainer or Filter
2. Pump Running Dry Reprime
3. Insufficient Liquid Supply Look for Suction Restriction or Low Suction Level
4. Suction Piping Not Immersed in Liquid Lengthen Suction Pipe or Raise Liquid Level
5. Liquid Vaporizing in Suction Line Check NPSH.
Check for Restriction in Suction Line
6. Air Leakage Into Suction Piping or Shaft Seal Tighten and Seal all Joints;
Adjust Packing or Repair Mechanical Seal
7. Solids or Dirt in the Liquid Clean System
Install Filtration
8. Liquid More Viscous than Designed For Reduce Pumped Medium Viscosity;
Reduce Pump Speed;
Increase Drive HP
9. Excessive Discharge Pressure Check Relief Valve or By-pass Setting;
Check for Obstruction in Discharge Line
10. Pipe Strain on Pump Casing Disconnect Piping and Check Flange Alignment
11. Coupling, Belt Drive, Chain Drive Out of Alignment Realign
12. Rotating Elements Binding Disassemble and Inspect
13. Internal Parts Wear Inspect and Relace Worn Parts
14. Speed Too Low Check Driver Speed
15. Wrong Direction of Rotation Check and Reverse if Required
16. Coupling Out of Balance Balance
17. Relief Valve Improperly Seated Check and Repair
18. Packing too Tight Readjust or Replace
19. Bent Drive Shaft Repair or Replace
Other References
References – All Pumps

1. Cameron Hydraulics Data. Edited by C. R. Westaway, A. W. Loomis.
(Ingersoll-Rand, 1981).
A compact reference which includes discussion on hydraulics, fluid-flow data,
pipe friction, specific gravity, viscosity, conversions. Not specific to
petrochemicals.
2. Fundamentals and Applications of Centrifugal Pumps for the Practicing
Engineer. Alfred Benaroya. (Petroleum Publishing Company, 1978).
3. Hydraulic Institute Standards. (Hydraulic Institute, 1975).
An authoritative reference covering definitions, testing, applications, hydraulics,
materials. Background information with numerous graphics.
4. Pump Application Engineering. Tyler G. Hicks and Theodore W. Edwards.
(McGraw-Hill Book Company, 1971).
A basic reference to a broad range of subjects: Covers all pump types (excluding
downhole types), hydraulics, drivers, controls, applications, suction design, installa-
tion, operation, and maintenance.
5. Pump Handbook. Edited by I. J. Karassik, W. C. Krutzch, W. H. Fraser,
J. P. Messina. (McGraw-Hill, 1976).
A basic reference to a broad range of subjects: Covers all pump types (excluding
downhole types), hydraulics, drivers, controls, applications, suction design installa-
tion, operation, and maintenance.
References – Centrifugal Pumps

1. Centrifugal Pump Clinic. I. J. Karassik. (Marcel Dekker, Inc., 1981).
A thorough discussion of centrifugal pump problems and how to avoid or
solve them.
2. Centrifugal Pump Handbook. (Sulzer Brothers Ltd., 1987).
A technically oriented handbook.
3. Centrifugal Pump Problems. S. Yedidia. (Petroleum Publishing Co., 1980).
A thorough, practical treatment of all aspects of centrifugal pump problems.
June 1988 © 1988 Chevron USA Inc. All rights reserved. Reference-1
Other References Pump Manual
4. Centrifugal Pumps - Design and Application. V. S. Lobanoff, R. R. Ross. (Gulf

Publishing, 1985).
A basic text covering hydraulics, centrifugal pump design and application. Covers
vertical and horizontal pumps. Straightforward, slightly technical.
References – Electrical Submersible Pumps

1. Electrical Submersible Pumps. (Chevron USA, 1984).
A practical manual issued by Central Region (now Northern) to accompany a course
for engineering and field personnel. Covers sizing, auxiliary equipment, installa-
tion, evaluation, maximizing run life, failure analysis. (Material in Section 600 is
reproduced from this manual.)
2. Handbook for Oilfield Subsurface Electrically-Driven Pumps. (Centrilift-
Hughes, Inc.).
3. API RP-11R. Recommended Practice for Electric Submersible Pump
Installations.
4. API RP-115. Recommended Practice for Operations, Maintenance, and
Troubleshooting Electrical Submersible Pump Installations.
References – Sucker-Rod Pumps

1. Artificial Lift Sucker-Rod Pumping. (Royal Enterprises, Inc., 1979).
Provides a detailed look at all aspects of sucker-rod pumping systems.
2. Rod Pumping Equipment: Surface and Downhole Analysis and Inbal Programs.
(Chevron Oil Field Research Co.).
Analysis of rod pumping systems performance based on monitoring conditions at
surface.
3. API RP-11L. Recommended Practice for Design Calculations for Sucker-Rod
Pumping Systems (Conventional Units).
4. API Bul-11L4. Bulletin containing Curves for Selecting Beam Pumping Units.
5. API RP-11AR. Recommended Practices for Care and Handling of Subsurface
Pumps.
6. API RP-11BR. Recommended Practice for Care and Handling of Sucker-Rods.
7. API SPEC-1B. Specification for Oil-Field V-Belting (Includes a design proce-
dure for power application of V-Belts).
8. API RP-11G. Recommended Practice for Installation and Lubrication of
Pumping Units.
9. API SPEC-11AX. Specification for Subsurface Sucker-Rod Pumps and
Fittings.
Reference-2 © 1988 Chevron USA Inc. All rights reserved. June 1988
Pump Manual Other References
10. API SPEC-11B. Specification for Sucker-Rods (Pony rods, polished rods,
couplings, and sub-couplings).
11. API STD-11E. Specification for Pumping Units.
References – Mechanical Seals

1. Engineered Fluid Sealing. (John Crane, Inc., 1979).
A straightforward manual covering practical aspects of sealing rotating shafts and
gasketed joints. A general, useful reference despite being a manufacturer’s publica-
tion.
2. Mechanical Seals. E. Mayer. (Newnes-Butterworths, 1977).
A highly technical, thorough treatment of fundamentals, design, leakage, wear, and
special designs.
3. Mechanical Seals. E. M. Scheel. (Chevron USA, 1986).
A straightforward, practical manual covering basics of mechanical seals, design,
application, materials, and converting packed pumps to mechanical seals (Much of
Section 700 is reproduced from this manual.). Available from Chevron USA,
742 Bancroft Way, Berkeley, CA, 94710.
References – Metering Pumps

