EG50099 Flow Assurance T.

Baxter

Flow Assurance
Pressure Drop
Flow Patterns
Slugging

1 INTRODUCTION 3
1
EG50099 Flow Assurance T. Baxter

1.1 Typical Subsea Development 4

2 FLUID FLOW ENERGY EQUATIONS 4

2.1 Hold-Up 5

3 FLOW REGIMES 6

3.1 Horizontal Flow Regimes 7

3.2 Vertical Flow Regimes 8

3.3 Mandhane Flow Regime Map 9

3.4 Baker Flow Regime Map 10

3.5 Taitel Dukler Flow Regime Map 14

3.6 Govier Vertical Flow Map 15

3.7 Wet Gas Pipeline Pigging and Slugcatcher 16

3.8 Eaton Wet Gas Pipeline Hold-Up 18

4 PRESSURE DROP PREDICTION 19

4.1 Pressure Drop Models Category 1 20

4.2 Pressure Drop Models Category 2 20

4.3 Pressure Drop Models Category 3 20

4.4 Homogeneous Model – Category 1 21

4.5 Lockhart and Martinelli – Category 1 21

4.6 Beggs and Brill – Category 3 23

4.7 Mechanistic Models 23

5 SLUG FLOW 24

5.1 Hydrodynamic Slugging 25

5.2 Terrain Induced Slugs 26

5.3 Severe Slugging 26

6 PIPELINE TOPOLOGY 30

7 COUPLING OF HYDROCARBON RESERVOIR AND PRODUCTION WELLS 30

7.1 Tubing Diameter 30

2
EG50099 Flow Assurance T. Baxter

7.2 Gas Liquid Ratio 31

7.3 Surface Pressure 32

1 Introduction

3
EG50099 Flow Assurance T. Baxter

Flow assurance has become a critical technology area for the oil and gas industry. The
simultaneous flow of gas and liquid through pipes, often referred to as multiphase flow,
occurs in almost every aspect of the oil industry. Multiphase flow is present in well tubing,
gathering system pipelines, and processing equipment. The use of multiphase pipelines has
become increasingly important in recent years due to the development of marginal fields and
deep water prospects. In many cases, the feasibility of a design scenario hinges on cost and
operation of the pipeline and its associated equipment. Confidence in the flow assurance
analysis – pressure loss, heat transfer, fluid chemistry, transient effects – is essential to
developing a safe and operable design.
Multiphase flow in pipes has been studied for more than 50 years, with significant
improvements in the state of the art during the past 15 years. The best available methods can
predict the operation of the pipelines much more accurately than those available only a few
years ago. The designer, however, has to know the key areas requiring analysis and which
methods to use in order to deliver the most accurate results.

1.1 Typical Subsea Development

In addition to multi-phase pipelines, flow
assurance also involves the design of single
phase compressible and incompressible flow.
A typical subsea development with multiple
duties is shown in figure.
Flow assurance will involve the design and
operation of;
Single phase incompressible flow –
Water injection, oil export, chemical
injection (methanol, corrosion inhibition
etc.)
Single phase compressible flow –
Gas lift, gas export.
Multi-phase flow –
Production, test pipelines

2 Fluid Flow Energy Equations

4
EG50099 Flow Assurance T. Baxter

Recall Bernoulli’s equation and the lost work (friction) term.

dP ∆ (v ) 2
∫ ρ
+ g .∆X +
2
= −W f − W

Pressure Elevation Velocity Friction Work

change term term term term

ρ ⋅ f ⋅ L ⋅ v2
∆Pf =
2⋅ D

For multi-phase flow it is apparent that a single density no longer applies as does a single
velocity. Similarly other necessary properties viscosity, heat capacity etc. Immediately it is
clear multi-phase flow will be much more complex to analyse.

