Piping System Design
Piping System Design
Piping System Design
System
Design
IL Greeff
W Skinner
Department of Chemical Engineering
University of Pretoria
February 2000
TABLE OF CONTENTS
1. INTRODUCTION ..............................................................................................................1
5. COMPRESSIBLE FLOW................................................................................................ 39
~- -.-·--- ·-: ·-
LIST OF SYMBOLS
a acceleration, m/s2
A area, m2
c sonic velocity, m/s
c constant of integration
corrosion allowance
flow coefficient
c, specific heat capacity at constant pressure, kJ/kmol. K
C, specific heat capacity at constant volume, kJ/kmol. K
d pipe diameter, mm
D pipe diameter, m
F energy loss in the system due to friction, J/kg
safety factor
pulse factor
Fx force in direction x, N
r Darcy friction factor
g acceleration due to gravity, m/s 2
h specific enthalpy, J/kg
H enthalpy, J
k specific heat ratio
k' generalised rheological constant
J( resistance coefficient
K, constant for 2-K method
IC constant for 2-K method
ke specific kinetic energy, J/kg
KE kinetic energy
L length, m
I mixing length, m
m mass, kg
non-isothermal exponent
M molecular mass .
Ma Mach number
MEB mechanical energy balance
mol or kmol (number)
rheological constant
exponent
II generalised rheological constant
N revolutions per minute
NPSH net positive suction head
N, turbulent exponent
p, velocity head
p pressure
pb back pressure
P, vapour pressure
pe specific potential energy, J/kg
PE potential energy, J
q heat, J/kg
Q heat, J
r general radius, m
proportionality constant
- ---... · - - . -. - ---........,4"
R radius of pipe, m
ideal gas constant, J/mol. K
Re Reynolds number, dimensionless
s specific entropy, J/kg.K
s entropy, JIK
Sp entropy production, JIK
SG specific mass
t time, s
T temperature
linear point velocity in the x direction, mls
average linear velocity, mls
maximum linear velocity, mls
velocity fluctuation components
. ···: ..
--~--~~- --~---::
1. INTRODUCTION
Chemical engineering is the science of the industry's manufacturing processes. Raw materials
are transformed into more valuable products by means of chemical, physical thermal,
biochemical and mechanical processing. It is performed by companies which are collectively
referred to as the process industry and in equipment collectively referred to as a plant. Planning
and evaluations associated with a specific envisaged manufacturing process are collectively
known as chemical engineering design or simply process design.
• PROCESS DEVELOPMENT
Material balances - correlate feed and product flow rates of processing units; render
flow rates necessary for dimensional design of equipment; also composition of fluids.
Energy balances- render temperatures at different stages of manufacturing; needed for
evaluation of physical properties of fluids.
Compositions and temperatures -selection of suitable construction materials.
• PLANT DESIGN
• INVESTMENT EVALUATION
After completion of the plant design, information is available to do a proper cost analysis.
This will determine whether the project is economically viable and whether the company
will proceed with the building of the plant.
1.1.2 DIAGRAMS
Several types of diagrams can be found for example architectural diagrams, instrumentation
diagrams, plant diagrams, diagrams of subsurface constructions, services diagrams (water, gas,
steam), pipe diagrams, pipe and instrumentation diagrams (P & ID) and flow diagrams.
1
Abbreviations, symbols for processing units and so called equipment tables are used to provide
maximum information on flow diagrams.
1.1.3 REPORTS
The basic principles of report writing are applicable. A typical layout of the main division is the
following:
• MATERIAL AND ENERGY BALANCES. Tables and figures (Bulk calculations in the
appendix).
• SAFETY. Dangers associated with operation and properties of materials (solids, liquids,
gases).
The design of piping systems can involve various engineering disciplines. It is based on a sound
knowledge of fluid dynamics combined with practical design guidelines and procedures.
Chemical engineers are usually involved in dimensional design and specification as well as the
selection of suitable construction materials. This course will focus on dimensional design. The
methods will either be applied to design new systems or to analyse existing systems. Analysis of
existing systems is necessary if unknown parameters need to be determined or if modifications
need to be implemented.
Components include pipe sections, tubes, valves, elbows and T-pieces, equipment like pumps,
compressors and flow measuring instruments.
• The main distinctions between pipes and tubes are in the methods of fabrication, finishing
2
and in their codes of standards. The surface roughness of pipes varies to such an extent
that differences must be taken into consideration in the calculation of friction factors.
Tubes are taken as smooth with minimal surface roughness.
• VALVES
• Gale valve- The closing element operates at right angles to the fluid flow, the flow is
straight through. The gate wedges into the body. With the gate in the fully open position
there is little resistance to the fluid flow. Not suitable for flow rate control, used only as a
stop valve. May be difficult to open should the downstream pressure fall with the valve in
the closed position.
• Globe valve- The closing element is parallel to the fluid flow, the body has a globular
shape. The disk can have various profiles to provide different controlled flow
characteristics. The head loss across these valves are large. Due to relatively large
disks the valves are limited to smaller sizes.
• Plug valve - The closing element is a quarter turn plug, straight or tapered, with a
rectangular opening through its centre. The flow is straight through and there is little
resistance to flow or head loss. The valve is suitable for coarse throttling. The valve can
be subject to sticking if not lubricated suitably.
• Ball valve- The closing element is in the shape of a ball with a hole through its centre.
The valve opens and closes through an angle of 90 degrees. Straight through flow
provides minimal resistance to flow. The ball can be manufactured with a contoured
cavity to give some degree of controlled flow characteristics.
• Butterfly valve- The valve disk opens and closes through a 90 degree angle. Often used
for the control of gas and vapour flows. May be used as a stop valve except under
severe unbalanced pressure conditions. Should not be used as a terminal valve except
at very low pressures.
• Check valve -Also known as a non-return valve. Used to prevent backflow of fluids in
process lines, closure effected by zero or reverse flow. Available in various patterns such
as lift disk or swing check. Produces resistance to flow, can cause water hammer if not
well designed and positioned.
• Diaphragm valve -The closing element is a diaphragm clamped between the body and
cover of the valve and separating the fluid from the operating mechanism. Little
resistance to flow. Diaphragm can be manufactured of rubber or various elastomers.
Was invented in 1929 by a SA engineer- PK Saunders.
• PIPING COMPONENTS
• Elbows differ in their angle and also in the ratio bend radius : pipe diameter.
• T-pieces include the following types- soft T (no flow in branched leg), hard T (no flow in
one main leg), reduced T (three flow streams are relevant and the diameters may differ).
3
• Various types of reducers and enlargers are used. Changes in diameter can be gradual
(a typical example is the ASA type) or sudden.
• EQUIPMENT
Strainers are used to remove solid particles larger than certain dimensions from fluids.
Flow measuring instruments are used to measure flow. They also provide signals for
flow regulating equipment. A typical example is the orifice plate. (Control valves are
sometimes also classified as equipment).
Fluid movers are used to move fluids from one process unit on a plant to another-
pumps for liquids and compressors for vapours or gases. To limit shut downs due to
faulty fluid movers, standby units are often installed in parallel.
1.2.2 STANDARDS
Standards are used for specification purposes. Important standards for piping systems are
standards for dimensions, material composition and material properties. Various standard
systems are in use. Examples are ASA, ASTM, API, AWWA, AISI, DIN, CABRA and BSJ. Each
system is subdivided into sections for pipes and tubes of different construction materials and also
for the various types of piping components and fittings. The general format for pipe system
specification is MATERIAL STANDARD, DIMENSIONAL STANDARD.
The material standard is a reference for information relating to the composition and properties of
the relevant construction material. Several of the given standards are used. In 1974 the UNS
(unified numbering system) was introduced and will hopefully become the only system to be
used. The latest edition of Perry's Chemical Engineers' Handbook is still using different systems.
Examples for mild steel are AISI1 020 and UNS G1 0200.
The dimensional standard is a reference for information relating to the diameter and wall
thickness of the pipe, tube, component or fitting.
The most important dimensional information for pipe and tube sections is that of nominal
diameter, outside diameter and wall thickness. The inside diameter can be calculated as
(outside diameter- 2 x wall thickness). Different formats are in use for pipes (variables are
material of construction and units of fabrication) and tubes (the same variables as for pipes).
Examples for mild steel pipes are 6" nom sch 40 and 150 mm nom 5 mm wall. Standards for
British pipes are in Perry. The standard for mild steel Sl pipes is given in appendix A1. In the
case of metal tubes formats are independent of the metal type; examples for metal tubes are 2"
nom BWG 14 and 50 mm nom 2 mm wall. The standard for British metal tubes is in Perry and
for Sl tubes in appendix A2.
Examples of dimensional standards for fittings and other pipe components are available in Perry.
The format is nominal diameter and class number. The class number refers to the maximum
allowable operating pressure (psig) and relates to the wall thickness. Plastic piping are also
4
specified according to class numbers. Dimensions for polypropylene and polyethylene pipes are
given in appendix A3.
1.2.3 DIAGRAMS
Piping systems are shown in various degrees of detail on design diagrams. Examples are pipe
diagrams, pipe and instrumentation diagrams and engineering flow diagrams. Pipe related
diagrams may be presented in orthographic and isometric formats. Use is made of various
abbreviations and symbols for process units. Process streams may also be abbreviated, e.g. 0
for oil and S for steam.
1.2.4 SYMBOLS
Various symbols are used to show piping components or process units on a diagram. Some
of the symbols to be used in this course are shown below.
-D-- Reducer
-E_G Orifice
[_] Column
5
1.3 LITERATURE
1. Walas, S M, Rules of thumb - selecting and designing equipment, Chern Eng, 75,
March 16, 1987.
2. Spitzgo, C R, Guidelines for overall chemical plant layout, Chern Eng, 103, Sept 27,
1976.
5. R. H. Perry, Green, W.G., Perry's Chemical Engineers' Handbook, 7th ed: McGraw-Hill,
1997.
6
2. PHYSICAL PROPERTIES AND UNITS
The most important physical properties for flow applications are density and viscosity. In
compressible flow applications thermodynamic properties also become important. Physical
properties are one of the most important aspects in chemical engineering design and the
literature is vast. Perry's Chemical Engineers' Handbook is a useful source for physical
properties. Another source is the databanks of flowsheeting simulators. If the physical
properties of a component is not known various estimation methods can be used to find the
properties e.g. group contribution methods. In the case of mixtures of liquids or gases properties
are estimated with thermodynamic methods. Modern ftowsheeting simulators are also very
useful in this regard.
2.1 UNITS
Sl units will mainly be used. Pipes and tubes of certain construction materials are however only
available in British units and some data in reference books are given in British units. Useful
conversion factors relevant to the course material are given in appendix B.
2.2 DENSITY
LIQUIDS
For design purposes liquids are considered to be incompressible and densities are functions of
composition and temperature but not of pressure. The density of most organic liquids, other than
those containing a "heavy atom" usually lies between 800 and 1000 kg/m 3 . Density can also be
calculated from specific mass if the latter is known:
PL = 1000 SG kg/m3
where SG=SG (tWCJ
For general engineering purposes it is sufficient to consider gases and vapours as ideal. Density
is then calculated using the ideal gas law:
P'V=nRT
MP
p= RT kglm'
where Tis always in Kelvin.
7
P'V=znRT
Pv=zRT
MP k I 3
P =-- g m
zRT
The compressibility factor can be estimated using an equation of state for real gases such as the
Pang-Robinson equation or the Redlich-Kwong equation. A generalised compressibility plot,
which gives z as a function of reduced pressure and temperature can also be used. For mixtures
of gases the pseudo critical properties ofthe mixture should be used to obtain the compressibility
factor:
P~.~m= Pc.a Ya + Pc.bYb + ···
T,,m =Tc.a Ya + T,;, y, + ...
where Pc and Tc are critical pressure and temperature, y is mol fraction, m refers to mixture and a
and b to the components.
2.3 VISCOSITY
Units for viscosity are cP for absolute/dynamic viscosity and eSt for kinematic viscosity. The
relation between the two viscosities is:
p=pv
The Sl units for viscosity are as follows:
1 cP = 1x 10'3 Pa.s
1 eSt = 1x10'6 m2/s
Viscosity of liquids vary with temperature and pressure but the pressure effect is not significant
except at very high pressures. Viscosity of liquids tend to decrease with an increase in
temperature whereas the opposite effect is found in the case of gases.
Specific heat capacities are required to find the specific heat ratio for gases which is used in
k=Cp
C,.
8
For a gas in the ideal state the specific heat capacity at constant pressure is given by:
CP =a+ bT + cT 2 + d T 3
Values for the constants in the equation are available in handbooks. Several group contribution
methods are also available for estimation of these constants.
The three-term Antoine equation can be used to determine vapour pressure for a pure
B
h1p.=A---
, T+C
2.6 LITERATURE
9
3. BASIC FLUID DYNAMICS FOR PIPE FLOW
Fluid dynamics is the branch of fluid mechanics that is concerned with the motion of fluids.
Previously fluid dynamics existed as two separate disciplines namely hydrodynamics and
hydraulics. Hydrodynamics is a mathematical science based on the equations of motion of an
imaginary ideal fluid. Results of hydrodynamic studies are of limited practical value. For this
reason empirical formulae were developed from experimental studies on fluid flow. When
dealing with liquids this subject is called hydraulics. In fluid mechanics today, the basic
principles of hydrodynamics are combined with experimental data which satisfy the need for
a broader treatment.
In an ideal fluid the effects of viscosity are completely neglected. When such a fluid flows in
a pipe, the shear stresses are absent and only the pressure and inertia forces are considered.
There is also no velocity variation in the direction perpendicular to the pipe axis and as a
consequence the fluid must slip past the solid boundary of the pipe wall. Ideal fluid flow is
frictionless, without any losses and is also referred to as the reversible flow of
thermodynamics.
When considering real fluid flow, the tangential stresses due to shear as well as the normal
stresses due to pressure are taken into account. It is especially in the region near a solid
boundary that real fluid flow differs from ideal fluid flow. A real fluid adheres to the solid
boundary and does not slip. This no slip condition at the solid boundary causes the velocities
of the different fluid layers to vary across the cross-section of a pipe between a zero velocity
at the wall and a maximum velocity in the centre.
The three-dimensional motion of any fluid can be described by the fundamental laws of fluid
dynamics and thermodynamics. These laws are mathematically formulated by the continuity
equation, the momentum equation and the energy equation. In addition to these laws certain
other relations are also employed in describing a fluid, for example the ideal gas law and
Newton's viscosity relation.
When the above laws are applied to flow in a straight pipe a one-dimensional approach works
very well since there is no curvature in the streamlines. The complete one-dimensional
differential equations describing flow in a pipe, applied in the x-direction, are the following:
ap
u- a
+ -(pu) = 0 ..... (3.1)
at ax
In the case of steady flow, equation (3.1) becomes:
10
a
- (pu) ~ 0 ..... (3.2)
ax
or more simply
which describes the conservation of mass in a system through which a fluid flows.