1. Metering Pumps Handbook. R. E. McCabe, P. G. Lanckton, W. V. Dwyer.
(Industrial Press, Inc., 1984).
A basic reference covering all aspects. (Authors affiliated with Pulsafeeder, Inc.).
2. Metering Pumps - Selection and Application. J. P. Poynton. (Marcel Dekker,
Inc., 1983).
Also a basic reference covering all aspects of metering pumps. (Author affiliated
with Milton Roy Corp.).
References – Miscellaneous
1. Grouting Handbook. (U. S. Grant Corporation, 1983).
A useful book for understanding grouting basics (Oriented to the publisher’s prod-
ucts.).
2. Practical Machinery Management for Process Plants. H. P. Block,
F. K. Geitner. (Gulf Publishing, 1982).
A thorough, four-volume set covering a broad range: design, repair, reliability
factors, troubleshooting, and failure analysis of all machinery, including pumps.
June 1988 © 1988 Chevron USA Inc. All rights reserved. Reference-3
Other References Pump Manual
3. IMI Candidate Manual. Integrated Machinery Inspection Staff. (Chevron USA,

1981).
A straightforward, practical manual outlining the IMI Program. Covers all aspects
of an overall program of machinery reliability in process plants. This program has
been implemented throughout CUSA Manufacturing. Applicable to all pumps other
than downhole, but concepts apply to all pumps.
4. Total Alignment. V. R. Dodd. (Petroleum Publishing Co., 1975).
A thorough treatment of alignment fundamentals, ambient and service condition
alignment, and computer alignment.
5. Machinery Reliability Procedures. (Chevron USA, 1986).
A collection of detailed procedures for machinery reliability and repair. Includes
graphics, checklists, and step-by-step procedures for all types of machinery.
Published by the IMI Group, El Segundo Refinery.
6. Engineering Data Book. (Gas Processors Association, 1987).
A two-volume set containing information on all process equipment. The section on
pumps is simple and concise, covering centrifugal pumps and hydraulics.
7. Fluid Flow Manual. (Chevron Corporation, 1980).
A practical manual specific to the petroleum industry. Covers fundamentals and
includes numerous charts for calculating hydraulics based on common fluids (steam,
water hydrocarbons).
References – Positive Displacement Pumps

1. Positive Displacement Pumps. (Howell Training Co., 1969).
A two-volume set covering reciprocating and rotary pumps. Although very basic,
may be helpful to entry-level personnel.
2. Reciprocating Pump Manual. J. Cooper, W. K. Lee. (Gasco Pump, Inc.).
A convenient reference covering fundamentals, design and troubleshooting. Manu-
facturer’s publication.
Reference-4 © 1988 Chevron USA Inc. All rights reserved. June 1988
Appendix A Conversion Tables and Formulas
Contents Page
Conversions, Constants, and Formulas A-2
Conversion Factors and Formulas (Courtesy of GSPA) A-3
Conversion Chart: Gallons per Minute to Barrels per Day A-4
Decimal and Millimeter Equivalents of Fractions A-4
Temperature Conversion Table A-5
June 1999 © 1999 Chevron USA Inc. All rights reserved. A-1
Appendix A Conversion Tables and Formulas Pump Manual
Fig. A-1 Conversions, Constants, and Formulas
A-2 © 1999 Chevron USA Inc. All rights reserved. June 1999
Pump Manual Appendix A Conversion Tables and Formulas
Fig. A-2 Conversion Factors and Formulas (Courtesy of GSPA)
Fig. A-3 Conversion Chart: Gallons per Minute to Barrels per Day
Fig. A-4 Decimal and Millimeter Equivalents of Fractions
Pump Manual Appendix A Conversion Tables and Formulas
Fig. A-5 Temperature Conversion Table
Appendix B Viscosity Data
Contents Page
Viscosity Conversion Table (Courtesy of Hydraulic Institute) B-2
Viscosity of Common Liquids (Courtesy of Hydraulic Institute) B-4
Friction Loss for Viscous Liquids (Courtesy of Hydraulic Institute) B-7
Viscosities of Hydrocarbon Liquids (Courtesy of GSPA) B-8
Approximate Viscosity–Temperature Relations for
Liquid Petroleum Fractions (Courtesy of Hydraulic Institute) B-9
Viscosity Unit Conversion Chart (Courtesy of Hydraulic Institute) B-10
June 1999 © 1999 Chevron USA Inc. All rights reserved. B-1
Appendix B Viscosity Data Pump Manual
Fig. B-1 Viscosity Conversion Table (1 of 2) (Courtesy of Hydraulic Institute)
B-2 © 1999 Chevron USA Inc. All rights reserved. June 1999
Pump Manual Appendix B Viscosity Data
Fig. B-1 Viscosity Conversion Table (2 of 2) (Courtesy of Hydraulic Institute)
Fig. B-2 Viscosity of Common Liquids (1 of 3) (Courtesy of Hydraulic Institute)
Fig. B-3 Friction Loss for Viscous Liquids (Courtesy of Hydraulic Institute)
Viscosities of Hydrocarbon Liquids (Courtesy of GSPA)
Fig. B-4
Fig. B-5 Approximate Viscosity–Temperature Relations for Liquid Petroleum Fractions