2.1 Hold-Up
Liquid hold-up is the ratio of the volume of liquid in the system to the total volume. It is an
important feature of multi-phase flowlines as it will affect pressure drop, slug generation, rate
of pipeline cooldown and other design and operational features.
If an equal volume rate of gas and liquid are introduced into a horizontal pipeline it might be
expected that the liquid
hold-up i.e. the fraction of
liquid volume to the total
volume, would be 50%.
This is unlikely to be the
case as the gas is much
more mobile than the liquid
and will slip through the pipeline resulting in a higher than 50% liquid hold up.
The level of hold up will also be influenced by the pipeline angle. In upflow the gas will slip
much more readily through the liquid resulting in higher hold-ups. In downfall the liquid will
start to run downwards due to gravity hence the hold up will be smaller.
Furthermore, as the liquid viscosity is significantly greater than the gas velocity, a pressure
force will give a larger gas velocity than liquid velocity. This will cause a slip between the
(average) gas and liquid velocity. This slip (hold-up)is not caused by gravity, and thus is also
present in horizontal pipe sections.
Terminology used in the development of multi-phase flow correlations follows;

5
EG50099 Flow Assurance T. Baxter

Liquid Holdup Hl
Ratio of the volume of a pipe segment occupied by liquid to the volume of the pipe segment
Hl = volume of liquid in a pipe segment/volume of pipe segment
Value varies from 0 to 1
Gas Holdup or gas void fraction Hg
Hg = 1.0 - Hl
No-Slip Liquid Holdup – λl
Ratio of volume of liquid in a pipe segment divided by volume of pipe segment which would
exist if the gas and liquid travelled at the same velocity (no-slippage)
λl = ql / (ql + qg)
Where ql and qg are liquid and gas volumetric flow rates
No-Slip Gas Holdup – λg
λg = 1.0 - λl = qg / (ql + qg)
Two-Phase Velocity/ Mixture Velocity - vm
The velocities of the gas and liquid in the pipe are prime variables in the prediction of the
behaviour of the multiphase mixture. Most multiphase flow prediction methods use the
superficial gas and liquid velocities as correlating parameters. The superficial velocities are
defined as the in situ volumetric flowrate of that phase divided by the total pipe cross
sectional area. Or, the velocity in the pipe if the gas or liquid were flowing alone.
The following can be readily shown;
vm = vsl + vsg mixture velocity, vm, is the sum of the two superficial velocities
vsl = ql / A – superficial liquid velocity vl = ql / (A.Hl) – actual liquid velocity
vsg = qg / A – superficial gas velocity vg = qg / (A.Hg) – actual gas velocity
λl = ql / (ql + qg), but vsl = ql / A and vsg = qg / A, so
λl = vsl / (vsl + vsg) = vsl / vm
Note, some researchers refer to superficial velocity as j, the volumetric flux.

3 Flow Regimes
In multiphase flow, the gas and liquid within the pipe are distributed in several fundamentally
different flow patterns or flow regimes, depending primarily on the gas and liquid velocities,
physical properties and the angle of inclination. Observers have labelled these flow regimes
with a variety of names. Over 100 different names for the various regimes and sub-regimes
have been used in the literature.

6
EG50099 Flow Assurance T. Baxter

A number of workers have generated flow regime maps which can be used to determine
which flow regimes are likely to occur within a particular pipeline (or parts of the pipeline)
under various operating conditions. It is important to know which flow regime since the
regime will determine the pressure drop along the pipe, the liquid hold-up within it and the
nature of the fluid mixture which will have to be treated on arrival. Research on two phase
flow regimes goes back many decades (Baker 1954). Papers by Mandhane et al (1974), Taitel
and Dukler (1976) are often quoted.
With the exception of Taitel and Dukler, these researchers have used an empirical approach to
interpreting experimental results. Taitel and Dukler produced a more theoretical or
mechanistic approach and they compared this successfully with published results. This was a
landmark advance for the science.

3.1 Horizontal Flow Regimes

Though other patterns exist, generally four horizontal regimes are defined: Bubble, Slug,
Annular and Mist flow.
Plug/Bubble Flow
− Pipe is almost completely filled with liquid and the free gas phase is present in small
bubbles.
− Pipe wall is always contacted by the liquid phase
Slug or Intermittent Flow
− Liquid phase with accompanying gas bubble gas
Stratified Flow
− Pipe acts as a separator
− Liquid flows along bottom with gas above
Annular Flow
− High velocity gas core
− Liquid annulus surrounds a gas core
Dispersed Bubble Flow
− Bubbles of gas flow within a predominantly liquid phase
− Nearest regime to no slip conditions

7
EG50099 Flow Assurance T. Baxter

3.2 Vertical Flow Regimes

For vertical flow similar regimes are defined. However, as can be seen stratified flow does not
exist.