- - - g - + ~ a u ~ au +
2
1 aP az au du
!l- ..... (3.4)
p ax ax p ax 2 at ax dt
the first term represents the normal stresses due to pressure and the third term the tangential
shear stresses due to viscosity. The second term represents the gravitational forces, which
are zero in the x-direction if the pipe is horizontal. For an ideal fluid the viscosity is zero and
equation (3.4) reduces to the well known Euler equation:
1 ap
- - - g-
az ~
au
+
au
11- ..... (3.5)
pax ax at ax
The momentum equations in three-dimensional form are also known as the Navier-Stokes
equations.
The one-dimensional energy equation for fluid flow is derived here using the basic
thermodynamic balances.
where E = End and B = Beginning and refer to the system, I= In and 0= Out and refer to the
11
mass streams entering and leaving the system. Note that the sign convention for work is taken
as positive when done on the system and negative when done by the system. For steady flow
the energy of the system does not change and the balance reduces to:
La (H + PE + KE) - L 1
(H + PE + KE) = oQ + oW'
or
11H + 11PE + ME = oQ + oW 1
Potential energy and kinetic energy are calculated with the following well-known equations:
11pe = g/1z
Me =
1'1112
11h +g/1z + - - = q +w 1 ..... (3.8)
2
The entropy balance over the system, in terms of specific properties, is:
..... (3.9)
where Lis is the change in entropy of the mass streams and Sp is the entropy production. The
following thermodynamic relation is now used:
12
which on integration from inlet to outlet under the assumption of constant temperature yields:
Outlet
q ~T/I..s-TS
b p ..... (3.12)
Substituting equations (3.11) and (3.12) into (3.8) and noting that v ~ 1/p :
Outlet dP
J- + g/l..z + ..... (3.13)
Inlet p
where TJ:fp is the frictional loss of energy of the flowing fluid and is now substituted by F. For
incompressible flow p is constant and equation (3.13) reduces to:
dP
+ gdz + udu + oF ~ OW ..... (3.14)
p
For an ideal fluid (i5F =D) without work being done on the fluid, equation (3.14) reduces to the
well-known Bernoulli equation:
dP
+gdz+udu~o ..... (3.15)
p
Since it is inconvenient to deal explicitly with the variations in the flow and fluid properties that
occur at a pipe cross-section, average flow quantities need to be defined. In this regard it is
only the linear velocity that is considered to vary significantly over the pipe cross-section for
a real fluid. It is obvious that the average linear velocity is dependent on the velocity profile.
In order to express equation (3.14) in terms of the average linear velocity, a kinetic energy
correction factor aKE• is introduced:
dP
+ gdz + aKE u du + oF ~ oW ..... (3.16)
p
13
The kinetic energy equation can be determined with the following equation:
R l/. 3
(~) rdr
J0
11
..... (3.17)
where u1 is the linear velocity at a point and u is the average linear velocity.
M !<.u 2
+ gf'..z + u.K1<.-- + F = w ..... (3.18)
p . 2
Multiplication of each term with density transforms equation (3.18) to a pressure balance in
units of Pascals:
!<.u2
M + pgf'..z + pu.KE-- + pF = pw ..... (3.19)
2
Equation (3.19) is now expressed using the following symbols and this is referred to as the
Mechanical Energy Balance (MEB):
..... (3.20)
Equation (3.19) can also be expressed in units of metres when divided by the term p g, such
an equation is known as a head balance (HB) :
..... (3.21)
The MEB form of the energy equation will mainly be used in this course. The first three terms
of the MEB deal with pressure energy, potential energy and kinetic energy respectively. The
frictional pressure loss term (AP1 ') deals with losses due to components in a pipeline as well
as losses in straight pipe sections due to internal fluid friction and friction between the fluid and
the pipe wall due to viscosity. LIP" is the energy added to or removed from the system due to
a fluid machine of some kind.
A variety of flow features, including energy losses and velocity profiles, are affected by
14
whether the flow is laminar or turbulent. The Reynolds number is the criterion used to
distinguish between different flow regimes namely laminar, transitional and turbulent. This
dimensionless number is named after Osborne Reynolds, who was first to describe the
existence of laminar and turbulent flow quantitatively in 1883. The Reynolds number is the
ratio of the inertia forces to the viscous forces which are the only significant forces affecting
the flow pattern. Gravity and capillary forces do not have any effect on the flow pattern of a
fluid in a completely filled pipe. The general formula for the Reynolds number is the following:
Re = xup
~~
In the case of a pipe flowing full, the pipe diameter, D, is used as the linear dimension:
For flow in a straight circular pipe laminar flow exists for Reynolds numbers below 2000. It is
however subject to slight variations. In laminar flow, movement of the fluid appears as the
sliding of thin laminations over adjacent layers. The particles move in definite paths or
streamlines with relative motion occurring at a molecular scale.
i5u
't = ~~~ ..... (3.23)
oy
When laminar flow is developing in a pipe, a laminar boundary layer, or annular outer zone,
starts to grow against the pipe wall until the boundary layers from opposite sides meet at the
pipe axis. At this point, the flow is termed fully developed. Theoretically an infinite distance
is required for the flow to become fully developed, although it has been established that the
maximum velocity in the centre of the pipe will reach 99% of its ultimate value in the distance
L = 0.058ReD.
Between Reynolds numbers of 2000 and 4000 there exists a transition region in which the flow
can be laminar or turbulent. The value of 4000 is also known as the upper critical Reynolds
number, above which flow is normally turbulent. This value is indeterminate, laminar flow has
been maintained in circular pipes for values of Reynolds number up to 50 000. The value of
2000 is much more definite and that is why it can be defined as the true critical Reynolds
number.
15
In turbulent flow the velocity at a point in the flow field fluctuates in direction and magnitude.
These fluctuations are caused by a multitude of small eddies created by viscous shear
between adjacent particles. When turbulent flow is developing in a pipe, a laminar boundary
layer starts to grow at the pipe wall up to a point when transition occurs and the boundary layer
becomes turbulent. This turbulent boundary layer increases in thickness much more rapidly
than the growth of the laminar boundary layer until the layers from opposite sides meet at the
pipe axis to result in fully developed turbulent flow. At a smooth pipe wall, the fluctuation in
velocity in the direction of the wall must be zero causing the turbulence to be inhibited. This
results in a laminar-like sublayer next to the wall. This is not a true laminar layer since it is
momentarily disrupted by the adjacent turbulent flow. Because shear in this layer is
predominantly due to viscosity alone, it is called the viscous sub/ayer. Fully developed
turbulent flow will be found at about 50 pipe diameters from the pipe entrance for a pipe with
no special disturbance at the entrance; otherwise, the flow will be fully developed within a
shorter distance.
In turbulent flow the shear stress is made up of two components: the viscous stress and the
Reynolds or inertia stress due to the turbulent fluctuations:
' =
ou
11- - p 11
"v ..... (3.24)
oy
The concept of a boundary layer, within which viscosity is important, was advanced by Ludwig
Prandtl in 1904. According to Prandtl's hypothesis, the viscous effects of fluid friction at high
Reynolds numbers are limited to the boundary layer. In all the flow outside the boundary layer
(the core flow) the viscous effects can be ignored at high Reynolds numbers. Thus the core
flow can be considered ideal and is well described by the ideal Bernoulli or Euler equations.
For incompressible, laminar flow in a pipe the well-known Hagen Poiseuil/e Jaw can be used
to compute the frictional pressure drop through a straight pipe section:
..... (3.25)
Equation (3.25) can be derived from basic principles and is not empirical. The empirical
equation for the prediction of the frictional energy loss which is valid for both laminar and
turbulent flows is known as the Darcy-Weisbach equation:
- f'L pu2
M--- ..... (3.26)
f D 2
The equation expresses the loss in terms of an empirical friction factor, j', known as the Darcy
16
friction factor, which corresponds to fully developed flow.
3.5 RELATION BETWEEN THE FRICTION FACTOR AND SHEAR STRESS
It will now be shown how the friction factor is related to the shear stress. It is valid for laminar
and turbulent flow. Consider a fluid element and the various forces acting on it in figure 3.1.
~~~-Pipe wall
~-,____ {
~~ Surface Area, A,
k--- --T-~-----~----
r
Area ofplane, AP --
If acceleration forces can be considered neglible, a simple force balance results in:
or
PA-PA
I p ~.._p
-TA s =0 ..... (3.28)
..... (3.29)
f'...Pr
T = ..... (3.30)
2L
17
Combining equation (3.26) with equation (3.30) and simplifying, yields and expression for the
friction factor in terms of the shear stress :
..... (3.31)
Due to viscosity the different layers of a fluid travel at different velocities resulting in a velocity
distribution over the cross section of a pipe. In the case of fully developed laminar flow it can
be shown that the velocity profile is a perfect parabola:
..... (3.32)
where u 1 is the velocity at one point and urn"" is the maximum velocity in the centre of the pipe.
This profile is always valid for laminar flow and is independent of the condition of the pipe wall
i.e. whether the wall is smooth or rough. The average linear velocity in laminar flow can be
shown to be precisely one half of the maximum centerline linear velocity :
The kinetic energy correction factor can now be calculated for laminar flow by inserting
equations (3.32) and (3.33) into equation (3.17) and integrating. This renders a value for aKE
of exactly 2.
Since both the Darcy-Weisbach equation and the Hagen Poiseuille law are valid for laminar
flow, a simple friction factor relation can be derived by equating these equations, solving for
j' and substituting for the Reynolds number:
This result shows that for laminar flow the friction factor is dependent only on the Reynolds
number. In the critical region, between laminar and turbulent flow, values for the friction factor
are uncertain.
18
3.8 VELOCITY DISTRIBUTION AND aKe FOR TURBULENT FLOW
In turbulent flow the condition of the pipe wall influences the velocity profile. The roughness
of commercial pipe walls is described by the absolute roughness e, which is an indication of
the size of the projections on the pipe wall. It has been found experimentally that a pipe with
a given wall roughness will sometimes behave as a smooth pipe and other times as a rough
pipe depending on the Reynolds number. The velocity profiles encountered are different.
Whether a pipe is smooth or rough is determined by the size of the absolute roughness (height
of the projections) with respect to the thickness of the viscous sublayer (ii), which decreases
with an increase in Reynolds number. If the roughness does not project through the viscous
sublayer (e < ii) in turbulent flow, the surface is said to be hydraulically smooth. It is then the
viscous shear alone that determines the flow resistance, and roughness has no effect on the
flow. In rough pipe flow the roughness projections protrude through the layer (e > ii) causing
a broken up viscous sublayer. It is now the form drag of the protrusions that determines the
flow resistance.
The oldest representation of a smooth pipe velocity profile is the so-called power Jaw:
~'
11. = ( 1-- '')~ I ..... {3.35)
Umax R
A newer approach, which is mostly used for smooth pipe velocity profiles, is the Jaw of the wall
which is semi-empirical. This law states that there is a wall layer where most of the velocity
variation occurs. Three regions are apparent in this layer namely the viscous sublayer which
is laminar, a transition zone (also referred to as the buffer layer) and a turbulent core where
the velocity profile is described by a logarithmic function. Since the equations describing the
velocity profile is very complex, the profile is only described qualitatively with the aid of figure
3.2. The vertical scale is exaggerated so that the three zones are clearly visible.
19
Turbulent zone
~----------
Logarithmic velocity profile
y
/
/
/
-- ----~+~-
I
~~/_/_/_/
Viscous sublayer
In the viscous sublayer the flow is assumed to be laminar and the Reynolds stress due to
turbulent fluctuations is neglible. The following expression for the thickness of the viscous
sublayer can be derived:
SuD
0 = -'-----
..... (3.36)
Re J-c 0
/p
SD
0 =
~~
..... (3.37)
Re
The Prandtl mixing length theory is used to predict the logarithmic function describing the
velocity profile in this region analytically. In this region it is assumed that the viscous shear can
be neglected over most of the flow area and that the Reynolds stress dominates.
20
The Transition Region
The intersection of the velocity profiles for the viscous sublayer and the logarithmic layer
introduces an abrupt change in the velocity profile which is not realistic. The one curve must
rather merge gradually into the other, introducing a transition region where the viscous and
turbulent stresses are of the same magnitude.
In the case of rough pipe flow, the classification can be broken down into fully rough and
transitionally rough. In the first case the roughness protrudes right through the viscous
sublayer and the transition region to render a totally broken up wall layer. The resulting flow
is then completely turbulent. In transitionally rough flow the roughness projections protrude
only partially into the transition region. It is then both viscous shear and form drag determining
the flow resistance.
When < < o the roughness protrusions are contained within the viscous sublayer and the
roughness has no effect on the flow. When<> 14o the protrusions extend into the turbulent
layer and foro<<< 14o there is a transition region where flow resistance is a result of viscous
shear and form drag.
Since the development of the law of the wall by Prandlt, other authors have attempted to
describe the complete velocity distribution in a single equation. Experimental data have also
shown that the law of the wall equation for the logarithmic layer deviates from the real profile
near the centerline.
When the law of the wall for the logarithmic layer is assumed to be valid over the entire cross-
section of a pipe the following approximate relation can be derived for the kinetic energy
correction factor for turbulent flow:
In laminar flow the Hagen-Poiseuille law can be used to determine the relation for the friction
factor. In turbulent flow no such law exists and use is made of the velocity profiles to predict
the forms of the equations for the friction factor.
For smooth pipe flow this was first done by Prandtl in 1933 and is known as the theoretical law
of friction. The experimental work of Nikuradse contributed to the exact values of the constant
21
giving an equation known as Prandtl's smooth pipe equation for the friction factor:
By similar analyses von Karman developed the following equation for turbulent flow in fully
rough pipes:
_I_ = 2 log (
ll
.!2)
28
+ 1.74 ..... (3.40)
In 1939 Colebrook combined equations (3.39) and (3.40) to yield an equation that is
applicable over the entire range, smooth, transitionally rough and fully rough:
The problem with using more recent equations for the velocity distribution to derive equations
for the friction factor, is that they are complex and have to be numerically integrated. Churchill
and Chan derived an improved theoretically based expression for the friction factor but, as
stated in their article : "the net numerical corrections to the friction factor are too small to be
of practical interest".
The measuring and specifying of the roughness of commercial pipes remains a problem. The
roughness projections vary in size, shape and distribution and cannot be quantified in terms
of a single number. The experimental work done by Nikuradse entailed the coating of different
sizes of pipe with sand grains that had been sieved to ensure uniform diameters. The
diameters of the sand grains are represented by c, the absolute roughness. The correlations
for the friction factor are all based on Nikuradse's sand roughness scale. This means that in
order to use these correlations, one is bound to express any pipe roughness data on this
scale. Nikuradse did experimental work with values of the relative roughness (c/O) ranging
form 0.000985 to 0.0333.