(Courtesy of Hydraulic Institute)
Fig. B-6 Viscosity Unit Conversion Chart (1 of 2) (Courtesy of Hydraulic Institute)
Fig. B-6 Viscosity Unit Conversion Chart (2 of 2) (Courtesy of Hydraulic Institute)
Appendix C Vapor Pressures
Contents Page
Vapor Pressure—Various Liquids (Courtesy of Goulds Pumps, Inc.) C-2
Low-Temperature Vapor Pressures for Light Hydrocarbons
(Courtesy of GPSA) C-4
June 1999 © 1999 Chevron USA Inc. All rights reserved. C-1
Appendix C Vapor Pressures Pump Manual
Fig. C-1 Vapor Pressure—Various Liquids (1 of 2) (Courtesy of Goulds Pumps, Inc.)
C-2 © 1999 Chevron USA Inc. All rights reserved. June 1999
Pump Manual Appendix C Vapor Pressures
Fig. C-1 Vapor Pressure—Various Liquids (2 of 2) (Courtesy of Goulds Pumps, Inc.)
Low-Temperature Vapor Pressures for Light Hydrocarbons (1 of 2) (Courtesy of GPSA)
Fig. C-2
Pump Manual Appendix C Vapor Pressures
Low-Temperature Vapor Pressures for Light Hydrocarbons (2 of 2) (Courtesy of GPSA)
Fig. C-2
Appendix D Properties of Water
Contents Page
Properties of Water at Various Temperatures From 32 to 705°F D-2
Atmospheric Pressure, Barometer Reading and Boiling Point of
Water at Various Altitudes D-3
Saturation: Temperatures, Steam Data (Courtesy of Goulds Pumps, Inc.) D-4
Vapor Pressure—Various Liquids (Courtesy of Goulds Pumps, Inc.) D-6
June 1999 © <<1st published year (or latest revision)–year of latest revision>> Chevron USA Inc. All rights re-
Appendix D Properties of Water Pump Manual
Fig. D-1 Properties of Water at Various Temperatures From 32 to 705°F
D-2 © <<1st published year (or latest revision)–year of latest revision>> Chevron USA Inc. All rights reserved.
Pump Manual Appendix D Properties of Water
Fig. D-2 Atmospheric Pressure, Barometer Reading and Boiling Point of Water at Various Altitudes
June 1999© <<1st published year (or latest revision)–year of latest revision>> Chevron USA Inc. All rights re-
Fig. D-3 Saturation: Temperatures, Steam Data (1 of 2) (Courtesy of Goulds Pumps, Inc.)
Pump Manual Appendix D Properties of Water
Fig. D-3 Saturation: Temperatures, Steam Data (2 of 2) (Courtesy of Goulds Pumps, Inc.)
June 1999© <<1st published year (or latest revision)–year of latest revision>> Chevron USA Inc. All rights re-
Fig. D-4 Vapor Pressure—Various Liquids (Courtesy of Goulds Pumps, Inc.)
Appendix E Specific Gravities
Contents Page
Approximate Specific Gravity of Petroleum Fractions
(Courtesy of GPSA) E-2
API and Baume Gravity Tables and Weight Factors
(Courtesy of Goulds Pumps, Inc.) E-3
June 1999 © 1999 Chevron USA Inc. All rights reserved. E-1
Appendix E Specific Gravities Pump Manual
Fig. E-1 Approximate Specific Gravity of Petroleum Fractions (Courtesy of GPSA)
E-2 © 1999 Chevron USA Inc. All rights reserved. June 1999
Pump Manual Appendix E Specific Gravities
Fig. E-2 API and Baume Gravity Tables and Weight Factors (Courtesy of Goulds Pumps, Inc.)
June 1999 © 1999 Chevron USA Inc. All rights reserved. E-3
Appendix E Specific Gravities Pump Manual
E-4 © 1999 Chevron USA Inc. All rights reserved. June 1999
Appendix F Mechanical Seals Bellows Failure
Contents Page
Mechanical Seals Bellows Failure F-2
June 1999 © 1999 Chevron USA Inc. All rights reserved. F-1
Appendix F Mechanical Seals Bellows Failure Pump Manual
Mechanical Seals Bellows Failure (1 of 2)
Fig. F-1
F-2 © 1999 Chevron USA Inc. All rights reserved. June 1999
Pump Manual Appendix F Mechanical Seals Bellows Failure
Mechanical Seals Bellows Failure (2 of 2)
Fig. F-1
June 1999 © 1999 Chevron USA Inc. All rights reserved. F-3
Appendix F Mechanical Seals Bellows Failure Pump Manual
F-4 © 1999 Chevron USA Inc. All rights reserved. June 1999
Appendix G Hardness Tables
Contents Page
Approximate Conversion Table for Hardness Numbers
Obtained by Different Methods G-2
June 1999 © 1999 Chevron USA Inc. All rights reserved. G-1
Appendix G Hardness Tables Pump Manual
Fig. G-1 Approximate Conversion Table for Hardness Numbers Obtained by Different Methods*
G-2 © 1999 Chevron USA Inc. All rights reserved. June 1999
Appendix H Factory Test Procedure
Contents Page
H1.0 Scope and Purpose H-2
H2.0 How to Use H-2
H3.0 Documents to Have on Hand for Test H-2
I. Preliminary Instructions and Definitions
II. Test Preparation
III. Conduct of Test and Data To Be Recorded
IV. Data Development
V. Other Tests
VI. References
VII. Attachments
June 1999 © 1999 Chevron USA Inc. All rights reserved. H-1
Appendix H Factory Test Procedure Pump Manual
H1.0 Scope and Purpose

This factory test procedure is a general guide for testing centrifugal pumps in the
factory. It assumes that the pump has been constructed in accordance with the
purchase specifications. Pump performance can be evaluated from data collected by
the testers using the purchasing standards. Knowledge of the kind of data collected
during performance tests is helpful for using this guideline, but necessary defini-
tions, calculations, and references are included. Also covered briefly are procedures
for conducting cavitation (NPSH) and hydrostatic tests.
H2.0 How to Use

This guideline is a general description of the steps required to judge a pump’s
performance. Each statement is self-explanatory and should be completed or
acknowledged before moving on. It is divided into procedural sections:

II. Test Preparations
V. Other Tests
VI. References
VII. Attachments
The required equations are in the Data Development section or are referenced when
mentioned. A computer program to perform the common calculations is available.
See the Data Development section for instructions.
H3.0 Documents to Have on Hand for Test

• Purchase Order
• Pump Data Sheet (attached to Purchase Order)
• Purchase Specifications (attached to Purchase Order)
• Industry Specifications referenced
H-2 © 1999 Chevron USA Inc. All rights reserved. June 1999
Pump Manual Appendix H Factory Test Procedure