8
EG50099 Flow Assurance T. Baxter

Experimental studies of flow regime transitions have shown that each of the flow regime
boundaries reacts differently to changes in the system variables. The following table shows
the sensitivity of the transitions to changes in the major system variables:

Many researchers have attempted to develop simple flow regime maps, usually using
dimensionless parameters on each axis (e.g. Baker, Beggs & Brill). These methods can be
inaccurate since no single parameter can model the sensitivity effects indicated above.

3.3 Mandhane Flow Regime Map

Mandhane proposed a fluid property correction to the superficial velocities, but concluded
that the fluid property effects are insignificant compared to the errors in the empirical map.
The map reports the flow regimes: stratified, wavy, annular mist, bubble, slug, and dispersed.
Care should be taken in the interpretation of these flow maps as the regime boundaries are
strongly affected by pipe inclination. Clearly, horizontal flow regime maps must not be used
for vertical flow, and vertical flow regime maps must not be used for horizontal flow. The
Mandhane map was developed for horizontal lines flowing air and water at near atmospheric
pressure. As is seen later, inclinations in the range of 0.1-1.0 degrees can cause substantial
regime boundary movement.
In addition, flow regime boundary adjustment has been observed due to fluid pressure, pipe
diameter, viscosity and surface tension. The gas density increase caused by high pressure acts
to move the slug-mist boundary to lower superficial gas velocities, while increased pipe
diameter acts to increase the stratified wavy flow regime at the expense of the slug flow
regime. In addition, foamy fluids having a high surface tension have been observed to flow in

9
EG50099 Flow Assurance T. Baxter

the dispersed flow regime even though Mandhane would have predicted superficial liquid
velocities too low to cause dispersed flow.

3.4 Baker Flow Regime Map

The Baker Flow Regime map incorporates physical properties not included within the
Mandhane chart. In the Baker chart, GL and GG are the liquid and gas mass velocities, μ lL is
the ratio of liquid viscosity to water viscosity, μ lG is the ratio of gas density to air density, ρlL
is the ratio of liquid density to water density, and σlLis the ratio of liquid surface tension to
water surface tension.
The reference properties are at 20°C (68°F) and atmospheric pressure, water density 1,000
kg/m3 (62.4 lbm/ft3), air density 1.20 kg/m3 (0.075 lbm/ft3), water viscosity 0.001 Pa × s,
(1.0 cp) and surface tension 0.073 N/m (0.0050 lbf/ft, 73 dyne/cm). The empirical parameters
λ and ψ provide a crude accounting for physical properties.

10
EG50099 Flow Assurance T. Baxter

These two methods can indicate completely differing flow regimes. The engineer now has to
make a decision on which method he/she considers more accurate. The simplistic single
variable axis with Mandhane would be a concern suggesting Baker to be more realistic.
Most early researchers studied multi-phase flow in either horizontal or vertical with low
pressure air, water and steam systems. The inherent concern was what effect will different
properties have on the flow maps. Some of the effects are illustrated in the following figures.
As can be seen, and probably as expected, the effects can be very significant.

11
EG50099 Flow Assurance T. Baxter

12
EG50099 Flow Assurance T. Baxter

Such significant effects raised concerns that flow regime models were likely to be flawed if
scaled to high pressure oilfield conditions.

13
EG50099 Flow Assurance T. Baxter

3.5 Taitel Dukler Flow Regime Map

Taitel and Dukler made the first attempt at understanding the underlying the physics of flow
transitions. The map uses the Martinelli parameter X, the gas Froude number, Fr and the
parameters T and K. The map is composed of three graphs as shown later.
The Froude number (F or Fr), is a dimensionless quantity used to indicate the influence of
gravity on fluid motion. It is generally expressed as Fr = v/(gd)0.5, in which d is depth of flow
or diameter, g is the gravitational acceleration (equal to the specific weight of the water
divided by its density, in fluid mechanics), v is the velocity of a small surface (or gravity)
wave, and Fr is the Froude number. When Fr is less than 1, small surface waves can move
upstream; when Fr is greater than 1, they will be carried downstream; and when Fr = 1 (said
to be the critical Froude number), the velocity of flow is just equal to the velocity of surface
waves.