3.11 LITERATURE
1. R. P. Benedict, Fundamentals at Pipe Flow: John Wiley & Sons, Inc., 1980.
22
2. J. B. Franzini, Fluid Mechanics with Engineering Applications, ninth ed.
Belfast: McGraw- Hill, 1997.
3. J. R. Welty, Fundamentals of Momentum, Heat, and Mass Transfer, third ed. New York:
John Wiley and Sons, Inc., 1984.
4. A. J. Ward-Smith, Internal Fluid Flow, 1 st ed. New York: Oxford University Press,
1980.
6. L. Prandlt, "Uber Flussigkeitsbewegung bei sehr kleiner Reibung," Verhandl. Ill Int.
Math. Kongr., Heidelburg, 1904.
7. L. Prandtl, "Uber die Ausgebildete Turbulenz," Proc. II Int. Congr. Appl. Mech., Zurich,
p. 62, 1926.
9. T. von Karman, "Uber laminare und turbulente reibung," Z. Angew. Math. Mech., vol.
1, p. 233, 1921.
10. T. Von Karman, "Aspects of turbulence problems," Proc. IV Int. Congr. Appl. Mech.,
Cambrigde, England, 1934.
11. J. Nikuradse, "Laws of turbulent flow in smooth pipes," Forsch. - Arb. lng. - Wesen,
val. 356, 1932.
12. J. Nikuradse, "Laws of flow in rough pipes," Forsch.-Arb. lng.-Wesen, vol. 361, 1933.
13. S. W. Churchill, "Friction factor equation spans all fluid-flow regimes," Chemical
Engineering, pp. 91-92, 1977.
14. S. W. Churchill, "Improved correlating equations for the friction factor for fully
turbulent flow in round tubes and between identical parallel plates, both smooth
and naturally rough," Industrial and Engineering Chemistry Research, val. 33, pp. 2016-
2019, 1994.
15. C. F. Colebrook, "Turbulent flow in pipes, with particular reference to the transition
region between the smooth and rough pipe laws," Journal of the Jnsliluteof Civil
Engineering, London, val. 11, pp. 133-156,1938-1939.
23
4. PRACTICAL APPLICATION OF THE ENERGY BALANCE ON PIPING SYSTEMS
The various terms of the MEB are now looked at in more detail:
..... (3.20)
..... (4.1)
where
AP1 = friction pressure loss in pipe sections, components such as valves (except control
valves) and fittings such as elbows.
APcv = friction pressure loss in control valves.
APEQ = friction pressure loss in equipment such as reactors, columns, filters and flow
measuring instruments.
The MEB can be applied between any two reference points in a piping system. Such
reference points are referred to as terminal points. Consider the system shown in figure 4.1.
The pressures P 1 and P 2 will normally be fixed and independent of flow rate. A MEB between
points 1 and 2 will contain all the terms in the given balance. A typical application of such a
balance
will be to establish the required APa for the system.
24
When the balance is applied to the reference points 3 and 4, the AP. term become irrelevant:
The pressuresP3 and P4 are not fixed and will be a function of flow rate. Such reference points
are referred to as intermediate points.
For the three integral terms (APE, LiPEP• APKJ, AX= XJa'""''"ampaint - X),P'"'"mpai"'· Their values can
be positive or negative. The three friction terms (L1P1, LiPcv. APEQ) are calculated by means of
specially developed methods. Apart from rare exceptions their values are positive.
For design purposes the terms L1P8 r and LIPEL between two terminal reference points are fixed
and values are independent of flow rate. In certain applications it is handy to group those
terms, that do vary with flow rate:
..... (4.2}
where TV= Total Varying. Due to the control valve flow regulating mechanism, LIPcv increases
with decreasing flow rate. The other three terms decrease with decreasing flow rate. In certain
applications these three terms are grouped:
..... (4.3}
where STV = Subtotal Varying. The MEB forms the basis for most piping system evaluations.
Different methods have been developed to calculate the different MEB terms for different flow
systems. These will be discussed in the following sections.
25
4.2.1 STRAIGHT PIPE SECTIONS
Frictional pressure loss in straight pipe sections are calculated using the Darcy equation:
Pa ..... (3.26)
The following forms of the Darcy equation are also very handy:
The friction factors are a function of Reynolds number. Reynolds number also indicates
laminar or turbulent flow. The following equations are handy for the calculation of Reynolds
number:
Re ~ Dup
all parameters in Sl units
J.l
Re ~ 354W
with Win kg/h, din mm and f1 in cP
dJ,I
354Vp
Re with Vin m3/h, din mm and f1 in cP
du
26
Laminar I Turbulent flow
The critical zone is avoided in design since the flow is unstable and friction factors are
uncertain.
The difficulty in calculating accurate friction losses lies in the selection of a value for the
absolute pipe roughness. Pipe roughness generally increases with age due to corrosion,
pitting, etching or deposition of sediment or attachment of algal or bacterial slime to pipe walls.
Typical values for absolute roughness for various pipe materials are given in appendix C1.
If roughness values for a given pipe material are not available in literature it will have to be
determined experimentally. The roughness projections cannot be measured physically and
an experiment to obtain the value will involve a flow test to generate values of the friction
factors for various Reynolds numbers. Roughness can then be calculated using the Colebrook
equation rendering a value that corresponds to the Nikuradse sand roughness scale.
In turbulent flow the relations for friction factor is much more complicated. The Colebrook
equation has long been accepted as the formula to use in design for turbulent flow:
The only disadvantage of the equation is the implicitness of the friction factor,}. A trial and
error method is required to calculate] from given values of Re and e/D. In order to overcome
this difficulty many explicit equations have been proposed to use in the place of the Colebrook
equation. One such an equation that seems to be widely accepted was proposed by Haaland
in 1983:
27
In the past numerical values for the friction factor were read from a chart prepared by Moody
in 1944, also referred to as the Moody diagram. This diagram has been plotted with the aid
of equations (3.39) and (3.40) and is shown in appendix C2. The beauty of the Moody
diagram is that the different zones of importance are clearly visible, namely the laminar, critical,
transitional and fully rough zone. The chart has been plotted to cover the range of roughness
values and Reynolds numbers that would typically be encountered in practical situations,
rendering values for the friction factor varying between 0.008 and 0.1. These correspond to
values for the kinetic energy correction factor for turbulent flow ranging between 1.0216 and
1.27, according to equation (3.38):
The dotted line separating the transition region from the fully rough region was suggested by
R.J.S. Pigott and the equation for this line is:
Re = 3500
..... (4.5)
E/D
In 1977 Churchill introduced a friction factor equation that is valid for both rough and smooth
pipes as well as for the full range of laminar, transitional and fully turbulent flow regimes :
12 ] 1
j'-8 8 1 12 ..... (4.6)
- [ ( Re ) + (A +B)u
A = ( -2.457 In [ ( ;e r 9
+ 0.27 (;) lr
and
The Churchill equation eliminates the need to test for the flow regime and is very suitable for
use in computer programming codes.
28
4.2.2 PIPING SYSTEMS
Piping systems consist of pipe sections (often with more than one diameter), fittings and other
components. The Darcy equation can not directly be applied for the evaluation of friction
pressure losses in fittings and other components - the contribution of form friction is larger
than in pipe sections and length and diameter dimensions are complex. Other methods were
developed for such evaluations. They are based on equivalent lengths and resistance
coefficients.
The equivalent length of a fitting or component is that length of pipe section which will render
the same friction pressure loss as the restriction. Equivalent lengths are determined
experimentally and are tabulated in literature as the number of equivalent pipe diameters (LID).
The equivalent length of the restriction (Lr) can be calculated as follows:
The inside diameter of the pipe in which the unit will be mounted must be used for the
evaluation.
The friction pressure drop through the restriction can be calculated using the Darcy equation:
f' Lr pu 2
;.,pi ~
kPa ..... (4.8)
D 2000
In the calculation of the friction pressure loss of a pipe system it is standard practice to first
calculate a global equivalent length of the system. In the case of a single diameter system:
Then
~ f' Le pu
2
kPa ..... (4.10)
D 2000
Literature data for LID values are subject to inaccuracies. In reality the value for a restriction
is a function of certain dimensions as well as Reynolds number. These dependencies are
often omitted and only average values are given. The data given in appendix C3 are valid for
Re > 1000. When used for calculations where Re < 1000 the results should be treated as rough
estimations. This is sometimes acceptable in piping system evaluations.
29
Resistance coefficient method
It is also known as the velocity head method. The following format of the friction pressure drop
equation was used in the development:
f'L pu2
D.P = - · kPa
f D 2000
..... (4.11)
pu2
Pv = = velocity head ..... (4.13)
2000
The friction pressure drop through a restriction can be calculated with any of the given
equations. In the case of a pipe system with only one diameter the global resistance
coefficient can be calculated as:
where
_ f'Lp
Kp - -
D
and
..... (4.15)
Literature data forK values are subject to inaccuracies for the same reasons as those of LID
values. Depending on the type of restriction, values for K may be more dependent on
30
Reynolds number than on LID values, or less dependent; note thatK ~ (LID)f'. The Kvalues
given in appendix C4 are valid for Re > 2000. For Re < 2000 the K values become highly
inaccurate; for rough estimations it may be assumed that the LID values would be
independent of Reynolds numbers; this enables the evaluation of K values at low Reynolds
numbers by prorating:
For more accurate calculations the two-K method can be used. The two-K method takes the
dependency of K-values on Reynolds number and exact geometry of the fitting into account
in the following equation:
Kt I
K~ +K(l+-) ..... (4.17)
Re o• d
where
In the cases where restrictions are associated with changes in diameter (e.g. reducers and
enlargers) and/or changes in flow rates (e.g. T pieces) different possibilities for the calculation
of velocity head exist. The data source must specify which velocity head should be combined
with the given K value for the calculation of friction pressure drop.
• First consider those restrictions to be treated with the K method or two-K method and
calculate EKr
• Transfer EKr to 2:Lr!Dsflbtoraz
• Add all the LID values of those restrictions to be treated with the LID method to find
LIDTotal
• Calculate ELr and then Le ~ ELp + ELr for calculation of the friction pressure loss in the
usual way.
31
Systems with more than one diameter
Various approaches can be followed. Each diameter can be evaluated on its own. The friction pressure
Joss of the system is obtained by summation:
Mf ="
L..J !'J.P!;
In certain applications it is more convenient to combine the different diameter systems into a single system,
based on one diameter with a single Le or Ke. The friction pressure drop of the system can then be
evaluated in a single calculation with any of the given equations.
Prorating is used in combining the different diameter systems. Consider a two diameter system for
illustration purposes:
The term (Keb + (d/d) 4 Kej is known as the resistance coefficient of the system based on db. Similarly an
equation can be derived for calculations based on da :
kPa
32
The term (Le• + Le.(djd,J' (f'jf'~) is known as the equivalent length of the system based on diameter "b".
Similarly with base diameter "a":
In practice the diameters normally do not differ substantially,.('.~ f'• and the friction factor ratio term is
omitted in the correlations.
Division of Flow
Division by a T-piece renders two flow systems between two sets of reference points. Each must be
analysed individually. Calculation of friction pressure loss can be performed by summation or by using
combined equations. In prorating for combined equations it is necessary to also take into account the
difference in flow rate in the relevant sections. Examples of combined equations (for division of flow with
aT-piece) where the main line has a diameter "b" and the branched line a diameter "a", are
2
62544(/W
· b b kPa
5
pdb
Individual restrictions
In the case of restrictions in which division of flow occurs and/or where there is a change in diameter, the
literature source forK-data will specify the base for the relevant p,. calculation. In certain applications it
may be convenient to make use of the other p,. value; such evaluations require transformation of Kvalues.
The principle is prorating; examples are:
33
Diagrams
In older literature diagrams are often used to do friction pressure drop calculations. Evaluations are based
on certain equivalent lengths like 1 m (for AP, ...) and 100ft (for LiP 100}. The friction pressure loss of a
system can then be calculated as follows:
Le
:::::; MlOO X-~
100
Safety factors
Calculations of friction pressure loss are subject to various inaccuracies such as approximated correlations
for friction factors, average values for LID and K and also changes in relative roughness with time.
Calculated values may be either too low or too high. Applications where these inaccuracies may cause
operating problems, e.g. the specification of a fluid mover which will not be able to deliver the required
flow rate, must be identified; the use of suitable safety factors is recommended for such systems. A
typical safety factor for non-complex flow systems is F = 1.2 :
Special equations are available to calculate the friction pressure loss through control valves (LiPcvl· This
will be dealt with in section 10.
4.4 EQUIPMENT
In practice AP,Q for various types of equipment are reported as part of design results and will be
associated with a certain flow rate. It is often necessary to calculate LiPEQ at other flow rates.
34
This is done by prorating:
2
.dPEQ a W
.. L1PEQ,2 ~ (W/WJ 2 .;JPEQ,J
Reported L1PEQ values will already include exit and entrance pressure drops. Exit and entrance pressure
drops of such equipment must not again be incorporated into L1P1 calculations!
Elevations on a plant are normally given relative to a common reference elevation. In equipment where
fluid levels may vary, the most conservative level is used in calculations. Equations for calculation of L1P,"
are the following :
..... (4.18)
L1PEL for gases and vapours are normally considered as small (low density) and omitted in evaluations.
L1PEL for liquids can be large; if L1PEL is large and negative, it may be possible to obtain the desired flow
rate by means of gravity flow- no pump is needed.
..... (4.20)
In situations where one of the two reference pressures is not known, it may be calculated by application
of the MEB. In the case of gases or vapours, the calculation of an unknown P 1 will require a trial and error
approach. If L1PEP (terminal) is large and negative, it may be possible to obtain the desired flow rate
without the use of a fluid mover.
35
it will be theoretically possible to obtain the associated flow rate in the absence of a fluid mover; it may
however require impractically large pipe diameters.
with the following values for the kinetic energy correction factor:
Liquids are incompressible and densities are fixed. In the cases of gases and vapours for systems which
can be approximated as incompressible, calculations may be based on p 1 for both end points. In the case
of compressible flow the density in the energy equation:
dP
- + gdz + u.KE 11 du + oF = oW ..... (3.16)
p
should be integrated using an appropriate equation of state to describe the relation between density and
pressure. Special equations will be derived for compressible flow in section 5.
LIPKE is often relatively small in relation to other terms in the MEB; in some applications it is convenient
to approximate LiPKE as zero in initial calculations; this simplifies trial and error calculations; the
assumption is checked in later calculations. Linear velocities of fluids in processing units with relatively
large diameters are taken as approximately zero; if such a unit is one of the reference points, the
associated kinetic energy is zero. It is important to realise that LiPKE is applicable between the two
reference points only; all kinetic energy changes in between are irrelevant.