The following information will normally be determined from a pump performance
test:
Data Units
Flow Rate gallons per minute (GPM)
Head foot-pounds/pound (feet)
Power Input horsepower (BHP)
Efficiency percent (%)
Speed revolutions per minute (rpm)
Vibration — Velocity inches per second peak (IPS)
— Displacement thousandths of an inch peak-to-peak (mils)
Bearing Oil Temperature degrees Fahrenheit (°F)
Mechanical Seal Leakage Rate drops per minute
Flow Rate is usually measured with orifice or venturi meters, which require unob-
structed flows for specified minimum lengths in the discharge or suction line. These
are connected to either a manometer or differential pressure gage and the readings
are converted to a flow rate using a calibrated chart. Capacity measurements by
change of weight or volume over a specified time period are also reliable.
Head is defined as the energy content per unit weight of fluid. The energy is in foot-
pounds and the unit weight is the pound, resulting in feet as the unit of head. Energy
in a liquid can take the form of pressure, kinetic energy (velocity), or potential
energy in height above a datum point. The datum point is an arbitrary elevation,
usually at the shaft centerline of a horizontal pump or at the inlet eye of the first
stage of a vertical unit. Instruments commonly used are calibrated pressure gages
and manometers. All pressure readings are converted to feet. (See Section IV, Data
Development, for equations.)
Power Input to the pump shaft from the driver is called brake horsepower (BHP). It
may be determined with a transmission dynamometer, torsion dynamometer, strain
gage set-up or with a calibrated motor and wattmeter. The first three devices are
delicate instruments with high degrees of accuracy which should be checked before
and after each test for valid results. The calibrated motor is the most practical and
most often used means of measuring brake horsepower in shop tests. Determine the
current flow (average in each phase winding for a polyphase motor), voltage, and
power factor at the motor terminals. Then using the motor efficiency, horsepower
can be calculated from the equation given in Section IV, Data Development, of this
guideline.
Pump output, water horsepower (WHP) or hydraulic horsepower is proportional to
the capacity (Q), the total head developed (H), and the specific gravity (SG) of the
working fluid. (See equations in Section IV, Data Development.)
Pump Efficiency is the ratio of WHP to BHP. It is always less than 100%, and the
value changes as the pump operating conditions are varied.
Pump Speed (rpm) can be monitored by a tachometer, a revolution counter and

stop watch, or a stroboscopic device with a frequency indicator.
Vibration measurements are usually taken on the pump bearings in both the vertical
and horizontal planes during the shop performance test. The instruments needed are
a seismic vibration pickup, or a proximity probe if available, and the appropriate
readout and analysis equipment.
II. Test Preparation

Test Setup
Proper instrumentation must be available which will allow all required parameters
to be recorded. Make sure that the tester records the manufacturer’s serial number,
type, size, dimensions, and physical conditions of the pump and all associated
equipment (meters, tanks, etc.) to clearly identify the specific units used.
Shop tests will normally be conducted with a calibrated test motor equipped to
determine the pump input horsepower. The job driver cannot be used without the
purchaser’s approval.
The inspector should thoroughly review and understand the test system. Examine
the test setup to verify the adequacy and accuracy of the testing apparatus. Note on
the test data log anything unusual, significant, or questionable in the test system.
Shop gages should be certified accurate to 1/4% of scale, or a calibration curve
should have been prepared within the last 6 months prior to a test. For best accu-
racy, gages should be selected so that the expected readings are 50% to 75% of the
full scale values.
Attached are sketches of a typical horizontal pump performance test setup (Attach-
ment A). The test system may vary somewhat with different pump types, but should
retain the same essential elements. Prior to starting the shop test, be sure to check
for correct direction of rotation.
Special Considerations
Test performance should closely approximate the rated performance. This can
normally be ensured by matching the speed and capacity, and applying the neces-
sary viscosity and specific gravity corrections (see Section IV, Data Development).
Under certain special conditions, the pump test may be performed at a speed other
than rated. This is only allowed with prior consent of the purchaser.

General
A suction system booster pump may be used to slow roll the pump under test. This
is done to check to see if the rotor is free rolling before full power is applied.
Purge all air from instruments, case and system, and start with the discharge valve
set to prevent driver overload. No suction valve shall be throttled during pump oper-
ation, except for NPSH tests.
Do not accept shop tests if the mechanical seals show leakage. In packed pumps, the
packing should allow some leakage for packing lubrication. Normally, the job
packing will not be used in shop tests.
In performance tests, the bearing oil temperature must remain within the purchase
order specifications. As a general rule, a pump will be unacceptable if the lube oil
temperature exceeds the lesser of 180°F, or 70°F above ambient. For circulating
lube systems, the maximum differential (T - T) should not exceed 50°F at the most
adverse specified operating condition. At the beginning and end of a test, record the
bearing oil (in and out if pressure lubricated), suction side water, and ambient air
temperatures. Record the lubrication flow rate, which should be the same as
expected at service conditions.
In some cases, the purchase order may require some minimum seal cavity pressure
(typically 25 psi) above the pumped fluid vapor pressure. The required pressure
difference between suction pressure and stuffing box pressure will be different if the
rated fluid specific gravity is different from the test fluid (water). See Section IV,
Data Development, for instructions on how to calculate the required pressure.
Watch pump shutdown for smoothness. If any irregularity appears (binding, exces-
sive vibrations or sounds) it must be investigated, repaired, and the pump retested.
The pump shall be repaired and retested if any rubs are evident during an open
pump inspection, if such inspection is required in the purchase order. Do not allow
any significant mechanical alterations without retesting the pump.
At the conclusion of the performance test, sign the test log “Witnesses by ...” and
date. Obtain a copy of it at that time. Also, ask for a copy of the performance
curve. Most vendors can produce one “while-you-wait.” If not, insist on one within
1 week.
The results of the test at the rated point should be computed and found acceptable
before the test equipment is disassembled, unless specific approval is given by the
purchaser.
Performance Data
Shop performance test data will be taken at a minimum of five flow rates, including
rated (as on pump data sheet), and at least 120% of rated flow. The remaining test
points should be chosen to display the full range of the pump’s performance.
Usually readings are also taken at shutoff position (zero flow).
At each flow rate (data point), the inspector must read all instruments and agree
with the values recorded by the technician which will be as a minimum:
• Suction pressure (psig)
• Discharge pressure (psig)
• Flow rate (GPM — may require some conversion)
• Input power (kW — convert to horsepower)
• Speed of pump shaft (rpm)
• Vibration (with filter in and out; any unusual values should be reported to
purchaser)
Rapid fluctuations are short-term oscillations of values on the recording meters in a
single observation. The mean value is to be estimated and recorded, but for shop
tests at near best efficiency, the Hydraulic Institute Standard rapid fluctuation limits
are:
Data % of Value
Differential across pump (ΔP) ± 2%
Discharge pressure (Pd) ± 2%
Suction pressure (Ps) ± 3%
Rate of flow (Q) ± 2%
Speed (N) ± 0.3%
Power input to pump (BHP) ± 1%
Vibration Data
For pumps with sleeve bearings, shaft vibration measurements should be made with
a proximity probe, or a seismic vibration pickup with a shaft rider. For antifriction
(ball) bearings, a seismic vibration pickup will be used on the bearing housing. The
measurements should be recorded in vertical, horizontal, and axial planes. The
following vibration limits for readings filtered at running speed frequency apply to
API 610 and PMP-EG-983 pumps operating at a rated speed and at ±10% from
rated capacity in shop tests.
Peak-to-Peak Vibration Limits (API Fifth Edition)

Speed (rpm) Antifriction (inch/second) Sleeve Bearings (mil)
1800 and below 0.2 2.0
3600 0.2 1.75
3600 and above 0.2 Consult API 610
In shop tests, the unfiltered vibration velocity readings should also be measured
with a seismic pickup on the bearing housing. If the values at any operating point
exceed the limits, if any, specified on the purchase order, report the corresponding
frequency and amplitude to a mechanical specialist before accepting the pump. As a
general guide, the unfiltered velocity should not exceed 0.3 inch-per-second (IPS) at
rated capacity, and any other capacity within the normal operating range.