14
EG50099 Flow Assurance T. Baxter

3.6 Govier Vertical Flow Map

The correlation by Govier, et al. (Can. J. Chem. Eng., 35, 58–70 [1957]), may be used for
quick estimate of flow pattern.
Rv = Vsg/Vsl
Film flow is the same as annular flow.
15
EG50099 Flow Assurance T. Baxter

Froth flow is churn flow

3.7 Wet Gas Pipeline Pigging and Slugcatcher

Pipelines are pigged for several reasons. If liquid water is present (holdup) in the line, it must
be removed periodically in order to minimize corrosion. This water accumulates in dips in the
pipeline, and these low spots are particularly susceptible to corrosion.
Pipelines are also pigged to improve pressure drop-flow rate performance. Water or
hydrocarbon liquids that settle in the dips constitute constrained areas that increase pressure
drop and reduce flow rates. Pigging or sphering removes these liquids and improves pipeline
efficiency. The sphere will propelled at the pipeline mixture velocity.
Pigging can also be used as a means of limiting the required slug catcher size. By pigging at
frequent intervals, liquid inventory buildup in a pipeline can be reduced, and the maximum
slug size can be limited. The required downstream slug catcher size must take into account
pigging frequency. The amount of liquid in the pipeline is identified by calculating the liquid
hold-up.
Slug catchers are devices at the downstream end or other intermediate points of a pipeline to
absorb the fluctuating liquid flow rates. Slug catchers may be either a vessel or constructed of
pipe. They provide residence time for vapor-liquid
disengagement. Particularly for high pressure service, vessel
separators may require very thick walls. In order to avoid
thick wall vessels, slug catchers are frequently made of
multiple pipes. Lengths of pipe many tens of meters long are

16
EG50099 Flow Assurance T. Baxter

used as slender horizontal separators. The pipe is generally inclined from one to ten degrees
and banks
of these slightly inclined pipes are frequently manifolded together. The are often described as
finger type slug catchers.

The following figure shows the variation in predicted liquid hold-up for a large wet gas
transmission line. As can be seen the correlation predictions can differ by an order of
magnitude. Once gain the engineer is concerned about method accuracy and flow assurance.
Use of an inappropriate correlation could result in a significant under or over prediction of
slugcatcher size.

17
EG50099 Flow Assurance T. Baxter

3.8 Eaton Wet Gas Pipeline Hold-Up

A correlation developed by Eaton is widely used for estimating hold-up in pipe lines flowing
stratified. The Eaton holdup correlation is shown in the opposite figure. In this figure, the
holdup fraction is plotted directly as a function of a dimensionless group. The dimensionless
groups are of the form:

18
EG50099 Flow Assurance T. Baxter

A worked example is provided at the end of the lecture notes. Note units of d are mm.

4 Pressure Drop Prediction

Accurate prediction of pressure drop in a two/multi-phase system is an essential element of
flow assurance. Inaccurate analysis of pressure loss will mean that the pipeline is over or
undersized which on both counts could lead to
loss of production and loss of revenue together
with serious operational problems.
The boundary pressure is often an existing
pressure on a host installation. Hence, the sub
sea architecture must be designed to ensure
that the production profile can be delivered to
the host pressure inlet requirements.
Using the host pressure as a boundary,
calculations are undertaken to quantify
required flowing wellhead pressures. These

19
EG50099 Flow Assurance T. Baxter

are compared with flowing wellhead pressures established from well lift curves to ensure
there is adequate wellhead pressure to deliver the production fluids.
There are three categories of flow correlations

4.1 Pressure Drop Models Category 1

− No-slip, no flow regime considered
− Mixture density and viscosity based on input average phase densities
− Gas and liquid assumed to be travelling at same velocity in pipe
− Only require a two-phase friction factor correlation
− No distinction made for different flow regimes.
Examples:
− Lockhart Martinelli
− Poettmann and Carpenter
− Baxendell and Thomas
− Fancher and Brown
− Homogeneous Flow Model

4.2 Pressure Drop Models Category 2

− Slip considered, no flow regime considered
− Gas and liquid travel at different velocity in pipe
− Requires a two-phase friction factor correlation and liquid holdup correlation
− No distinction made for different flow regimes.
Example:
− Hagedorn and Brown

4.3 Pressure Drop Models Category 3

− Slip considered, Flow Regime considered
− Requires method to predict flow regime
− Requires specific two-phase friction factor and liquid holdup correlations for each
flow regime
− Acceleration pressure gradient depends on flow regime
Example:
− Duns and Ros
− Orkiszewski
− Aziz, Govier and Fogarasi

20
EG50099 Flow Assurance T. Baxter

− Beggs and Brill

The correlations reflect the complexities of multi-phase flow with all work being very
empirical in nature.