Two types of pipe exits are encountered in the case of liquids - below the liquid level and above the
liquid level:
36
Below:
kPa since u2 = 0
Above:
The energy supplied to a system by a fluid mover is represented by the LIP. term in the MEB. The LIP.
of a fluid mover is a function of flow rate. The relation between LIP. and flow rate is a unique
characteristic of a fluid machine and this information is supplied by the manufacturer in the form of a
graph, usually as LIZ. plotted against volumetric flow rate, V in m3/h.
In the design of new systems, the required LIP. is calculated using the MEB:
..... (3.20)
The terms on the left hand side of the MEB can be plotted against flow rate and such a curve is called the
system curve. The intersection of the system curve and characteristic pump curve is known as the
operating point and indicates the flow rate that will achieved. This is illustrated in figure 4.2.
37
----J~ump curve
Pressure head -~--
ljl__
Flow rate
Figure 4.2 Pump curve and system curve
4_9 LITERATURE
1_ J. B. Franzini, Fluid Mechanics with Engineering Applications, ninth ed. Belfast: McGraw-Hill,
1997_
2_ L F_ Moody, "Friction Factors for Pipe Flow," Transactions of the American Society of
Mechanical Engineers, vol. 66, PP- 671-684, 1944.
3. S. E. Haaland, "Simple and explicit formulas for the friction factor in turbulent pipe flow,"
Journal of Fluids Engineering, vol. 105, 1983.
4. Hooper, W B, "The two-K method predicts head losses in pipe fittings", Chemical Engineering,
96, Aug 24, 1981.
38
5. COMPRESSIBLE FLOW
5.1 INTRODUCTION
One-dimensional gas flows through nozzles, orifices and in pipelines are the most important applications
of compressible flows in chemical processing. With compressible fluids density and hence velocity may
vary considerably in a pipeline. In engineering applications liquids are considered to be incompressible.
Although the flow of gases and vapours are always compressible, for design purposes it is considered
incompressible if
..... (5.1)
where LIPsrv is calculated with the fluid density at the upstream reference point. Otherwise gas and
vapour flow in piping systems must be treated as compressible.
Isothermal or adiabatic conditions are assumed in the calculation of compressible pipe flows. Most real
situations are polytropic which increases the complexity of calculations tremendously. The isothermal
and adiabatic models for pipe flow fortunately provide bounds for the range of real behaviour and in
many cases the two models provide similar results.
In the case of flow through nozzles the flow is assumed to be adiabatic and reversible, or isentropic.
Ideal gas behaviour is assumed in the derivation of all the models. To consider non-ideal gases the
compressibility factor, z, can be included when deriving the equations.
In older literature graphical solutions for compressible flows were given, examples are the method of
Lobo, Friend and Skaperdas for isothermal flow and the Crane method for adiabatic flow. Graphical
solutions were very handy since the models are complicated to solve by hand. With the development
of computer software the emphasis in literature has now shifted to algorithms that can easily be
translated to computer code or implemented on a spreadsheet. The subsequent sections will therefore
concentrate on the fundamental equations and the necessary algorithms to solve them.
The necessary thermodynamic principles used in the derivation of the various models are only
summarised here:
cp I
R ) CP> Cv' k=-
C > P > T' v=-
'v p
39
• Ideal gas:
Pv ~ nRT
p ~-
MP
RT
Pv n ~ constant
• Isothermal process :
!1T ~ 0, 11 ~ 1
• Adiabatic process :
q ~ 0
11s ~ 0, 11 ~ k
• Definition of enthalpy:
h ~ 11 + pv
h ~ 11 + RT for ideal gas
l
cv ~ ~;) v cP ~ l dh)
dT p
40
• Relations between specific heats for ideal gases:
cp - c" = R
C = kR
p k-i
R
c =-
" k-i
II
Ma = - ..... (5.2)
c
where c is the sonic velocity. This is the velocity at which a pressure wave will travel through a
compressible fluid. The sonic velocity is calculated with the following equations for an ideal gas:
..... (5.3)
ForMa< 1 the flow is termed subsonic, sonic flow occurs if Ma = 1 and supersonic flow if Ma > 1. Sonic
flow causes shock waves and vibrations in a system and should be avoided. Sections of possible sonic
flow should be identified and tested. Such sections include small diameters and areas of low density,
since u ~ 1/pd.
The isothermal model is used for flow in long, uninsulated pipelines. Consider the energy balance
applied between two points within a single diameter pipeline without a fluid mover:
dP
- + gdz + UKE II d11 + oF = 0 ..... (5.4)
p
41
Equation (5.4) now becomes:
dP j'dL u 2
= udu + ---- ..... (5.5)
p D 2
At normal pressures the viscosity of a gas or vapour is only a function of temperature and therefore the
Reynolds number as well as the friction factor can be considered constant in isothermal flow.
MP
p = - ..... (5.6)
RT
II =
w
..... (5.7)
pA
WRT
II--- ..... (5.8)
PAM
2 _ 2 f 1dL
- - dP - -d11 + - - ..... (5.9)
pu2 11 D
p2 11 2 L,
2 f PdP f !;d11 = + r~ dL ..... (5.11)
P1 u1 L,
42
..... (5.12)
Substitution of the ideal gas law for density into the above equation yields:
MP 1 MP 2
--II = - - 1 12
RT I RT
Substitution of the above result into equation (5.12) gives the isothermal model:
L ----
The adiabatic model is most appropriate for shorter, insulated pipelines. The derivation starts with
equation (3.8):
..... (3.8)
The equation is applied between two points in a pipeline without the incorporation of a fluid mover. The
elevation term is neglected since the fluid is a gas and q~O since the flow is adiabatic:
43
II 2
l
+- ..... (5.14)
2
u2 2 - ut 2 = 2(h t - h2) = M p 2c
(Tt - T2 ) ..... (5.15)
where the right hand side of the equation is divided by molar mass to convert the units of Cp from
kJ/kmol. K to kJ/kg. K. Substitution of
T = PvM
R
and
kR
cp =
(k-1)
Substitution of
11 = -
w
pA
2k
+ --(Pv) = C ..... (5.17)
(k-1) 2 2
where C is a constant evaluated from known conditions at point 1. From the above relation the
following can be derived:
44
2
P =
k-1( W
2k Cp - pA
)
2
2pdP = k- 1 [ __!!'____
2 ( P22 - P2)
1
!
1
2k p2A2
1
2
dP fdL u 2
+ udu + - - - = 0 ..... (5.5)
p D 2
w) 2 ( w)2 dL=O
!'DA
pdP+ (A du+ ..... (5.19)
2
45
Since pAu =constant from continuity, it can be shown thatpdu = -udp. Substitution of this into equation
(5.19) gives:
(w) -dp w) dL -
2 2
pdP - - !' ( -
+ - _ 0 ..... (5.20)
A p 2/J A
Integrating and assuming that f' is constant (this is not really true for adiabatic flow) :
2 2 2
J pdP = ( W) p 2 L I (
A In"P;" - f D 2 A W) 1
..... (5.21)
I
(Aw) \/2p ( w) w)
2 2
_ !'.!:.. !._
lJ 2 A
= !:.::i(
4k A
( p;
p~
_
1
2( w) \n .':.!_ - !:.::i( w) P; - I 2 (
A p2 2k A P~
- k- 1 ( p; - I + 2In p2 ]
2k p; p1
+ k-1. (1-p;]-
2k 2
p1(~)
W
2
( 2- 2)
p2 Pt
Pt p,
.. j'!:... = k+ lin P2
D k p,
k-1 2
+ --2 (pl -p2) + -
2kp2
2 P, ( A
p,
w r 2 2
(p2 - p,)
.. !'.!:...
D
= k+ 1 In P2
k Pt
+ ( p~ ~ p; l ( k-1 ( A
2k + PtPt W
r)
46
Which leads to the adiabatic pipe flow model:
[ ---
D
:. .fL = k+1 lnP2 + ( 1
k p1
In both isothermal and adiabatic flow in a pipe of constant diameter there is a limiting pipe length at
which flow choking takes place. This happens because as P 2 decreases along the pipe, p2 decreases
and since pu = constant, u increases. Since it is physically impossible for P 2 to drop to zero, there is a
choking of the flow that limits the mass flow rate. For isothermal flow this occurs at
1
Ma =
Iii
and for adiabatic flow at
Ma = 1
It is possible through successive calculations to plot a curve such as shown in Figure 2 for any assumed
flow and initial conditions where P 2 represents any pressure along the pipe at any distance x2 . In the
case of isothermal flow equation (5.13) is only valid for
1
Ma<-
/1(
whereas equation (5.22) for adiabatic flow apply equally well for supersonic flow.
Adiabatic flow with friction is also termed Fan no Flow. If a gas entering a duct is flowing at subsonic
velocity, friction will have the effect of accelerating the flow so that sonic velocity is approached;
likewise, if the flow at the entrance is supersonic, the gas will be decelerated, also approaching Mach
1. In each case, when Mach 1 is reached choking of the flow occurs.
Figure 1 shows the conditions along the pipe length for isothermal as well as adiabatic flows.
47
p
u
T
I
I
I
1
Distance along pipe
48
5.8.2 APPLICATION OF THE COMPRESSIBLE PIPE FLOW MODELS
The equations are only applicable between two reference points inside a single diameter pipe. In the
case of piping systems, evaluations must be done incrementally for different diameters. Frictional losses
due to pipeline fittings which do not significantly reduce the pipe cross-sectional area may be added to
the velocity head term, ]LID, otherwise incremental evaluation of the sections upstream and
downstream of the restriction is necessary.
The MEB is applied between two points within a single diameter pipeline:
/';.PEP + Mf,K = 0
p2 - PI + Mf,K = 0
:. Mf,K = PI - p2
Either P 1 or P2 will be known, the unknown will be calculated using one of the models.
Consider for illustration purposes the calculation of the storage tank pressure (1'2) for the flow system
shown in Figure 2. 1'1 is known.
The first step will be calculation of 111}; 1 _2 and then checking for compressible flow according to !JPsr,IP1
> 0, 1. Assume the control indicates compressible flow. Incremental evaluation is necessary:
49
Normally, for the pipe entranceLIPSTI/1'1 « 0,1 and LIP1andLIPKE may be calculated as for incompressible
flow; confirm after the evaluation.
LJPJ,K
P;- LJPJ.K
It should be noted that in this evaluation LIPSTI/P, may be < 0,1; calculation of Pj will then not require a
compressible analysis.
-LJPEP
Because L1P1
p2
All components over which changes in diameter and/or flow rate occur must be analysed incrementally!
ISOTHERMAL MODEL
..... (5.13)
• Calculation of APJ.K
Either P1 or P2 will be known, the unknown will be calculated using a trial and error procedure or the
"solver" function of a spreadsheet. The term on the right hand side of the equation can be neglected as
a first approximation. A typical algorithm for this calculation is given in appendix D1.
In the case of a flow rate calculation a trial and error procedure is necessary to find the friction factor.
The von Karman equation can be used as an initial estimate for the friction factor. An algorithm for this
calculation is given in appendix D2.
50
AD/ABA TIC MODEL
..... (5.22)
• Calculation of APr,K
Either P 1 or P2 will be known, the equation is in terms of p 1 and P2 and therefore the ideal gas law must
be used to calculate the relevant pressures. Since temperature is not constant, T2 must also be
calculated to find P2 from p 2 . An algorithm is given in appendix 03.
In the case of a flow rate calculation a trial and error procedure is necessary to find the friction factor.
The von Karman equation can be used as an initial estimate for the friction factor. An algorithm for this
calculation is given in appendix 04.
This theory is highly relevant to the design of relief valves or bursting discs which are often incorporated
into pressurised systems in order to protect equipment and personnel from the dangers which may arise
if equipment is subjected to pressures in excess to design values.
-------
----------------
---------
Throat
Inlet
51
Isentropic conditions is assumed since frictional loss is neglected (reversible) and the area of heat
transfer is small (adiabatic).
v :::: v
p
- 2
)-Ilk ..... (5.23)
2 I( p
I
u2-u2
2 I
1/ 2
2
112
2
- ut 2 ~ -( I - ( -p 2 ) (k-IYk)
-p I -2k ..... (5.24)
Pt k-1 PI
..... (5.25)
The velocity of approach (u 1) can be considered neglible compared to the outlet velocity:
52
..... (5.26)
A2 ~
w
makes it possible to find the required nozzle throat area for the pressure to be reduced to P 2•
Noting that
The velocity at the throat, u2 , depends on the ration P/P2- If there is a large enough pressure differential
betweenP1 and the back pressure Pb, sonic velocity will occur in the nozzle throat. With further increase
in the pressure differential, the flow rate will increase {due to the density increase) but the velocity at the
throat will remain sonic.
During subsonic flow, P2 ~ P• ~ back pressure, but if the flow in the throat is sonic 1'2 :> P•.
(;~r-1)/k k+l
2
53
Equation (5.28) is the critical back pressure for sonic flow:
p crit
2
p1 ..... (5.28)
..... (5.29)
A converging nozzle is a very handy device for the measurement of flow rate. Flow rate can be
calculated as follows:
P2 _3/!_(( P 1
1) (k- )'k _
1)
p2 k-1 p2
kp
2_. ((p1)(k- 1)lk- 1) ..... (5.30)
W = A2 k-1 2p2 p
2
and because
p 1v1k = p 2 v2k
54
Substitution of the above expression into equation (5.30) yields:
1
W = A2 2_.kp
k-1 1p1
((p2)2/k-
p
(p2)k;
p
) ..... (5.31)
I 1
The maximum flow rate occurs at sonic flow in the throat, therefore substitution of equation (5.28) into
(5.31) yields:
2 )(k+1)1(k-l)
W =A Pp -
max 2 1 1 k+
. l
or '
Note that the square root expression on the right hand side of equation (5.32) only depends on the
properties of the gas. By measuring the pressure and temperature in the tank, the flow rate through
the nozzle can be calculated.
At the point where 1'2 I P 1 reaches the value of PjP 1 ,·rir, the flow in the nozzle throat is sonic. As P 1 is
increased beyond the threshold point, P jl' 1 maintains the value of P JP 1 crir and u remains sonic.
However, W increases directly with 1'1 , as shown in equation (5.32).
The flow through a converging-diverging nozzle is shown in figure 4. If a diverging section is placed
after a converging nozzle, it is possible to attain supersonic velocities in the diverging section if sonic
flow exists in the throat. The gas will continue to expand in the diverging section to lower pressures and
55
the velocity will continue to increase. If the velocity at the throat is not sonic, the gas will behave in the
same manner as a liquid : it will accelerate in the region up to the throat and decelerate in the diverging
region. This is shown by the dashed lines ABO in figure 4.