The performance curve shall be prepared from the test data log. (See Attachments B
and C for sample test data log and performance curve form.) Check the curve at
rated capacity and compare values of head, efficiency, and brake horsepower with
the design values given on the pump data sheet. These results can vary from the
rated conditions within the following tolerances for shop tests.
API Tolerances at Rated Speed and Rated Capacity (Seventh Edition)

Rated Differential Heat Guaranteed Point (%) Shutoff (%)
0–500 ft –2 + 10
+5 – 10
501–1000 ft –2 +8
+3 –8
Over 1000 ft –2 +5
+2 –5
Rated efficiency – ½ point of efficiency
Rated horsepower +4
NPSHR +0
For rated differential head of 0-500 feet at shutoff position, the value of –10% is
only allowed if, for a specified rising head capacity curve, the test curve at this point
still shows a rising characteristic.
For all non-API shop tested pumps, Hydraulic Institute Standards should apply
(Fourteenth Edition):
At rated head, + 10% of rated capacity, or
At rated capacity, + 5% of rated heads under 500 ft
+ 3% of rated heads above 500 ft
No minus tolerance allowed for capacity, total head, or efficiency at specified
conditions.
Equations commonly used to develop test data are given below
Head (Fluid Energy Content)

1. Pressure (Gage Head - feet): Hg = P × 2.31/SG
2. Kinetic Energy (Velocity Head - feet): Hv = V2/2g = V2/64.4
where:
D = Nominal pipe diameter (inches)
Q = Capacity (GPM)
P = Pressure (psig)
SG = Specific Gravity (SG)
V = Flow Velocity (ft/sec) (V = 0.41 × Q/D2)
3. Potential Energy (Static Head - feet): Z

where:
Z = Upward distance from datum elevation
Total Differential Head (Change in Fluid Energy Content)

H = (Hg + Hv + Z) discharge - (Hg + Hv + Z) suction
Power
Input:
BHP = Volts × Amps × Power Factor × Motor Efficiency- 0.5

--------------------------------------------------------------------------------------------------------------------------------------------- × [3]
746
(For a single phase motor, remove [ 3 ]0.5 from the above expression.)
or:
BHP = Kilowatts × Motor Efficiency-

-------------------------------------------------------------------------
0.746
Output:
Q ( GPM ) × H ( feet )
WHP = ------------------------------------------------- × Specific Gravity
3960
Efficiency
E pump = WHP
--------------
BHP
Affinity Law Corrections

When a pump is tested with speed or impeller diameter other than rated, values must
be converted to predict actual performance using the affinity laws.
For speed changes (N):
NS
Q S = Q T × -------
N T
2
⎛ N S⎞
H S = H T × ⎜ -------⎟
⎝ N T⎠
3
⎛ N S⎞
BHP S = BHP T × ⎜ -------⎟
⎝ N T⎠
or changes of impeller diameter (D):

DS
Q S = Q T × -------
D T
2
⎛ D S⎞
H S = H T × ⎜ -------⎟
⎝ D T⎠
3
⎛ D S⎞
BHP S = BHP T × ⎜ -------⎟
⎝ D T⎠
where:
S (subscript) = Service
T (subscript) = Test
H = Head (ft)
Q = Capacity (GPM)
BHP = Horsepower
D = Impeller diameter
The customer’s specifications may restrict the use of affinity laws for test data
correction. This is especially true for reductions of impeller diameter. Be sure to
review applicable specifications if impeller machining is required (API 610, Section
4.3.3.4.1, Seventh Edition).
Viscosity Corrections
If the pump is rated for conditions of fluid viscosity significantly different than
the test, the test results must be adjusted to reflect the expected performance. For
viscosity corrections, see Hydraulic Institute Standards, Fourteenth Edition,
pages 112–113.
Seal Cavity Pressure Corrections

To calculate the seal cavity pressure where applicable:
1. From the pump data sheet (purchase order), calculate the minimum seal cavity
pressure required.
Minimum Pressure = Vapor Pressure + Minimum Differential
For example:
Vapor Pressure = 10 psia
Minimum Differential = 25 psi
Minimum Pressure (Pminimum) = 10 psia + 25 psi = 35 psia
2. Determine the required differential between suction pressure and the seal cavity
pressure by taking the difference in Pminimum and the rated suction pressure.
For example:
Psuction = 20 psia
Pminimum = 35 psia
Suction Pressure and Seal Cavity Pressure Difference = P minimum - Psuction or:
P = 35 psia - 20 psia = 15 psi
Therefore, the difference between suction pressure and seal cavity pressure is
15 psi for rated conditions.
3. The differential pressure for test conditions depends on relative specific gravi-
ties.
1.0 ( specific gravity water )

P test = P rated × ----------------------------------------------------------------
SG ( specific gravity rated )
If the rated gravity is 0.5, then
1.0
P test = 15psi × ------- = 30 psi differential for water
0.5
Therefore, the seal cavity pressure should be 30 psi above suction pressure. The
seal cavity pressure should be measured in the stuffing box and not in the seal
recirculation line. The required seal cavity pressure is equal to the test suction
pressure plus the required differential.
V. Other Tests
Cavitation Test
When excessive pump noise, vibration, or unusual impeller eye damage is evident,
cavitation may be responsible. Cavitation occurs when the absolute pressure in the
pump inlet falls below the vapor pressure of the working fluid, and the fluid then
vaporizes rapidly. As these vapor pockets collapse in higher pressure areas, the
pump may be damaged by high resultant forces. And flow and pressure pulsations
will be produced in the discharge line.
The pump manufacturer will specify a head required above the fluid vapor pressure
that is the minimum to prevent cavitation. If this specified Net Positive Suction
Head Required (NPSHR) is to be checked, a test can be made by keeping differen-
tial pressure and capacity constant while the suction pressure is reduced. A drop in
differential head of 3% is an indication that damaging cavitation is present.
For a water NPSH test, the pump should operate without harmful cavitation as long
as the vacuum at the suction flange does not exceed Ps.
Ps = Pb - (2.04 Pv + [(NPSHR - h)] SG/2.311