4.4 Homogeneous Model – Category 1

The simplest approach to the prediction of two-phase flows is to assume that the phases are
thoroughly mixed and can be treated as a single-phase flow. This homogeneous model will
obviously work best when the phases are inter-dispersed. This type of model utilises a two-
phase friction factor calculated from a two-phase Reynold’s number.
In determining two-phase Reynolds number the following equations is often used to calculate
a two-phase viscosity.

μn = μl. λl + μg. λg
Similarly a no slip two-phase density is utilised.

ρ n = ρ l. λl + ρ g. λg
Recall λl and λg are the no slip liquid and gas hold up.
The two-phase Reynold’s number becomes;
Ren= ρn.vm.d/μn
Where vm is the no slip mixture velocity.
The two – phase friction factor can be determined using the Blasius expression;
ftp = 0.079. Ren-0.25
Two-frictional pressure drop becomes;
ΔPf = 2. ftp. ρn.vm2 /d

4.5 Lockhart and Martinelli – Category 1

Rapid approximate predictions of pressure drop for fully developed, incompressible
horizontal gas/liquid flow may be made using the method of Lockhart and Martinelli (Chem.
Eng. Prog., 45, 39–48 [1949]). First, the pressure drops that would be expected for each of
the two phases as if flowing alone in single-phase flow are calculated. The Lockhart-
Martinelli parameter X is defined in terms of the ratio of these pressure drops:

21
EG50099 Flow Assurance T. Baxter

The two-phase pressure drop may be then be estimated from either of the single-phase
pressure drops, using;

where YL and YG are read from the accompanying figure as functions of X. The curve labels
refer to the flow regime (laminar or turbulent) found for each of the phases flowing alone. The
common turbulent-turbulent case is approximated well by;

Lockhart and Martinelli correlated pressure drop data from pipes 25 mm (1 in) in diameter or
less within about 650 percent. In general, the predictions are high for stratified, wavy, and
slug flows and low for annular flow. The correlation can be applied to pipe diameters up to
about 0.1 m (4 in) with about the same accuracy.

22
EG50099 Flow Assurance T. Baxter

4.6 Beggs and Brill – Category 3

This is perhaps the best known and widely used correlation. It was developed from
experimental data from a small scale test facility and field data and compared well in many
oilfield applications. The correlation presents different values for liquid holdup for each
horizontal flow regime. Initially, the liquid holdup is calculated assuming the pipe is
horizontal and then the value is corrected for the actual pipe inclination.
The correlation is modified to include transition zone between segregated and intermittent
flow regimes.
A two-phase friction factor, independent of flow regime but a function of holdup is calculated.
Flow Regime Determination
The following variables are used to determine which flow regime would exist if the pipe were
horizontal. Note, this flow regime is a correlating parameter and does not reflect the actual
flow regime.
Froude Number: Nfr = Vm2 / (g.d)
Liquid Velocity Number NLv = vsl . (ρl .σ)0.25
λl = vsl / vm and λg = vsg / vm
A worked example is included at the end of this section – for reference.

4.7 Mechanistic Models

The older correlation methods such as Beggs and Brill are now seldom used by Flow
Assurance engineers. The mechanistic models developed by SINTEF are widely accepted
across the industry as state of the art. The OLGAS model is recognised as an industry leader.
OLGAS is based on data from the SINTEF two-phase flow laboratory near Trondheim,
Norway. The test facilities were designed to operate at conditions that approximated field
conditions.
− The test loop was 800m long and of 8”
diameter and operated at pressures between 20 and
90 barg.
− Gas superficial velocities of up to 13 m/s,
and liquid superficial velocities of up to 4 m/s were
obtained.
− Different hydrocarbon liquids were used
(naptha, diesel, and lube oil) in order to simulate the

23
EG50099 Flow Assurance T. Baxter

range of viscosities and surface tensions. Nitrogen was used as the gas.
− Pipeline inclination angles between 1° were studied in addition to flow up or down a
hill section ahead of a 50m high vertical riser.
− Over 10,000 experiments were run on this test loop during an eight year period
running in both steady state and transient modes.
− OLGAS considers four flow regimes and uses a unique minimum slip criteria to
predict flow regime transitions.
OLGAS has become an industry accepted tool for the analysis of steady state multi-phase
systems. Also offered by the same group is OLGA – a state of the art transient mutli-phase
simulator. OLGA and OLGAS also have OVIP (OLGA® Verification an Improvement
Project) . This is a multi-client joint industry research and development program for
validation and improvement of OLGA®. The OVIP program is supported by BP, Chevron,
ENI, ExxonMobil, Petrobras, Statoil, Total and Shell. OVIP embodies by far the largest
multiphase flow data collection ever assembled. Containing both experimental and field data,
including gas/condensate pipelines, lower GOR pipelines, well, and transient data. Transient
modelling is a critical aspect of flow assurance and requires highly sophisticated simulators
for accurate analysis.