Suppose the back pressure P, in figure 4 is reduced gradually while P 1 remains constant. Then P1 = P,
and the pressure at the throat decreases while the velocity at the throat increases until the limiting sonic
velocity is reached. The pressure plot is ACE. If the back pressure is now further reduced to H, the
pressure plot is ACFGH; the jump from F toG is a pressure shock, or a normal shock wave (normal to
the approaching flow), which is analogous to the hydraulic jump, or standing wave, often seen in open
channels conveying water. Through the shock wave the velocity is reduced abruptly from supersonic
to subsonic, while at the same time the pressure jumps as shown by the lines FG, F'G' and F"G". The
flow through the shock wave is not isentropic, since part of the kinetic energy is converted to heat.
Further reduction of the back pressure causes the shock wave to move further downstream until at some
given value H"' the shock wave is located at the downstream end of the nozzle. If P, is lowered below
the level of H"' the shock wave occurs in the flow field downstream of the nozzle exit. Such flow fields
are either two or three dimensional and cannot be described by the foregoing one-dimensional
equations.
56
If the back pressure is lowered to H"", the flow will proceed isentropically to supersonic throughout the
entire region downstream from the throat, the velocity will increase continuously from 1 to its maximum
value at 3 and the pressure will drop continuously from 1 to 3. As long asP, is above H"" then P 3 ~ P•
; but if P, drops below H"" then P 3 > P • and supersonic flow occurs through the entire length of the
divergent portion of the nozzle.
If the back pressure is above E, the flow rate through the nozzle is given by equation (5.31). In this
instance theP2 of equation (5.31) must be replaced by the P , of Figure 3. If the back pressure is below
E, critical pressure, as defined by equation (5.29), will occur at the throat and the flow rate will be given
by equation (5.32).
If P 1 is increased, the sonic velocity may be shown to remain unaltered, but since the density of the gas
is increased, the rate of discharge will be greater. The converging nozzle and the converging-diverging
nozzle are alike insofar as discharge capacity is concerned. The only difference is that with the
converging-diverging nozzle, a supersonic velocity may be attained at discharge from the device, while
with the converging nozzle, the sonic velocity is the maximum value possible.
5.10 LITERATURE
1. R. K. Sinnott, Coulson and Richardson's Chemical Engineering, Design, vol. 6, 2 ed: Pergamon,
1993.
2. J. B. Franzini, Fluid Mechanics with Engineering Applications, ninth ed. Belfast: McGraw-
Hill, 1997.
3. R. H. Perry, Green, W.G., Perry's Chemical Engineers' Handbook, 7th ed: McGraw-Hill, 1997.
57
6. NON-NEWTONIAN FLOW
Gases and simple low molecular weight liquids are all Newtonian and viscosity may be treated
as constant unless there are significant variations in pressure and temperature.
Fluids which do not adhere to Newton's viscosity law are classified as non-Newtonian. Examples
are colloidal suspensions, emulsions like certain paints, sewerage sludge, melted polymers and
melted metals. Non-Newtonian fluids deviate in different ways from Newton's viscosity law.
Viscosities are functions, not only of temperature, but also of shear stress and shear rate. In
evaluations use is made of so called apparent viscosities.
Non-Newtonian flow are much more likely to be laminar due to high apparent viscosities
compared to the viscosities of simple Newtonian fluids. In order to predict the transition from
laminar to turbulent flow it is necessary to define a modified Reynolds number. The transition
from laminar to turbulent flow is not always sharp as in the case of Newtonian flow.
The terms LiP,, LiPEv L!PEP, L1Pc"" L!PEQ and LiPKE of the MEB are calculated in the usual way.
Special correlations were developed for the calculation of L1P1
6.1 RHEOLOGY
Rheology is the science concerned with the flow of both Newtonian and non-Newtonian fluids.
A Newtonian fluid at a given temperature and pressure has constant viscosity which does not
depend on shear rate and obeys Newton's viscosity equation :
T =
ou
J.t- ..... (6.1)
0)'
The apparent viscosities of non-Newtonian fluids may depend on the rate they are sheared and
on their previous shear history. At any position or time in the fluid the apparent viscosity is
defined as the ratio of the shear stress to the shear rate at that point:
T
J.l =-- ..... (6.2)
a 8u/8y
When the apparent viscosity is a function of the shear rate, the behaviour is said to be shear-
dependent; when it is a function of the duration of shearing at any time it is said to be lime-
dependent. Any shear-dependent fluid must to some extent be time-dependent because the
apparent viscosity does not change instantaneously, in many cases the effect of time-
dependence is negligible.
Typical forms of curve of shear stress versus shear rate are shown in figure 6.1. Such a plot is
known as a rheogram since it represents the rheological properties of a fluid.
58
/
Newtonian
Shear-thickenin ~""""""Bingham-plastic
_"/shear-thinning
';;:..-:..----
,#/>;;;:::>' '
ouloy
Figure 6.1 Shear stress versus shear rate
A general plot of apparent viscosity versus shear rate is shown in figure 6.2. This plot describes
the behaviour of dispersions, emulsions, polymer solutions or slurries in general. At low enough
shear rates the viscosity is constant and relatively high (Newtonian behaviour). As the shear rate
increases the viscosity begins to fall (shear-thinning). Eventually the curve becomes a straight
line when plotted on log-log axes (power-law region). At even higher shear rates the viscosity
usually begins levelling out, falling towards a constant level.
105
-,~~~
1000
~
10
0.1 ~ ~
-----
59
Two exceptions to the general behaviour described by figure 6.2 are possible. First the existence
of a yield stress (Bingham plastic fluids) and secondly the appearance of shear thickening
(dilatancy) at the high end of the curve, both shown in figure 6.3.
Time dependency is another exception to the general viscosity curve. The behaviour described
so far relates to steady state behaviour. Some materials take a long time to achieve steady state,
and during the unsteady state period they can either show a continual decrease (thixotropy) or
an increase (rheopexy) in viscosity when sheared at constant shear rate/stress.
Yield stress
107
10'
1000
10
0.1
_j _ __l__ _)___ _[__
Apart from typical viscous behaviour described above some liquids also show the elastic
response usually associated with solids. Materials which behave like this include concentrated
solutions of high molecular weight polymers, shower gels, shampoos and polymer melts.
The ideal elastic solid obeys Hooke's law in which the relation between distortion and stress is:
where G is Young's modulus and dxldy is the ratio of the shear displacement of two elements to
their distance apart. Materials that exhibit some properties of both a solid and a liquid are termed
viscoelastic.
60
The viscosity function shown in figure 6.2 is well described by the Cross model:
..... (6.4)
The Cross model can characterise the complete flow curve if the fluid does not show a yield
stress of shear thickening. For the higher shear rates of more interest to the chemical engineer,
equation (6.4) simplifies to:
..... (6.5)
A simple redefinition of some of the terms in equation (6.5) allows a rearrangement to give the
Sisko model:
..... (6.6)
If the extrapolated viscosity at infinite shear rate is negligible compared to the viscosity at the
shear rate of interest, the Sisko model reduces to the well-known power-law model:
du n-1
f.l = k- ..... (6.7)
a dy
du"
1 = k- ..... (6.8)
dy
n> I, f.la increases with increase in shear rate and shear thickening behaviour is described
n<l, f.la decreases with increase in shear rate and shear thinning behaviour is described
n~1, f.la is constant and equal to the Newton's viscosity of the fluid.
..... (6.9)
61
..... (6.10)
where T0 is the yield stress and !lP is the plastic viscosity. This equation is known as the Bingham
equation.
Some materials give more complex behaviour and the plot of shear stress against shear rate
approximates to a curve, rather than a straight line with an intercept T0 on the shear stress axis.
The following equation, known as the generalised Bingham equation or Herschei-Bulkleyequation
can then be used:
..... (6.11)
Bingham shear thickening or Bingham shear thinning behaviour can be described using equation
(6.11).
NEWTONIAN FLUID
Hagen-Poiseulle law:
..... (6.12)
In order to make use of the Darcy-Weisbach equation, which is always valid, the following relation
for the Darcy friction factor is derived:
POWER-LAW FLUID
In this case the flow curve equation is more complicated, but still simple enough to derive
analytical equations. Relation between flow rate and pressure drop :
62
APf ~ ( 6nn+2)" 4 k L u" JY< n+ll ..... (6.14)
Using the Darcy-Weisbach equation the following relation for the friction factor is derived:
For a Newtonian fluid the friction factor is a function of Reynolds number. In the case of non-
Newtonian flow the Reynolds number changes with shear rate since it is a function of the
viscosity. It is therefore difficult to define an appropriate Reynolds number. Metzner and Reed
(1955) defined a Reynolds number Re,m for a power-law fluid so that it is related to the friction
factor in the same way as for a Newtonian fluid. It is derived by substituting equation (6.15) into
equation (6.13). The following expression is then found after simplification:
2
Re 1111 _
- 8 (
-11-) "pu -"D" ..... (6.17)
' 6n+2 k
The transition value is approximately the same as for Newtonian flow, although streamline flow
may in some cases persist to somewhat higher values. Setting n~l in equation (6.17) leads to
the standard definition of the Reynolds number.
The effect of the power law index on the velocity profile is that it is flatter for a shear thinning fluid
(n < 1) compared to the Newtonian parabolic profile and sharper for a shear thickening fluid (n
> 1).
In the case of a Bingham-plastic fluid the cross-section of flow in a pipe can be considered in two
parts:
1) A central unsheared core in which the fluid is all travelling at the centre-line velocity.
2) An annular region separating the core from the wall over which the whole of the velocity
profile is concentrated.
63
The relation between flow rate and pressure drop is derived by considering the two parts of the
flow mentioned above separately and then adding them. The relation is much more complicated
than for a power-law fluid:
For a Newtonian fluid X and ~0 is zero. Equation (6.18) is sometimes referred to as Buckingham's
equation.
Fluids whose behaviour can be approximated by the power-law or Bingham plastic models are
special cases. The rheology are frequently very complex and simple algebraic equations cannot
be fitted to the flow curves. A general method for time-independent fluids in fully developed flow
is given here. A general model with parameters that can be measured for any fluid is used.
The following general relation between pressure drop and flow rate can be derived:
l!.P
!
~ 4k'L (
D
8u)
D
n' ..... (6.19)
where k' and n' are generalised rheological parameters. These parameters are widely used in the
literature on rheology. Values for the generalised parameters for various fluids are given in
appendix E.
..... (6.20)
An equation for the friction factor can be found, similar to equation (6.15):
1
~ !"_) "
1
f' 8k ( ..... (6.21)
. pu2 D
pu 2 -n 'nn I
ReMR ~ ..... (6.22)
gn 1-lk I
64
6.3 TURBULENT FLOW CORRELATIONS FORLIP1
In the case of turbulent flow the relations for pressure drop can not be derived from first
principles, empirical relations are used. Many correlations have been published for different types
of non-Newtonian fluids.
The following relation proposed by Yoo (1974) gives values for the friction factor accurate to
within ±1 0%. The friction factor is expressed in terms of the Metzner and Reed generalised
Reynolds number and the power law index, n:
..... (6.23)
The above equation should be used with caution, particularly if the fluid exhibits any plastic
properties.
6.4 LITERATURE
1. H. Barnes, "Rheology for the chemical engineer," The Chemical Engineer, June 24, 1993.
2. D. C. H. Cheng, "Pipeline design for non-Newtonian flow," The Chemical Engineer, vol.
525, 1975.
65
7. MULTIPHASE FLOW
The complexity of multiphase flow is so great that design methods depend on an analysis of the
behaviour of such systems in practice and, only to a limited extent, on theoretical predictions.
For all multiphase flows it is important to understand the nature of the interactions between
phases and how these influence the flow patterns- the ways that the two phases are distributed
over the cross-section of the pipe. Pressure drop will depend on the flow pattern as well as the
relative velocity of the phases - slip velocity. This slip velocity will influence the hold-up, the
fraction of the pipe volume which is occupied by a particular phase. In the flow of a two-
component mixture, the hold-up of a component will differ from that in the mixture discharged
at the end of the pipe because, as a result of slip of the phases relative to one another, their
residence times in the pipe will not be the same.
Only liquid-liquid and gas(vapour)-liquid multiphase flows will be considered in this course.
Two phase liquid-liquid flow is encountered when two liquids which are relatively immiscible flow
in the same piping system. The method of Woods and Dukler is suitable for liquid-liquid flow.
It is less accurate for vapour-liquid flow but is however convenient for certain estimations in the
latter case. The same equations and methods as for single phase flow are used but with
weighted average values for physical properties, volume flow rate and linear velocity:
I = L xi ..... (7.1)
Prp pi
I:~
PrP = ..... (7.2)
'LWi /pi
I Lxi
= ..... (7.3)
f1rp lli
66
~w
vTP ~ --'
p ..... (7.4)
TP
~wi
u ~ -- ..... (7.5)
TP p A
TP
..... (7.6)
Two phase vapour-liquid flow is encountered in steam-condensate lines, reboiler lines and lines
from partial condensers. Three phase flow, with two liquid phases and a vapour phase, is
relatively common in the petroleum industry.
The MEB in its general format is applicable, special methods are used to predict AP1 . A fluid
mover may be part of a liquid-liquid two phase system but will not be mounted in a vapour-liquid
line; if a fluid mover is required for such a system, it will be placed in a line section where the
flow is still single phase. Similarly a control valve will also not be used in the two phase flow
section of a pipe system.
Vertical and horizontal flow patterns differ, in the case of vertical flow axial symmetry exists.
Principal characteristics of the flow patterns are described in figure 7.1.
The regions over which the different types of flow occur are conveniently shown on a flow
pattern map in which a function of the gas flowrate is plotted against a function of the liquid
flowrate and boundary lines are drawn. The distinction between the flow patterns are not clear
cut and several workers have produced their own flow maps. Figure 7.2 shows a flow pattern
map prepared by Chhabra and Richardson.
67
c .-
.. Bubble flow
~
~~d
Slug flow
Slug flow Annular flow
t====l]
Annular flow
Upward vertical flow
Mist/spray flow
Liquid Vapour
Bubble flow Gas bubbles dispersed throughout the liquid 1.5-5 0.3-3
Plug flow Plugs of gas in liquid phase 0.6 <1.0
Stratified flow Layer of liquid with layer of gas above <0.15 0.6-3
Wavy flow As stratified but with a wavy interface due to <0.3 >5
higher velocities
68
Slug flow causes unstable flow conditions with vibrations and is highly undesirable. In the design
of piping systems for two phase vapour-liquid flow, a check for possible slug flow must always
be done. If slug flow prevails, other pipe diameters must be considered. It is desirable to design
1.._,:
so that annular flow still persists at loadings down to 50% of the normal flow rates. Mist flow
should also be avoided since once mist flow is reached there is virtually no way of returning to
any other flow regime.
The most widely used method is that proposed by Lockhart and Martinelli and later modified by
Chisholm.