where:
Ps = Vacuum at the pump suction flange at which NPSH available
equals NPSH required (inches Hg)
Pb = Barometric pressure (inches Hg)
Pv = Absolute vapor pressure (psia)
NPSHR = Net positive suction head required by manufacturer (feet)
h = Suction velocity head (feet)
SG = Specific gravity of water at test conditions
The same relationship should hold for any fluid if the correct vapor pressure and
specific gravity are used. A negative Ps should be interpreted as a positive gage
pressure (in inches Hg).
Hydrostatic Test
A hydrostatic test insures that all parts of the pump can safely withstand their
maximum design pressures without leakage. API 610 pumps are to be subjected to
150% of the maximum design pressure for at least 30 minutes without external
leakage. Multistage horizontal pumps are to be tested segmentally. Auxiliary equip-
ment lines carrying process fluids are to be tested at 150% of design values with a
lower limit of 150 psi. Cooling passages are to be tested at 115 psi.
Non-API pumps shall conform to the Hydraulic Institute Standard and shall with-
stand for at least 5 minutes with no external leakage the greater of:
• 150% of the pressure which would occur in that part of the pump during service
at rated conditions; or
• 125% of the normal pressure in that part at rated speed with the discharge valve
closed.
VI. References
• API 610, Centrifugal Pumps for General Refinery Service
Available from: American Petroleum Institute
Refining Department
2101 L. Street, NorthWest
Washington, D.C. 20037
• Hydraulic Institute Standards for Centrifugal, Rotary, and Reciprocating Pumps

Available from: Hydraulic Institute 1230
Keith Building
Cleveland, Ohio 44115
VII. Attachments
Sample Test Set-Up Sketch (Figure H-1)
Sample Test Data Log (Figure H-2)
Sample Performance Curve (Figure H-3)
Sample Test Set-Up Sketch
Fig. H-1
Sample Test Data Log Courtesy of Flowserve Corporation
Fig. H-2
Sample Performance Curve Courtesy of Flowserve Corporation
Fig. H-3
Appendix I Presentation at the 2007 Mechanical Equipment
Roundtable (MERT)
This appendix contains the slides from the presentation at the 2007 Mechanical
Equipment Roundtable (MERT).
March 2009  2009 Chevron U.S.A. Inc. All rights reserved. I-1
Appendix I Presentation at the 2007 Mechanical Equipment Roundtable (MERT) Pump Manual
1
0HFKDQLFDO(TXLSPHQW5RXQGWDEOH
3UHVHQWDWLRQDWWKH
0(57
DOC ID
© Chevron 2005
I-2  2009 Chevron U.S.A. Inc. All rights reserved. March 2009
Pump Manual Appendix I Presentation at the 2007 Mechanical Equipment Roundtable (MERT)
1
Mechanical Equipment Round Table
Multiphase Pump Project
Main Pass 59
October 2, 2007
DOC ID
© Chevron 2005
I-4
Appendix I Presentation at the 2007 Mechanical Equipment Roundtable (MERT)

Project Scope
Increase production from 18 Gas Lifted
Oil Wells on a Gulf of Mexico Platform by
Installing a Multiphase Pump to Reduce

the Well Back Pressure from 350 to 150
psig.
Pump is the largest in Chevron’s Fleet of

MPP’s and the first to be driven by a
Natural Gas Engine.
Pump Manual
March 2009
© Chevron 2005 DOC ID 2

March 2009
Pump Manual
AGENDA
• MP 59 Background / History
• MP 59 Facility Bottlenecks

• Debottlenecking Alternatives
• Why Multiphase?
• Pump and Driver Selection
• Procurement and Package Delivery
• Control System
• Current Project Status
I-5
4
MP 59
Background - Main Pass 59 Location
DOC ID
© Chevron 2005
March 2009
Pump Manual
Background - Main Pass 59 Production History
7/89 - A3D 1998 – Chevron assigned
Group: Main Pass 59 Wells In Group: 24 Format:drilled
CDRATESLOG 100% WI for P&A liability
9/97 – Compressor
10000 targeting
4200M fire destroys A 1999 farmout to Ridgelake
platform
1977 – Amoco discovery
1980-82: Platform installed and 6 wells drilled

targeting 5500 M, 5000 M, and 4900 M sands
5500 M1 Sand drilled

1000
100
No Drilling Activity 17 capital projects to develop

No Drilling Activity for 10
the 6050 M and M2, 5600 M,
For 5 years years
5500 M1, and 5600 M3
10.0
82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03
CDOIL,BOPD,GrpTotal CDGAS,MCFD,GrpTotal CDWAT,BWPD,GrpTotal

VS Time

I-7
I-8

MP 59 Recent Production
Current Production Rates

2,561 BOPD, 1,963 BWPD,
Cum Prod 9/82 – Present
6,000 MCFD (Formation + Lift Gas)
10,203 MBO
7,905 MMCFG
10,262 MBW
Pump Manual
March 2009

March 2009
Pump Manual
Current Facility Setup and Bottlenecks
Capacity
Available in IP
Gas Sales Gas Sales
Separators
LP Separators
At Capacity.

42 Gas & Liquids 41 8" Wet 41
Oil P/L
D/L 3 Ph Separation
N 2 Ph Separation
B/O 2 Ph Separation
Compression Compression Compression
Oil Treating Pump Liquids to 8" Bulk
Oil D/L
Sales Water Treating P/L
Liquid Capacity
4" G/L
Problem: Production From PL
6,000 BFPD
MP 59 A Wells Constrained
by 350 psi Back Pressure
caused by Long 8" Bulk PL. 59
A 6 Pile Jacket
Well Headers
Test Separator

I-9
I-10

Main Pass 59 Prospects
Additional 6050 M well
Drilled A-15 & A-17

in 2006
A-17
Untested fault block
A-15
Future Prospects in 6050 M2

Sd (up to 3 more wells)
Pump Manual
March 2009

March 2009
Pump Manual
Production Forecast
MP 59 Facility Upgrade Project - Gross BOEGD


3500
3000
2500
2000
1500
1000
500
0
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Base - No Drills Total Prod w/Upgrade & Drills

I-11
I-12

Conventional Alternative 1 - Larger Pipeline
3 Ph Separation
Gas Sales Compression 2 Ph Separation Gas Sales
Oil Treating Compression
Water Treating Pump Liquids to