5 Slug Flow
Slug flow is particularly troublesome flow regime. It presents a difficult pattern for process
control systems.
When liquid and gas are flowing together in a pipeline, the liquid can form slugs that are
divided by gas pockets. The formation of liquid slugs can be caused by a variety of
mechanisms:
1. Hydrodynamic effects (surface waves)
2. Terrain effects (dip in pipe layout)
3. Pigging
4. Startup and blow-down
5. Flow rate or pressure changes
Hydrodynamic slugs, in horizontal and near horizontal pipes, are formed by waves growing
on the liquid surface for a height sufficient to completely fill the pipe. In vertical pipes the
hydrodynamic slugs are associated with Taylor bubbles. The hydrodynamic slugging is
difficult to prevent since it occurs over a wide range of flow conditions. The repeating

24
EG50099 Flow Assurance T. Baxter

impacts of hydrodynamic slugging can cause fatigue. It is therefore important to predict the
slug volume, velocity, and frequency of such slugs.
Furthermore, several hydrodynamic slugs can gather together due to terrain effects, creating
larger slugs – slug growth.
A particular form of terrain slugging is created where a pipeline slopes downwards and joins a
vertical pipeline (riser). In this flow phenomena liquid stalls in the riser, fills the riser and is
blown out when the accompanying gas bubble has sufficient pressure to overcome the
imposed hydrostatic head. The transient surges of liquid and gas associated with severe slugs
can cause major problems for topside equipment – typically separators and compressors.
Understanding and controlling the phenomena is essential for safe operation and to prevent
shutdowns and associated lost production.
Other types of slugging are initiated by pipeline operations. Pigging of a pipeline causes most
of the liquid inventory to be pushed from the line as a liquid slug ahead of the pig. Shut down
of a line will drain the liquid that is left in the line down to the low points. During restart the
accumulated liquid can exit the pipeline as a slug. Also, increasing or decreasing the flow rate
of either gas or liquid leads to a change in liquid holdup. This can come out in the form of a
slug, depending on the rate of change of flow rate.

5.1 Hydrodynamic Slugging

Hydrodynamic slugs occur spontaneously as a result of steady state operation in the slug flow
regime (see flow regime map). These slugs occur due to an instability of the fluids gas/liquid
interface causing liquid waves to reach the top of the pipe. The slugs are dependent of gas
and liquid flowrates, pipe inclination, surface tension, densities and viscosities of the fluid.
Hydrodynamic slugs are typically 30 – 100 pipe diameters and length.

25
EG50099 Flow Assurance T. Baxter

5.2 Terrain Induced Slugs

As the pipe angle changes (up), the liquid hold-up increases reducing the area available for
gas flow.
This reduced area causes the gas to accelerate creating waves on the film surface which grow
to form slugs.

Slugs then grow as they pick up liquid in low spots or join with an adjacent slug.
The empirical Scott, Shoham and Brill correlation takes account of slug growth and can be
used as a first approximation.

ln(Lm) = -2.663 + 5.441 [ln(d)]0.5 + 0.059 [ln(Vm)]

where
Lm = mean slug length(ft)
d = pipe diameter (in)

5.3 Severe Slugging

Severe slugging can be seen as a special case of terrain slugging – the geometry here is a
downward sloping flowline feeding a vertical riser (the a low spot being the riser base,
followed by a vertical section). Low spots in the pipeline topography can lead to pooling and
build up of liquid, which will eventually get swept out.
Terrain slugs differ from hydrodynamic slugs by the fact that they are influenced by the
terrain undulation and can occur at flow conditions either within or outside the theoretical
borders of the intermittent /slug flow regime. The pipe inclination has a strong influence on
the transitions between most flow regimes. Therefore with varying undulations, slugs formed
on an upward slope may be found to progress and persist further into a downstream section
operating in the stratified flow pattern area of the flow map. In general, slugs tend to increase
in size at upward slopes and decrease in size or even disappear on downward slopes.