The method requires the evaluation of friction pressure drop for the liquid phase as if it were the
only fluid in the system (LIPt.Ll and also the evaluation of friction pressure drop for the vapour
phase as if it were the only fluid in the system (LIPt.al· The two-phase pressure drop (APt.rP) is
taken as LIPJ.L or LIPr.a multiplied by some factor <D0 2 or <DL2 where
69
f>,.pf,TP
<1>2 ..... (7.7)
G
MG
Mf,TP 2
= <l>L ..... (7.8)
D.PL
<I>G2 and <I>~.2 are given graphically as functions of the hold-up parameter X, where:
..... (7.9)
Reynolds numbers for distinguishing between turbulent and viscous flow are calculated as if
only the fluid of the case in question flowed in the pipe.
Strictly speaking the method is only applicable for isothermal, non-flashing, incompressible flow
in horizontal pipe. It was shown that the method renders relatively low values for annular flow
and large values for stratified, wave and slug flow.
c 1
+- + ..... (7.10)
x x2
where c has a value of 20 for turbulent/turbulent flow, 10 for turbulent liquid/laminar gas, 12 for
laminar liquid/turbulent gas and 5 for laminar/laminar flow.
7.3 LITERATURE
1. R. Kern, Piping design for two-phase flow, Chem Eng, 145, June 25, 1975.
2. W.W. Blackwell, Calculating two-phase pressure drop, Chem Eng, 121, Sept 7, 1981.
70
8. OTHER TYPES OF FLOW
8.1.1 GRADUALAT
Gradual temperature changes occur when hot or cold pipelines are not effectively insulated and
in tubes of shell and tube heat exchangers. The normal MEB is used with the following
modifications:
2
62544fL W q> kPa ..... (8.1)
pds
rp ~ (fliflJ"'
For laminar flow m = 0,25 and for turbulent flow m = 0.14. Subscript)w refers to wall conditions.
The evaluation of mass and wall temperatures are dealt with in heat transfer literature.
LiP,, AP8 Q. LiPEP and APcv are calculated as for isothermal systems. Strictly speaking LiPEL must
be calculated by integration. In most applications variation in density is ignored and the
calculation is based on the density at the upstream reference point. LIPKE is calculated as A(X)
with properties and other variables evaluated at the conditions of the two relevant reference
points.
8.1.2 STEPWISE AT
Sudden changes in fluid temperature are normally caused by heat exchangers which form part
of the piping system. LIP,, LiPEpLiPcv and LiPKE are calculated as discussed in 2.4.1.2. In
calculations of LiP!!Q for heat exchangers non-isothermal flow is taken into consideration. Strictly
speaking, in prorating for other flow rates, densities should not be taken as constant; constant
densities are however often acceptable for design calculations. For the evaluation of L1P1 and
APEL the system is subdivided into sections of isothermal flow; then
71
iJP1 ~ I:iJP!.,
iJPEL ~ 1L1PEL,i
8.1.3 LITERATURE
1. K.N. Murty, Assessing effects of temperature and flow rate of an incompressible fluid,
Chem Eng, 101, Jul25, 1983.
Pulsating flow is the result of fluid being moved by reciprocating fluid movers. See figure 8.1.
Many variations of cylinder, piston, valve and drive mechanisms have been designed. In fluid
dynamics it is necessary to distinguish between single acting (single piston or plunger with only
one delivery per cycle), double acting (single piston or plunger with two deliveries per cycle);
simplex (one cylinder), duplex (two cylinders), etc. From a flow dynamic point of view double
acting simplex = single acting duplex because both render two deliveries per cycle.
72
Cylinder movement in the case of crankshaft movers is simple harmonic:
ll
s,m
= Oll' cos e = Oll' cos rot
=
2 1tN rad!s
Ol
60
11s,max Oll'
us,min = 0
The continuity equation gives a correlation between linear velocity of the cylinder and linear
velocity of an incompressible fluid in the pipeline; for simplification the subscript)p for pipeline
is ignored:
The concept of average linear velocity still applies and is correlated with the production rate:
W = pV = puA
puA = Pllvll ,A,
Because 11,,, varies between zero and u '·"""' 11, and .dP f.m also vary between zero and their
maximum values. Average values are however required for application in the MEB.
GASES AND VAPOURS : For purposes of piping system design calculations it is assumed
that the compressibility of gases and vapours to a large degree absorbs the pulsating action.
Friction pressure losses are calculated as for non-pulsating systems. Where relevant, a safety
factor of 1.3 is recommended.
LIQUIDS: No absorption of the pulsating action takes place and the approximation used for
gases and vapours is not valid. An equation can be developed to correlate the average friction
pressure drop of pulsating flow with the equivalent friction pressure drop of non-pulsating flow.
Consider the fluid mover shown in figure 8.1:
73
us,m ::::: corcos8
x2
..
2
us,m = oh 2cos28 = ro 2r 2(1-sin28) = ro2r\l-.:...)
r
:. X
2
= lf/(11,~,) = lfl(u,;,) = lfl(M'f,m)
r X
L1P1 .,~ can be calculated with one of the forms of the Darcy equation:
74
2
1 wrA
fL . ~ .
2
= 1l, ' kPa
D 2000 ( uA )
2
'l wrA
= /I.Pf,ass non-puls (
"uA ' kPa
)
2
2 'l wrA
where Fp = - " ' = pulsating factor ..... (8.2)
3( uA )
Where relevant the normal safety factor of 1.2 is used. Calculation of pulsating factors is
dependent on the type of fluid mover.
u = 'l,ll ,A ,I A
2rA/f - 2rN
us = ---
60A, 60
2
2nN
2 --·r
F = .3_ ( 'l,wrA,IA) = .3_ ( 'lvwrA,IA) = rur) - -2 2
2( 2
60
P 3 II 3 Tj.,II,A,/A 3 u, 3 2rN
--
60
.. F =
p
3.3 n2 = 6.58 ..... (8.3)
75
MOVERS WITH TWO DELIVERY STROKES PER CYCLE
11
= 2 ( 2rA,N) = 4rN
' 60A s 60
..... (8.4)
It follows that for the same flow rate friction pressure loss is much less for movers with two
delivery strokes per cycle than for movers with only one delivery stroke per cycle.
Air chambers may be included in delivery lines to reduce friction pressure losses. The use of
air chambers do not eliminate pulsating actions. The following pulsating factors are
recommended in the presence of air chambers.
Steam driven movers are in general use. Steam buffers the pulsating action. Pulsating factors
as for air chamber systems are recommended.
LiPKE is calculated as Fyl(X). Li(X) is calculated with the same equations as for non-pulsating
flow. Pulsating factors are the same as for friction pressure losses.
LiPa is calculated with the MEB.
8.2.4 LITERATURE
1. T.L. Henshaw, Positive displacement pumps, Chern Eng, Sept 21, 1981.
2. J.D. Ekstrum, Sizing pulsation dampeners for reciprocating pumps, Chern Eng, 111,
Jan 12, 1981.
76
9. PIPING SYSTEM DESIGN
For pipelines it embraces design of wall thickness and optimal economic diameter. Design of
wall thickness is based on mechanical engineering principles; designs for wall thickness and
diameter are however interrelated and are normally performed by the same engineer. A trial-
and-error approach is required. A typical procedure is an estimation of diameter, selection of
an industrially available pipe (nominal and outside diameter known), design of wall thickness and
check with optimum diameter criteria.
Nominal diameters of fittings and components are taken the same as that of the associated
designed pipe section; the maximum pressure which may occur determines the class type which
fixes dimensions like wall thickness. Design of a check valve diameter however requires an own
procedure.
The wall thickness must be sufficient to prevent failures which will result if stresses in the system
exceed the yield strength of the construction material. It is normal practice to allow for a certain
thickness of material loss due to corrosion. The system is designed to withstand the maximum
pressure which may be encountered.
The maximum pressure can be calculated with the MEB; at the beginning of the design, pipe
diameters necessary for MEB analyses are not available and a trial-and-error approach is
required. Any operation which may cause momentary pressure increases, like pulsating flow
and the occurrence of water hammer must also be taken into consideration. Water hammer is
only relevant with liquids and can be caused by the sudden closing of a valve or failing of a
pump; momentary pressure increases are functions of fluid momentum.
Momentary pressure increases due to water hammer may be calculated from basic principles:
AZ = -
af::..u
-Ill
D. ..... (9.1)
g
1 d ) -0,5
=
(
p(- + -)
K tE
77
K is the compressibility modulus of the liquid (N/m 2 ) and E is the elasticity modulus of the
construction material of the pipe (N/m2 ).
The momentary pressure increases may also be obtained from tables in literature. See table 9.1.
4 to 10 120
12 to 14 110
16 to 18 100
20 90
24 85
30 80
36 75
42 to 60 70
The maximum pressure to be used for wall thickness design is the sum of the maximum pressure
according to the MEB and any momentary pressure increase which may occur.
The method of wall thickness design is also a function of the type of construction material. In the
case of brittle materials, tables are available which correlate maximum system pressure and wall
thickness.
In the case of ductile materials wall thicknesses are calculated with the following mechanical
design equation
78
The equation is derived from basic principles; thus units of the variables in the right hand side
determine the unit for tmiw After evaluation of tmin an industrially available pipe is selected for
Which I :>c I min·
9.1.2 DIAMETER
The aim is to establish the optimum economic diameter of a piping system for the required flow
rate. Under favourable conditions for LIFE/' and LIPEt it may be possible to obtain the desired
flowrate in the absence of a fluid mover. It is most important to distinguish between systems
without fluid movers and systems with fluid movers.
Such systems are possible if (LIPEP + LIPEJ is negative and relatively large. It must be large
enough to render the relevant flowrate in a piping system for which the total cost will be equal
to or less than that of a piping system with a smaller diameter but which requires a fluid mover.
The choice of a system therefore actually requires a proper economic evaluation. A shortcut
method which is often acceptable is based on linear velocities. If, for Newtonian fluids, -(LIPEP+
LIPEJ is large enough to render a pipe diameter for which the linear flow velocity is larger than
1 m/s for liquids or 5 m/s for gases or vapours, the system without the fluid mover is considered
to be more economical; if the linear velocity is less than the given values, it is indicative of the
obtained diameter being too large and that a smaller diameter system, combined with a fluid
mover, may be more economical. For liquids with high viscosities and also for non-Newtonian
fluids this rule of thumb does not apply; systems without fluid movers with u « 1 may still be more
economical.
A suitable design method is the following. Determine LIFt,,..""''' by application of the MEB.
Estimate values for j' and Le and calculate d with the Darcy equation. Base calculations on the
maximum flowrate which may need to be processed. Choose an industrially available pipeline,
design for wall thickness and perform check calculations (Re, .f, Le, APr, 11). Prorate for other
diameters - the optimum economic diameter is the smallest diameter for which
LJPf S L1Pf,availabk
Linear velocities play important roles, in design of piping systems. Low velocities may be
indicative of an uneconomic system; it may also cause undesirable sedimentation of suspended
solids and problems with crevice corrosion. High velocities may cause problems with erosion
corrosion, vibrations and the development of static charges. Rough rules of thumb for systems
without fluid movers are the following.
Piping cost increases with pipe diameter and power cost decreases. The diameter with minimum
system cost is the optimum economic diameter. Different variables are important in the
79
evaluation of the optimum economic diameter. Various correlations have been developed for
its determination (see references). A simple but reliable method is based on recommended
values for API"' (AP 100 rJ combined with criteria for linear velocities. The following design
procedure is recommended:
• Obtain the relevant API, from the literature and calculate by means of trial and error the
associated diameter with the Darcy equation. Base calculations on the flowrate
associated with the planned production rate.
• Choose an industrially available pipe diameter and design for wall thickness.
• Do check calculations (Re J, APim, u). The optimum economic diameter is that diameter
which renders values for APim and u which correlate best with relevant criteria. Prorating
simplifies calculations.
Cavitation occurs when, due to pressure drop in liquid flow systems a fraction of liquid
evaporates and a two phase system results. At higher pressure zones the vapour suddenly
condenses and shock waves are formed. They cause problems with vibrations and erosion
corrosion. In the diameter design of piping systems with pumps it is necessary to distinguish
between systems where cavitation is a threat and systems where it is not.
They are all gas and vapour systems, all delivery lines where liquids are pumped as well as
suction lines for which the operating temperature is much lower than the bubble point
temperature, provided that the liquid is not saturated with dissolved gases.
Criteria for APim and AP100ft for mild steel systems are given in appendix G1. If the piping system
cost deviates substantially from that of mild steel, the criteria must be adjusted; for instance in
the case of UNS 30400 (austenitic stainless steel often used in the chemical industry)APvalues
= 2 xAP values for mild steel are recommended. The criteria can be used for all fluid flow types.
Ludwig criteria are used for linear velocities. Examples for different systems are given in
appendix G2. In most cases a designed pipe which complies with the relevant AP criterion will
also comply to the relevant linear velocity criterion. As a general rule, if a discrepancy does
occur, the AP criterion should be considered dominant; exceptions are steam (linear velocity
criteria are better developed), certain construction material-corrosive medium systems (e.g. for
mild steel and concentrated sulphuric acid for which u must be lower than 1.5 m/s to limit erosion
corrosion) and certain two-phase vapour-liquid systems (for vertical lines urp > 6 m/s to prevent
slug forming; slug flow in horizontal lines is also not acceptable but a different criterion is used).
For non-Newtonian flow and also for liquids with high viscosities the criterion of linear velocity
loses its relevancy.
The criterion APim- 0.000164 P"P"'"m may lead to two different sizes for the suction and delivery
lines for gas and vapour systems with AP" relatively large.
80
• SYSTEMS WITH CAVITATION
When a liquid is pumped at a temperature which is close to that of its bubble point at the
prevailing system pressure, a problem with cavitation may occur in the pump. Typical pressure
changes in a centrifugal pump is shown in figure 9.1. If the pressure drops to a value lower than
that of the bubble point at the prevailing system temperature, cavitation will occur. In the case
of pumps, cavitation not only causes problems with vibrations and erosion corrosion but also with
decreasing delivery head. According to the Hydraulic Institute a limited amount of cavitation is
allowable- the decrease in delivery head must however not be more than 3%. See figure 9.2.
Entrance loss
----~-__F_r~ction
P
______ _::_-::~Increasing P
due to im7elle
I -~
::::::::::~-:._:_----..--..._
1
Turbulence, friction, to cavitation
entrance loss at van _~
--~-~·------'-----
Point of lowest P where
vaporization starts Flow rate
When the liquid is nearly saturated with dissolved gases a less harmful type of cavitation may
occur at operating temperatures much less than bubble point temperatures. At zones of reduced
pressure two phase systems also form; however when pressure increases at the delivery side,
the return to a single phase system is gradual and less serious.
The concept known as the nett positive suction head (NPSH) is used in the analyses of
cavitation problems. It relates to the pressure at the suction flange of a pump. It is necessary
to distinguish between the available NPSH and the required NPSH.
NPSHA is a property of flow in the suction line and is defined as the available pressure difference
at the suction flange to limit cavitation:
= (total pressure which, at the conditions of flow, will exist at the suction
flange) - (effective vapour pressure of the liquid at the operating
temperature) in units of liquid head.