D/L
42 Gas & Liquids 41 8" Wet 41 Debottleneck Piping
Oil P/L
D/L N B/O 2 Ph Separation
Compression
Oil 8" Gas

Sales
New 12" P/L
Liquid PL
4" G/L
PL
59
A 6 Pile Jacket
Well Headers
Add 2 Phase Test Separator
Separator
Pump Manual
March 2009

March 2009
Pump Manual
Conventional Solution 2 – More Topside Equipment

G/L Hdr
Glycol Existing 4" PL
Compressor
System To MP 41 B
2 Phase
LP Separator
Bulk Test
Hdr Hdr
Surge
Vessel
MP 59 A Pipeline Existing 8" PL

Pumps To MP 41 B

I-13
I-14

Multiphase Pump Alternative
Multi- Existing 8" PL

150 Psi Phase 700 Psi To MP 41 B
Pump IP System
Bulk Test
Hdr Hdr
MP 59 A
G/L Hdr
Existing 4" PL
From MP 41 B
Pump Manual
March 2009

13
Multiphase Advantages
Lower Capital Cost

Smaller Footprint
Less Equipment
Less Weight
+
DOC ID
© Chevron 2005
I-16

MPP Design Basis
Suction Pressure: 150 psig
Discharge Pressure: 680 psig
Suction Temperature: 120°F

Flowrates: Gas – 8 MMSCFD
Oil – 4,000 BPD
Water – 4,000 BPD
Equivalent Flowrate: 152,000 BPD
Gas Volume Fraction: 95%
Pump Manual
March 2009

March 2009
Pump Manual
Pump Selection
Leveraged Learnings From Previous Chevron MPP
Installations
• Twin Screw Pump

• Suppliers: Leistritz and Bornemann
• Stellite Hardening of Pump Liner
• Borriding or Nitriding of Pump Rotors
• Seals on Suction Side of Pump
No History of Sand Production From Wells

• Single Mechanical Seals with API Plan 31 Flush System –
Use Process Liquids to Flush Seals

I-17
16
of Leistritz
Courtesy
(Courtesy Pumps)
of Leistritz Pumps
Twin Screw Pump
DOC ID
© Chevron 2005
March 2009
Pump Manual
Driver Selection
1700 BHP Requirement
No Electrical Power on Platform


Considered Gas Turbine but Economics Favored
Reciprocating Engine
Several Waukesha Engines Already in Field
Final Selection: Waukesha 9390 GSI Gas Engine
Two Alternators Mounted on Engine to Power MPP Lube

Oil Pumps and Platform SCADA System
Purchased Small Standby Generator to Power SCADA

System when MPP not Running
I-19
I-20

Procurement
Lump Sump Bid Package Issued to Leistritz and
Bornemann for Skid Mounted MPP Package which
includes the following:

• Waukesha Engine Driver
• RemVue 600/AS Control System with Remote
Start Capability
Project Awarded to Leistritz based primarily on Price

and Seals on Suction Side.
Due to Long Delivery Time on New Engine we

Elected to Use a Refurbished Engine from Universal
Compression
Final Package Price: $3,050,000
Pump Manual
March 2009

March 2009
Pump Manual
Pump Package Delivery
Leistritz used Fluid Power Equipment in Houston to
Design/Fabricate Pump Skid.

Quoted Package Delivery Time – 9 Months
Actual Package Delivery Time – 13 Months
Delivery Delays Primarily Due To:
• Skilled labor shortage at Packagers Shop
• Incomplete Understanding of Instrumentation
Requirements by Packager
Best Practice: Conducted Test of Entire

Package at Fluid Power’s Shop using Water.
Identified several deficiencies in the skid
which would have been costly to correct
offshore.
I-21
I-22

Process Control (Original Design)
From To
Wells Pipeline
Bypass Valve
SDV
Opens When Pump
Not Loaded
Recycle Valve and

Pump Speed Used to
Control Suction
Pressure PSHL
PSV
Recycle
Valve
TSH
Equalization
Valve
20x12
24x12 20x12
12x1 12x1
Discharge
SDV Boot
PSHL
Multiphase
24x12
Pump
Pump Manual
Strainer
March 2009

March 2009
Pump Manual
Seal Flush System To
Pipeline
Tangential
Entry

Seal Flush Boot
From Multiphase
Wells
Pump
LSL
3x1
1" Needle
Valve
3x1 RO FAL RO
2" Globe 3x2 3" Globe

Valve Valve
I-23
22
DOC ID
© Chevron 2005
23
Remote Start Capability
DOC ID
© Chevron 2005
I-26

Seal Pressure Issue
Just prior to start-up, we leaned that pump seals

have a limitation of 230 psi of Suction Pressure when
Pump is rotating.
During start-up and when wells are unloading the

pump suction pressure can exceed 230 psi.
Leistritz currently working with seal supplier to

determine if seals with a higher pressure rating can
be installed.
Lesson Learned: Include a dynamic seal

pressure rating in MPP Specification.
Pump Manual
March 2009

March 2009
Pump Manual
Revised Process Control
From To
Wells Pipeline
Bypass Valve
SDV
Opens When Pump

Not Loaded
Add PCV to Maintain
Suction Pressure
Below 230 psi
PSHL
PSV
Recycle
Valve
TSH
Equalization
Valve
20x12
24x12 20x12
12x1 12x1
Discharge
SDV Boot
PSHL
Multiphase
24x12
Pump
Strainer
I-27
I-28

Current Project Status
9/7/07 - Initial Pump Start-up.
9/10/07 - Pump Ran for 8 Hours. Shut-down due to

Vibration-Induced crack on the Engine Jacket Water

Manifold.
Repaired Manifold and modified Jacket Water Piping.
9/25/07 - Restarted Pump, Pump Ran for 6 hours and

developed another Jacket Water Manifold Crack.
Currently installing New Re-designed Jacket Water

Manifold
Observed Significant Oil Production Increase

when Pump is Running
Pump Manual
March 2009

27
Questions?
DOC ID
© Chevron 2005
0XOWLSKDVH3XPSLQJ
March 2009
Pump Manual
*20%80DLQ3DVV$0XOWLSKDVH3XPS
VW 1DWXUDO*DV(QJLQHLQ
&KHYURQ¶VIOHHWRI
033V
+DQGOHVRLOZDWHU JDV
*9) IURPZHOO