26
EG50099 Flow Assurance T. Baxter

For a terrain slug to be initiated a number of criteria need to be satisfied that are somewhat
similar to those for severe slugging. For example, the flow in the upstream section needs to be
stratified, whereas the flow in the downstream section needs to unstable. Although often not
very accurate, several correlations exist in the literature to predict the length and frequency of
hydrodynamic slugging. To predict the effect of topological variations on the downstream
development of these slugs a transient simulation can be made with a dynamic tool utilising a
slugtracking option (such as available in OLGA). This option includes a model to insert slugs
in dips, based on the local criteria. Thereafter the slug front and tail is tracked when the slug
travels through the pipeline. The slug can grow or shrink, can split up in two slugs, or can
merge with other slugs. Results obtained with slug tracking should be interpreted with care.
This is not exact science, and there is room for improvement (both with respect to the
numerical and physical modelling). Experience shows that size and frequency of the slugs that
appear at the outlet with the slug tracking simulation are often dependent on user-defined
parameters for slug initiation. Terrain slug can growth is very much dependent on the
topography and flow regimes encountered during the slug propagation along the pipeline.
Severe slugging is an extreme case of terrain induced slugging. The cycle of slugging is
described.
Pipeline of downward gradient meets riser base.
1. Liquid blocks the base of the riser.
2. Riser fills with liquid if accompanying gas bubble has insufficient pressure to overcome
liquid head. Riser continues to fill.
3. Once riser is full the backpressure stabilises however the accompanying gas bubble
pressure continues to rise
4 Once the gas bubble pressure is greater than the riser back pressure the riser starts to unpack.
5. Slowly at first then accelerating as the head in the riser reduces. The accompanying gas
pocket then discharges into the separator.
Severe slugging produces gas and liquid flow and pressure transients which are very difficult
to manage.

27
EG50099 Flow Assurance T. Baxter

Severe slugging can occur if the following 3 conditions are satisfied:

1 Liquid blockage at the riser base can only occur if;

(dp/dt)riser > (dp/dt) flowline
Here the latter pressure gradient is the rate of pressure increase in the pipeline upstream of
the riser foot due to gas compression in the pipeline, whereas the former pressure gradient is
the rate of pressure increase at the riser base due to an increasing liquid head caused by liquid
entering the riser.
The ratio of (dp/dt) flowline / (dp/dt)riser is known as the severe slugging number often labelled
Πss.
For severe slugging to occur Πss <1 (the Boe criterion). This criterion shows that a decreasing
GLR (gas liquid ratio) will enhance severe slugging. This can occur at late field life, where
water break through will decrease the GLR.
2. The pipeline topography has a low point at the riser foot where liquid blockage
may occur.

28
EG50099 Flow Assurance T. Baxter

3. The flow line is operated either in the stratified or annular flow pattern but not in
slug flow.
The criteria for severe slugging are summarized in the flow pattern map for the pipeline
upstream of the riser base, as depicted in the following. Severe slugging will typically occur at
turndown production, and/or at late field life when the riser is in the hydrodynamic slugging
regime and the gas production is low.
When severe slugging occurs, more energy is consumed due to the regularly varying gas and
liquid velocities in the line. Extra static head loss occurs when the riser is fully filled with
liquid.

The possibility of severe slugging can be determined using the following procedure;
1. Assume all the liquid in the system acts to fill the riser – calculate (dp/dt) riser. Calculate
rate at which riser fills in m/s and convert to a pressure rise.
2. Calculate the no slip hold up and, assume no slip exists in the downward sloping portion
of the flowline.
3. Calculate the gas volume in the downward sloping line assuming no slip.
4. Calculate the gas volume flow at average pipe line conditions .
5. Calculate (dp/dt) flowline the ideal gas law – rate of pressure rise will be linear with volume
flow into gas volume in the sloping line section.
6. Evaluate (dp/dt) flowline / (dp/dt)riser – less than one severe slugging possible.
7. Calculate superficial velocities and check for stratified flow or annular flow.

29
EG50099 Flow Assurance T. Baxter

6 Pipeline Topology
Note that Pipeline topography can have a large influence on slugging characteristics and also
on pressure loss. Accurate seabed bathymetry is very important for multi-phase analysis.