81
Consider the pump suction line system shown in figure 9.3.
!\I
I
I !J.Z
I
I
I ~----D---------
-··-·~-···-·---5-Q
"". (9. 3)
Use is made of "effective vapour pressure" to take into account the possible presence of
dissolved gases. If dissolved gas is not relevant the effective vapour pressure = vapour
pressure. If the liquid is at its bubble point at reference point 1, then P,!f= P 1• Various methods
are in use to calculate the effective vapour pressure if dissolved gases are relevant. A rough
method proposed by Whistler is
82
p_:_l_+_P_.:.:va:!::po:::"'--',.P::.:re:::"::::"r_:_e ..... (9.4)
pe/J = - 2
NPSHR is an intrinsic property of each pump and is defined as the required pressure difference
at the suction flange to limit cavitation to 3% loss in delivery head. It is determined
experimentally as
NPSHR (total pressure required at the suction flange to limit cavitation to 3% loss
in delivery head)- (effective vapour pressure of the liquid at the operating
temperature) in units of liquid head.
NPSHR is a function of flowrate as well as the geometry and relative roughness of the pump's
suction side. The NPSHR characteristic of a pump can be obtained from commercial pump
suppliers. It is normally determined with water as fluid. Correlations exist for transforming
NPSHR, wATER to NPSH R, FLu 10; in most design applications NPSH R, WATER is also used for
other fluids; results are normally conservative.
(c) DESIGN
To limit problems with cavitation, NPSHA for the suction line must be larger than NPSHR for the
pump.
That criterion which renders the largest NPSHA (in cases where the pump characteristic is fixed)
or the smallest NPSHR (in cases where the suction line characteristics are fixed) must be applied.
If the pressure at reference point 1 and the operating temperature are fixed the following three
variables play important roles:
(1) NPSHR : If NPSHR is small, the associated NPSHA is relatively small. This implies that
!JP1 may be relatively large; thus relatively inexpensive small diameter suction pipe may
be specified; pumps with small NPSH" values are however relatively expensive.
(2) Suction line characteristics : They determine !JP1 . Lines should always be as short as
possible to minimise !JP1 . Large diameters and large NPSHR go together; small
diameters and small NPSHR go together.
83
(3) L1Z : If L1Z is negative and large, large L1P1 values and NPSHR values can be
accommodated (thus inexpensive pipelines and pump). Large negativeL1Zvalues may
however require expensive constructions.
In the design of a suction line which is subject to cavitation problems a compromise is normally
made between pipe diameter and NPSHR. Diameters are designed according to criteria which
render larger diameters than criteria for pipe with no cavitation problems. See appendix G1.
Such designed lines are considered optimum economic for the relative costs of pipeline, pump
and elevation constructions. Linear velocities in these lines will not comply to Ludwig criteria.
It remains good practice to calculate linear velocities; with certain material-medium systems (e.g.
stainless steel with water containing dissolved chlorides) linear velocities « 1 m/s are not
acceptable because of sedimentation with crevice corrosion problems which may occur. After
obtaining the optimum economic suction line diameter, NPSHR can be calculated for pump
specification purposes.
9.1.3 CHECKVALVES
For effective operation the friction pressure drop across a check valve must be larger than a
certain minimum. Check valves are designed independently. Minimum pressure drop across
check valves are given in appendix C3.
9.1.4 LITERATURE
2. J. T. Petroskry, Determining economic pipe diameters, Plant Engineering, 114, June 24,
1976.
3. G. R. Kent, Preliminary pipeline sizing, Chern Eng, 119, Sept 25, 1978.
5. M.S. Peters and K. D. Timmerhaus, Plant design and economics for chemical engineers.
9.2 ORIFICES
Various instruments are available for the measuring of flowrates. Examples are orifices, venturi
meters, pilot tubes and rotameters. Orifices are in general use when pipe diameters ~ 50 mm;
pressure drop over an orifice plate is also often used as a signal for flowrate control.
An orifice is a plate through which a hole is drilled and which is mounted in a pipe between
flanges as shown in figure 9.4. The pressure differential is measured between two pressure
taps. A typical pressure profile in the vicinity of an orifice is shown.
84
Pressure along orific:e pipe·run
I
\L ---- _ -----=M±fo=ss)=;:::=;=
_l,_T_C>r-P"
·---- -- -~ ~ -~
~ \_ Flow--
--
.---+1-
h~
_L_ __
AP1 could be calculated with the Darcy equation if values for Kr or Lr!D for the orifice were
known. Calculation of APKE would require information of the flow profile (see figure 9.4) and is
a function of the positions of the pressure taps. Different arrangements in use are corner taps,
() radius taps, line taps, flange taps and vena-contracta taps. See figure 9.5. For flange taps
positions are one inch upstream and downstream of the plate. The point of minimum pressure
(maximum kinetic energy) is known as the vena contracta.
For analyses the terms AP1 and APKE are combined. Examples of equations in use are the
following:
hw = mm water
d = diameter in mm
fJ = diameter ratio = orifice diameter/pipe inside diameter
The basic correlation of APs-rv a w> still applies.
85
l..--------M----->.'__:_N-~
I' I I
"""' "
1-
r-2Y.. pipe di.,,..t<c------- 8 pipe dia. --~------> ~
1- ~
I f- '\
0 0 ~ r\.
1- '\
f- [\.
1:
0.3 '
0.2 0.3
"
0.4 0.5
..
0.6
''
0.7
..~ 0.8
tine Taps Hatio, d 0 /d 1
86
In practice 0.25 ~ p ~ 0. 75; p = 0. 7 is popular. Cis known as the coefficient of discharge. It is
not a constant but a function of the positions of the pressure taps, p and Re. See figure 9.6.
Positions of the pressure taps and pare fixed per orifice installation. However, because C varies
significantly with Re < 30 000, orifices are less suitable for flow measurement at low Re;
calibration is required. For Re > 30 000, C is approximately constant and flowrate is a function
of hw as shown in the equations.
0.90
v
lo.'
0.80
[.6: \ ~
"·~·"".'" [%::: 1':-:- l'i
.
u
PS I'-
~
,
.i!
g
;;
f' ~ 6::::
~---
Ji 1~-<0 '."m :'"~
~~
~~
'
~-·
.. ~
' 10 4 10
' 4 10
' ' 10 • ' 10
DESIGN
Various design approaches may be followed. One suitable approach is the following. Establish
Re for the designed pipeline. If Re > 30 000, the orifice may be mounted in it; if Re < 30 000 a
special section of smaller diameter pipeline must be incorporated for orifice mounting and for
which Re > 30 000. Certain lengths of pipe upstream and downstream of the orifice must be free
of restrictions. Perry gives more detail.
=
Decide on the following : Construction material, p (/3 0. 7 is popular), plate thickness (thickness
< 0.033 pipe diameters and < 0.125 orifice opening), type of pressure-taps (each has own
advantages and disadvantages, vena contracta is popular).
Calculate LiPEQ based on the normal flowrate (h w follows from the basic equation; it however
87
includes LiPKE which is recovered downstream; LiP EQ to be used in the global MEB can be
calculated from the following relationship:
Calculate the maximum pressure drop which may be recorded by the recording instrument; this
is h., (may be specified in mm water or kPa) based on the maximum flow rate which may be
encountered; report as the maximum instrument reading.
LITERATURE
1. J. Powers, Flow meter selection guide, Chem Proc, 79, Oct, 1979.
2. A. Noor, Sizing orifice and venturi meters, Chem Eng, 97, Aug 22, 1983.
88
10. CONTROL VALVES
10.1 INTRODUCTION
Control values are primarily used for the control of flow rates. Control of flow rates may indirectly
lead to the control of variables such as level and temperature. Flow rates may also be controlled
by regulating the rotational speed of fluid movers; this method of control is not dealt with in this
course. Various types of control valves are commercially available. See examples in figure 10.1.
Operating principles are the same- a signal from a flow rate measuring instrument (typically a
pressure drop across an orifice plate) is transmitted (pneumatic or electric) to the control valve
which then takes action. Closing causes an increase of LIPcvwith an equivalent decrease of LIPSTv
and flow rate; opening causes LIPcv to decrease and LIPSTv and flow rate to increase. This course
deals with size design of control valves; Masoneilen control valves are used for illustration
purposes.
A typical mounting of a control valve is shown in figure 10.2. Two gate valves are mounted on
either side of the control valve; it is necessary for maintenance purposes. For continuation of
operation during a maintenance period a loopline with globe valve is part of the system. The
diameter of the loopline is normally taken the same as that of the main line; it may also be
designed on its own; the principle for such a design would be LIP8 n~ loopu,, s LIPrv; m,1,u,, ucuon •
Diameters of control valves are often smaller than those of the pipelines; coupling of a control
valve directly to the gate valves establishes a sudden reducer and enlarger in the system; ASA
reducers and enlargers may also be used.
89
AIR PRESSURE
DIAPHRAGM
SPRING
STEM
PARABOLIC
INNER VALVE
IaI I bl IcI
90
10.2 FLOW RATES
NORMAL FLOW RATE (W,,) : It is the flow rate associated with the planned production rate. In
the absence of any additional requirements optimum economic diameters are designed with
normal flow rates.
MAXIMUM FLOW RATE (Wm): Some processings require flow rates larger than normal flow
rates for relatively short intervals. Examples are reflux lines of distillation columns and quench
lines for exothermic reactors. If such lines incorporate fluid movers, their optimum economic
diameters are still designed for w;,; the fluid movers are designed and specified to deal with the
occasional larger flow rates. If they are operated without fluid movers, the optimum economic
diameters are designed with Wm.
DESIGN FLOW RATE (WJ) : Control valves cannot control flow effectively when they are fully
open. For effective control of maximum flows, control valves must be " 95% open with maximum
flow. Design flow rates are approximately 5% larger than maximum flow rates and are used in
control valve design to enable the effective control of maximum flow rates.
Equations correlate flow rate and friction pressure loss over control valves and are based on the
Darcy equation. A control valve coefficient Ccv is used in stead of a Lr!D or Krvalue. Literature
values for Ccv are for fully open control valves. See appendix H. Ccv values are determined
experimentally as the volume flow rate (US gpm at 60 •F) through the fully open control valve
when the friction pressure drop over the valve is 1 psi (liPcv = 1 psi).
10.3.1 LIQUIDS
where
91
LIPcv = pressure loss across valve in psi
X = valve stem position
f(x) = fraction of the total flow area of the valve (the curve of f(x) versus x is called the
inherent characteristic of the valve)
v = flow rate in gpm
In the fully open position, at the design flow rate f(x) = 1, after the valve has taken action,J(x)
changes.
Pressure drops over control valves are normally relatively large and flow is mostly compressible.
The criterion for compressibility is L1Pcv/P1 > 0.1. See figure 10.3.
1 2
Two variations of the basic control valve equation are used to deal with compressible flow:
1) The first equation is in terms of the average density across the valve:
Expansion of a gas through a valve is polytropic. For purposes of design calculations, it may be
approximated as isothermal. Strictly speaking the calculation of LIPcv requires a trial-and-error
approach. The main application of the equation in this format is for the estimation of f(x) from
LIPCl\ availab/, when an approximated value for pav is acceptable.
2) More accurate evaluations require incremental solutions. Depending on the placing of the
compressor, either P1 or P2 can be accurately calculated from a MEB. Masoneilen derived an
92
equation which eliminates trial-and-error methods for the calculation of the other pressure (P2 or
P 1 ;L1Pcv~P 1 -P2):
520M
29ZT
..... (10.3)
Evaluations of LiPcv with these equations can only be performed if the control valve is fully open,
that is for f{x) = 1.
If flow in pipelines is identified as compressible, it is necessary to check for sonic velocities in the
lines. This can be done using the critical pressure ratio for sonic flow. Refer to section 5.
By changing the shape of the plug and the seat of the valve, different relations betweenj(x) and
x can be obtained. Common flow characteristics used are linear trim valves and equal-
percentage trim valves as shown in figure 10.4.
The reason for using different control valve trims is to keep the stability of the control loop fairly
constant over a wide range of flows. Linear trim valves are used, for example, when the pressure
drop over the control valve is fairly constant and a linear relationship exists between the controlled
variable and the flow rate of the manipulated variable. Equal percentage (increasing sensitivity)
control valves are often used when the pressure drop over the control valve is not constant. This
is best illustrated using pump and system curves. The pressure drop across the control valve is
not considered as part of the system curve, the pump and system curve is plotted and the
distance between the pump and system curve then represents the LiP"" The pressure loss across
the valve can then be determined very easily for a range of flow rates. This is shown in figure
10.5. In figure 10.6 the choice between linear and increasing sensitivity valve trims are shown.
93
Linear
f(x) =x
f(x}
Pressure head
Mev
li\--~-c-c-c::CC
Systelll~u~e __ -----~-----/~
_______________
L1Psrv
'V/__ ____ _
ljl_L____ ----------------
Flow rate
Figure 10.5 Pump curve, system curve and control valve pressure loss
94
n r--.
~ ,· '
"Tl
«5.
c
~
<1> Head
~ Pump Charactorlatlc Hoad
Pump CharacktriaUc
9(J) Ooveloped Do!v•loped
"
Aoqulred "
Roqulred
0 H
::::r H
0
cr
<1>
tr
<1>
~
<1>
<1>
::J Static I I -
Syatem Curve
<
"'<1>
<
.....
Delivery. Syatom Curve
Volumetric Aowrate, Q
-
:::!.
3(/)
Desired flowrate range
Use linear trim range
Volumetric Flowrite ,
Desired flowrate
Q
Desired flowrate range
Use linear trim
Desired flowrate
range
Use equal % trim
Use equal ~{, trim
(!)
"'
H•od Hoad
Pump Char.acterletlc Pump Ch.aractedatic
OovelopOO
.,
I
Oovoloped
"
Required
Required B ~~
Sysotem Curv•
H Syatom Curve
H
Static
.....
Delivery
Stdc
,., .
Oo!lvery
Standard optimum economic diameter design principles are applicable. Several operating
requirements are also relevant. Control valve rules were developed to take both economic and
operating requirements into account.
This rule takes care of the requirement that the chosen control valve must be able to
accommodate design flow rates when it is fully open; thus it will be able to control maximum flow.
Too much control valve action is associated with less effective control. This rule prevents
operating in relatively closed positions with normal flow.
Effective control also requires thatLJPcvmust be a substantial fraction of L1P7 v; the lower limit of
these two rules takes care of this requirement. Economic analyses require that LiPcv must not be
unnecessarily large relative to L1Prv; the upper limit of these two rules takes care of this
requirement.
Except for rule 1(d), the rules must not be applied rigidly. Exceeding a limit is an indication that
a better option most likely exists and that it should be investigated.