V\VWHP
033FKRVHQRYHU
WUDGLWLRQDOVHSDUDWRUZLWK
SXPSDQGFRPSUHVVRU
,QFUHDVHGSURGXFWLRQE\ EDVHGRQLQFUHDVHG
EEOVGD\DWSUHVHQW SURGXFWLRQDQGFDSLWDO
FRVWVDYLQJVDQGVPDOOHU
/DUJHVW/HLVWULW] 7ZLQ6FUHZ
SODWIRUPIRRWSULQW
033LQVWDOOHGRQ0DLQ3DVV
$LQ*20 033VNLGFRVWRYHU
PLOOLRQLQVWDOOHGFRVWUDQ
+31DWXUDO*DV(QJLQH PLOOLRQ
,QLWLDO6LPSOH3D\RXWLV
\HDUVZLWKRLODW
EEO
I-31
I-32

3URMHFW6FRSH
,QFUHDVHSURGXFWLRQIURP*DV

/LIWHG2LO:HOOVRQD*XOIRI0H[LFR
3ODWIRUPE\,QVWDOOLQJD0XOWLSKDVH
3XPSWR5HGXFHWKH:HOO%DFN
3UHVVXUHIURPWRSVLJ
Pump Manual
March 2009
March 2009
Pump Manual
Main Pass 59A Bottlenecks
Capacity
Available in IP
Gas Sales Gas Sales
Separators
LP Separators

At Capacity.
42 Gas & Liquids 41 8" Wet 41
Oil P/L
D/L 3 Ph Separation
N 2 Ph Separation
B/O 2 Ph Separation
Compression Compression Compression
Oil Treating Pump Liquids to 8" Bulk
Oil D/L
Sales Water Treating P/L
Liquid Capacity
4" G/L
Problem: Production From PL
6,000 BFPD
MP 59 A Wells Constrained
by 350 psi Back Pressure
caused by Long 8" Bulk PL. 59
A 6 Pile Jacket
Well Headers
Test Separator
I-33
I-34

&RQYHQWLRQDO$OWHUQDWLYH /DUJHU
3LSHOLQH
3 Ph Separation
Gas Sales Compression 2 Ph Separation Gas Sales
Oil Treating Compression
Water Treating Pump Liquids to

D/L
42 Gas & Liquids 41 8" Wet 41 Debottleneck Piping
Oil P/L
D/L N B/O 2 Ph Separation
Compression
Oil 8" Gas

Sales
New 12" P/L
Liquid PL
4" G/L
PL
59
A 6 Pile Jacket
Well Headers
Add 2 Phase Test Separator
Separator
Pump Manual
March 2009
March 2009
Pump Manual
&RQYHQWLRQDO6ROXWLRQ± 0RUH
7RSVLGH(TXLSPHQW

G/L Hdr
Glycol Existing 4" PL
Compressor
System To MP 41 B
2 Phase
LP Separator
Bulk Test
Hdr Hdr
Surge
Vessel
MP 59 A Pipeline Existing 8" PL

Pumps To MP 41 B
I-35
I-36

MPP Design Basis
6XFWLRQ3UHVVXUHSVLJ
'LVFKDUJH3UHVVXUHSVLJ
6XFWLRQ7HPSHUDWXUH)
)ORZUDWHV *DV± 006&)'
2LO± %3'
:DWHU± %3'
(TXLYDOHQW)ORZUDWH
%3'
*DV9ROXPH)UDFWLRQ
Pump Manual
March 2009
March 2009
Pump Manual
MPP Value Added
Initial Valued Added:

$80,000/day (crude at $80/bbl) based

on initial early increased production
(not all wells returned to service yet)
Simple Pay out: 0.17 years.
I-37
Multiphase Advantages
Lower Capital Cost

6PDOOHU)RRWSULQW
/HVV(TXLSPHQW
/HVV:HLJKW
+
=

Chevron Gray Pump Manual

Uploaded by

Copyright:

Available Formats

Chevron Gray Pump Manual

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chevron Gray Pump Manual

Uploaded by

Copyright:

Available Formats

100 General Information

140 Application and Selection Criteria 100-23

120 Pump Categories

Fig. 100-1 Pump Categories

122 Centrifugal Pumps

Fig. 100-2 Pump Categories Courtesy of the Hydraulic Institute (1 of 2)

123 Positive Displacement Pumps

Fig. 100-3 Pump Categories Courtesy of the Hydraulic Institute (2 of 2)

124 Oil Field and Special Service Pumps

• Jet pumps (fluid eductors)

125 Multiphase Pumps

130 System Hydraulic Design

132 Overview of the Hydraulic Calculation Sheet, PMP-EF-467

133 PMP-EF-467 Instructions

134 Calculation I (Initial Estimate)

135 Calculation II (Pre-purchase)

1. Enter the Pump Service and Liquid Properties

In addition to the effect on fluid properties from alternate service conditions, it is

2. Complete the System Sketch

3. Complete the Pressure Calculations

Fig. 100-9 Example: Pressure Calculations

Net Positive Suction Head Available (NPSHA)

H = Original Pressure. The lowest normal operating absolute pres-

4 × 1.84 × 360 × 0.066 = 3.88ft

20 × 0.811 × 360 × 0.066

h a = 3.88 + 8.55 = 12.4ft

Method A - Quick Estimation

Control Valve Pressure Drop Selection

Fig. 100-12 Method B for Calculating Flow Pressure Losses (1 of 2)

Method B - Detailed Approach

Combined System/Pump Curves

Fig. 100-12 Method B for Calculating Flow Pressure Losses (2 of 2)

b. Superimpose the two system curves from steps 1 and 2 as shown in

4. Establish Pressure Profiles

d. Downstream equipment design pressures, which determine minimum

136 Calculation III (Final Review)

140 Application and Selection Criteria

142 Pump Selection Fundamentals

100 General Information

LIMITED-SPACE APPLICATIONS. REVIEW

144 Centrifugal Pumps

• Available in many different materials of construction

Poor Applications for Centrifugal Pumps

145 Reciprocating Pumps

Good Applications for Reciprocating Pumps

Poor Applications for Reciprocating Pumps

146 Rotary Pumps

Fig. 100-16 Positive Displacement Pump Selection and Application

Fig. 100-17 Centrifugal Pump Selection and Application

Good Applications for Rotary Pumps

Poor Applications for Rotary Pumps

• Abrasive material or possibility of running dry. Rotary pumps are ordinarily

147 Miscellaneous Pumps

Fig. 100-18 Miscellaneous Pumps

150 Considerations for H2S Services

160 Suction Piping Design Considerations

161 NPSHA and Fluid Velocity

Suction Line Fluid Velocity for Reciprocating Pumps

Suction Line Fluid Velocity for Centrifugal Pumps

Boiler Feedwater Application