7 Coupling of Hydrocarbon Reservoir and Production Wells

A key use of fluid mechanics is the prediction of
flowrates from producing wells. Here the reservoir is
coupled to the well. A stable condition is reached where
the rate the reservoir can deliver is matched by the
resistance to flow in the well. A convenient way to
analyse the reservoir and well behaviour is by the
application of gradient or performance curves.
The reservoir is characterised by an Inflow
Performance Relationship (IPR). This describes the
flow through the rock into the well.
The well is characterised by a Tubing Performance
Relationship (TPR). This describes the fluid mechanics in
the well. This can change from single phase to two-phase
as the fluid pressures reduces moving up the well. The fluid
mechanics therefore becomes very complex.
The intersection of the IPR and TPR curves determines the
rate of stable flow that can be expected from a particular
well. The analysis is usually undertaken on specialist
software.

7.1 Tubing Diameter

The effect of changing tubing diameter is shown for a mixture of liquid and vapour in a well.
Increasing diameter increases the production rate until an optimal diameter is reached.
Increasing the diameter further results in a rate reduction. This is a consequence of the two
components of pressure drop – friction and elevation. At higher diameters the liquid tends to
‘slump’ within the well resulting in higher mixture densities. Hence elevation/hydrostatic
head increase is higher than the drop in friction losses. This effect would not be evident with
single phase flow.

30
EG50099 Flow Assurance T. Baxter

7.2 Gas Liquid Ratio

The effect of varying gas liquid ratio is shown for a mixture of liquid and vapour in a well.
The gas liquid ratio could be increased by injecting gas into the well. This is known as gas lift.
As can be seen there is an optimal GLR where further increases result in a drop off in
production rate.
The explanation for the optimal GLR relates to the two pressure loss components; friction and
elevation. Whilst injecting gas will increase the fluid velocity thus increasing the frictional
losses, the reduction in mixture density and elevation losses is more significant.
However there comes a point where friction losses start to dominate and the opposite holds.

31
EG50099 Flow Assurance T. Baxter

7.3 Surface Pressure

The effect of varying surface pressure is shown for a mixture of liquid and vapour in a well.
Reducing surface pressure increases the production rate .As the surface pressure is reduced
the back pressure reduces hence flow naturally increases. With two phase flow in the well
mixture velocities will increase and mixture density reduce as the surface pressures is dropped.
However, as the surface pressure drops gas volumes at surface increase resulting in large
compression requirements. There is an optimum surface pressure which balances improved
production rates with compression and ancillary costs.

32
EG50099 Flow Assurance T. Baxter

Key Learnings

1. Superficial and mixture velocity

2. Hold – up and phase slippage
3. Flow patterns/regimes – flow regime determination
4. Slug production from pipeline sphering
5. Multi-phase pressure drop methods
6. Slugging flow – form and impact on receiving facility
7. Severe slugging mechanism
8. Severe slug criterion, Πss
9. Reservoir Well Hydraulics

33
EG50099 Flow Assurance T. Baxter

Taitel Duckler Worked Example

34
EG50099 Flow Assurance T. Baxter

35
EG50099 Flow Assurance T. Baxter

Repeat with x=0.5

36
EG50099 Flow Assurance T. Baxter

37
EG50099 Flow Assurance T. Baxter

Beggs and Brill

The correlation presents different values for liquid holdup for each horizontal flow regime.
Initially, the liquid holdup is calculated assuming the pipe is horizontal and then the value is
corrected for the actual pipe inclination.
The correlation is modified to include transition zone between segregated and intermittent
flow regimes.
A two-phase friction factor, independent of flow regime but a function of holdup is calculated.
Flow Regime Determination
The following variables are used to determine which flow regime would exist if the pipe were
horizontal. This flow regime is a correlating parameter and does not reflect the actual flow
regime.
Froude Number: Nfr = Vm2 / (g.d)
Liquid Velocity Number NLv = vsl . (ρl .σ)0.25
λl = vsl / vm and λg = vsg / vm

38
EG50099 Flow Assurance T. Baxter

39
EG50099 Flow Assurance T. Baxter

40
EG50099 Flow Assurance T. Baxter

41
EG50099 Flow Assurance T. Baxter

42
EG50099 Flow Assurance T. Baxter

43
EG50099 Flow Assurance T. Baxter

Hold-Up Eaton Method

44
EG50099 Flow Assurance T. Baxter

45