Rules 1(d) and 1(n) are still applicable. In the absence of operating costs, the upper limits of rules
2(d) and 2(n) are less relevant. They do however, serve a handy purpose in dividing an available
L!Prv between L!Psrv (for pipeline design) and LiPcv (for control valve choice).
In the design of piping systems with control valves it is necessary to distinguish between different
systems. The main types are new systems with fluid movers, new systems without fluid movers,
new systems with fluid movers where the fluid mover supplies flow to two or more branched lines,
each with its own control valve and existing systems without control valves which must be
provided with control valves. Different design approaches may be followed. The following are
suitable:
96
10.5.1 NEW SYSTEMS WITH FLUID MOVERS
Design the pipeline on its own. Take as a first choice a control valve one size smaller. Test with
control valve rules. If the rules are satisfied, the system qualifies as optimum economic. If not,
the results will indicate whether a line size valve or a two diameters smaller valve should be
considered next; this system will most likely qualify; control valves more than two line sizes
smaller are rare exceptions; control valves larger than line size are never specified. If the initial
pipeline design indicated the possibility of a second line size which may also qualify as optimum
economic, it may also be investigated. Different procedures are recommended for the testing of
control valve rules for liquids and gases:
• LIQUIDS
Control valves are mounted in delivery lines. Mounting in suction lines promotes cavitation.
Assume a fully opened control valve with design flow - the method of calculation and
specification of LIP.justifies this assumption; calculateLIPc,~dwith the control valve equation. This
implies thatj(.'C)d ~ 1 and the system conforms to rule 1(d).
Calculate LIPsn~ .. and prorate for L!Psn~d· Check for rule 2(d). Consider another control valve if
necessary. Calculate LIPa.d with the MEB. Assume LIP a,n ~ LIP a,d - the accuracy of this
approximation depends on the head-flow characteristic of the pump which normally, at this stage
of the design, is not available. Calculate APe,~ .. with the MEB and check for rule 2(n). Adjust the
system if necessary.
Calculate f(.'C), with the control valve equation and test with rule 1(n). Adjust the system if
necessary.
• GASES OR VAPOURS
Control valves may be mounted in suction or delivery lines. Because flow over the valve is
normally compressible, it is necessary to divide the pipe system in two sections - upstream of
the valve and downstream of the valve. See figure 10.7.
97
SELDOM OFTEN
Consider the case with the valve in the suction line. Calculate P1.a with a MEB. Assume a fully
opened valve with design flow (see liquids) and calculate P 2,a with the control valve equation.
APc,~a ~ 1'1,"-
PM (This evaluation for the valve in the delivery line requires a trial-and-error
approach). Again, this implies thatj(."x:)a = 1 and the system conforms to rule 1(d).
Calculate APsr'~" and AP sr'~"; remember that strictly speaking prorating is invalid for compressible
flow. Check for rule 2(d). Adjust the system if necessary. Calculate AP."with the MEB. Assume
AP., ~ AP.a (see liquids). Calculate P 1,, with a MEB. Calculate P2,, with a MEB. Calculate APcv.n
~ 1'!_,- P2,, and check with rule 2(n). Adjust the system if necessary.
Calculate JM, with the control valve equation and test with rule 1(n). Adjust the system if
necessary.
Check for sonic flow in the control valve. If analysis proves that flow in the lines is also
compressible, tests for sonic flow at relevant points in the pipeline must also be performed.
NOTE : If APcv is large, it should be established whether the criterion L1P1 " ' - 0, 000164 P does
not render two different diameters for the two pipe sections upstream and downstream of the
control valve.
98
10.5.2 NEW SYSTEMS WITHOUT FLUID MOVERS
A typical example of a branched system with control is the top piping system of distillation
columns. One pump is used to serve both the reflux and product lines; each may be provided
with a control valve. One suitable design approach is the following:
Do designs for the two piping systems up to the point where LIP,a is calculated with the MEB.
Establish which system requires the largest LIP.d and design it fully.
Test the second system with the control valve rules. The format for these tests is as discussed
in 10.5.2. LIPn~amilaht< is calculated with the MEB.
For design purposes it is assumed that LIP""~ LIP• .,. In reality LIP,,, will be> LIP,,a· See figure 10.8.
99
Pressure head
·-- -"-----"-,---fu~ curve
~~
'"'-
I I
I
1 I /
l I/
~~r1
System curve ~II
I I
--- -- 1
--= -...=- _ _ - _ _ _ _ _ _ _ _ _ _ _I ___I_ ~ _ _ _ _ _ _
71\ --- : II
LlJ>EP + dJ>I!L : :
'!_ L____ _ _ _ - - - _ _ _ __,___,I I
10.7 LITERATURE
1. R. Kern, Control valves in process plants, Chern Eng, 85, April14, 1975.
2. M. Adams and D. Boyd, Control valves: time for review, Hydrocarbon Processing, 87,
May, 1984.
3. J.R. Connell, Realistic control valve pressure drops, Chern Eng, 123, Sept 28, 1987.
4. H.D. Baumann, Control valve versus variable speed pump, Chern Eng, 81, Jun 29, 1981.
5. W.L. Luyben, Process modelling, simulation and control for chemical engineers, second
ed. McGraw-Hill, 1990.
100
11. FLUID MOVERS
In general terms fluid movers for liquids are known as pumps and those for vapours and gases
are referred to as compressors.
11.1 TYPES
Radial flow
101
11.2 CHARACTERISTICS
The various fluid mover characteristics are mostly presented in diagram format and are
available from commercial suppliers. Most common is the head capacity curve. The
correlation differs substantially for kinetic and positive displacement movers. See figure 11.2.
r',;,
"""'
The operating point for a piping system is obtained when the head capacity curve of the fluid
mover and the system curve according to the MEB are combined.
Other important characteristics which are obtainable in diagram format include NPSHR, power
and efficiency vs capacity; influence of parameters like rotational speed and impeller size;
information regarding the surge zone for compressors.
102
0 50 U.S.qpm 150 100 250 JOO ]50 4()0 450 500 550
0 50 150 200 2y0 JOO 350 400 450
Lll I I I
"
'209. r+-+ I I I' 50
I I 1! I I
II II J. ro thsH '
14 I'
..:......t.....' ''
_2£.0~ '' 73 I
'.
.' I; 1 701
I
I'
30
8 'I
'' 'I
'I
II 601
' I
II
''
l" 1
6 20
' I'
11! I
I ''
I I
' ! 1I ''
4
l I I
! I I
''
I I
'' ,., 10
2
'' '' I
I
''''I
',' , ' '' < < ''I' I I'' ''' '''' '
:-'~5H -'~<f<lOon ,"nd.''~·~--:•tfe; ;~o•h••:-~~~log _.on.Q.5 "'.~1~·tl~
'' 12
--:--. ,- •• i!•i '•'' i;l\ ;o.: l'i 1 I
I I I!
J II '' I '' 10
5 :
~
' ' 'I
I':
TTl~
': '"
I
''
I '· I I '
I I
'
'• ' I'
' •
'
'
I I I
' I I ' ' ' I
I
i i I I I I
' ' I' ' i ; ! i I :'
I II I
'
I I
I I I i
' 'I ' ' I I
'
I I • '• '' I
'
''
6
'' I I I '
' I I ''I I I I I I I ' I '' I ' ·I
' '' '
' ' 'I
4 ' '
'II ' I I I I
'
: I I I I ' ' II I I ' I
' I
I
I '
' p
p '! II ' II I I I I I
'
II ' 1•: I
Ill I
' I i
:'
I
I
I
II
I
' ' ' II 'I ' I I ' ' ' :' I • ' hp.
kW I ! I I '! II
! ' ' ' '
I
' ' I I I
' ' I
,........,I ' '' '""'':I I II I
C'-• 3
I I I I
'II' I I
:
II :
I I .....' 'I ! '
:
'
I I I I I I•
4
c
I I' I I I I I I I
I
''
I
I 'I ' I ' I I
' '
' '...;........-.....-:-;; '
I
I '' I I' '
II I I I I ' ll- I I I
I
..,. J..-r' I I
'I I
r'l•'
I
·--~ I i 'I 3
- ·-
I
2 I I I '; I ' I '' II I I I
I I ' I
r ' I' 'I ' '
209' ' ' ' 'I
_,..,I I I
200' ..... ' 'I ' I ' i' I
''
I
' I' II I 'I I
I 190' _.. I ' ' ! '
:
I I I 'I ' I
I'
:
:
I I 2
I
180' ' II
I
I
' I
I I I
I 'I ' I
I
I' ! 'I
I '' I I
i I 1~01· I ::I
'
I I I
' '' I I '
I
U_l
' ' ' ' I I:
i I I I I I I I I ' I I
I
I ' I I '
I I I I II I I ' I I !
! '
I I
'I
! I I
'
: ''
I I I I
'! 'I I
' I
I I I
0 ' I
' 0
0 10 0m¥h JO 40 0 60 70 80 go 100 110 120 IJO
0 5 Otis 10 15 20 25 JO J5
103
11.3 SELECTION
Selection of the most suitable, cost effective fluid mover for chemical processing is normally the
responsibility of a fluid mover specialist with lots of experience. Variables that must be considered
include type, construction material, mechanical design, characteristics of the mover and
characteristics of the system. It is the responsibilities of the relevant engineering disciplines to
provide him with the necessary information. Information regarding the following is normally the
responsibility of the chemical engineer: fluid characteristics (temperature, type, viscosity, density),
flowrates (W•. W.n· Wa), terminal system pressures, LIP. (associated with w. and Wm in the case of
pumps), P at the suction flange and P at the delivery flange for both w. and Wm in the case of
compressors, NPSHA and NPSHR, construction material.
r.
',_ 1 Useful general guidelines are the following. Centrifugal movers are suitable for most ordinary
~·
applications. Application examples where other types of fluid movers are better choices are the
following. • Positive displacement movers for high delivery pressures (roughly LIP. > 3000 kPa)
• Gear and screw rotary movers for liquids with high viscosities (roughly v > 250 eSt) •
Diaphragm pumps for corrosive inorganic liquids e.g. certain acids • Positive displacement
movers where a fixed flowrate is of great importance (e.g. dosage pumps). Diagrams were
developed to simplify the selection of fluid movers. Examples are shown in figure 11.4, 11.5 and
11.6.
30
1111 _llllill
I Si -' Jmps
q' 10 :Dr~~
,· I 'I
• I flow
~) ~ 6
"E
~
'0 3
.2
:n• r...:::
c.
"' Jpump_
-
0.6
f::: 114
D, = DH 1.,(0, '\il
0.3 r-N =Speed, rpm 1 ii"' t-Ht+ "~T - ifT
r-- ZD' : ~~.~: ::'"Is,_ m.e ter, ,_+-+f-tll~' .. T_.Jj·._.L -Htttffii-~ ~"\"r
Httt--J;-;;-
0.1
1=
11
:mf~ i'1lll i" 1111 li1 ° "~~ I II
L__!_..L.LL!LillL..-.L....L.t...L:Ll.L!L--L--:':-.L!..:l:ll:-:-:--L:::::-.1....:':-::':!-'-!-'=---':::-::;:;~~
0.1 0.3 0,6 1 3 6 10. 30 60 100 300 600 1,000 3,000 10,000
Specific speed, N1
104
The diagram of figure 11.4 may be used for both liquids and gases. For liquids H = AH, as
calculated with the MEB. For gases and vapours H must be calculated as AH, (adiabatic).
L1Ha {adiabatic} =
[ (Pft' )fk - l!lk _ 1] ft ..... (11.1)
(k - 1)/k
)
1
= suction flange, )2 = delivery flange, T = oR, P = psia, k = Cp/Cv.
Application requires the calculation of N.; see equation in diagram; if N is not known, a typical
value must be assumed (in SA a typical value is 1 450 or 2 900 rpm). The type of fluid mover
follows from the diagram. If the mover is a kinetic type, the minimum impeller diameter associated
with maximum efficiency may also be calculated. With the type and approximate size known,
·;__ .
characteristic diagrams may be obtained from suppliers for final selection.
Figure 11.5 shows a second type of diagram which may be used for the initial type selection.
Head capacity curves for different sizes of centrifugal fluid movers may be combined in a single
diagram as shown in figure 11.6; it is convenient for size screening before final selections are
made.
I
' I I I
I I
I I I
' '
' ' '
I. .I.. II I I'
I . I I 'I ~-
Sp~Cf<lllh·g~·soeM
I
centnfu~ai-
I I" 'r~~
lI ~
0 I I II I I I ill I I I I I
"'0 20 0 I I I I i/ I II I . I ·1 ~ I I I ~ 1. I I I iI I
~
10 0
70
10 !;---_j___l__l_!----1-1-f:
Y1
___!_T-'' '1-'""I-bT_j_TlT:-'t~~_jlruJIJInt.4~~f¥~~~~~,t'~Et:Jo···_'''"_·1-L·~~::~~~~I_LI
0 tO 50 tOO 500 1.000 5.000 lO,OOO SO.OCO 100.009
CJoacity, gpm.
105
800
600
400
200
~
"
u
"0
c
}100 •c
~
-"
~
"'
20
MAXIMUM PRESSURE AT THE DELIVERY FLANGE: It must be calculated for specifying of the
class type (often A, B or C) of the fluid mover.
pmax = pfeed tank, max - LJPEL, suction line, max + LJPa, max
p mer< = p delivery tank, max + LJPEL. delivery line, max + LJP/. delil.'ery line, max
The maximum friction pressure drop must be calculated with the maximum linear velocity. Pmax is
the maximum pressure which will be encountered at the delivery flange under normal operation.
If however a delivery line valve in a positive displacement system is closed, the pressure will
increase until a mechanical failure (e.g. burst of pipewall) will occur. To prevent this a safety valve
with recycling facilities is normally used; a typical safety valve setting is 10% largerthanPm"""'"'"''
Class specification should be based on the pressure of the safety valve setting.
106
11.4 COUPLING IN SERIES AND PARALLEL
The capacity of an existing piping system may be increased by series or parallel coupling of
centrifugal fluid movers. See figure 11.7.
Pump I + Pump 2
/
r~ Flow
\ •.. :
Head
Pump I + Pump 2
Series
/
Q
o· 0 '
Q
Flow
Evaluation of operating points requires construction of the combined head capacity curve of the
fluid mover system. This can be done by simple addition of the relevant variables- note that in
the case of series coupling, flowrate is common; for parallel coupling, delivery head is common.
Hydraulic work is defined as the energy received by the fluid from the fluid mover. Real work is
the energy which must be provided to the fluid mover. Efficiency (tl) of a fluid mover is defined
as hydraulic work/real work. Power is work per unit time. Suitable equations for the calculation
of power are
Hydraulic power = L1Pa W/3600 p kW; Win kg/h
In the British unit system real power is also known as brake horse power.
107
11.6 LITERATURE
1. R.F. Neerken, Pump selection for the chemical process industries, Chern Eng, Feb 18,
1974.
.
C·•
·'
108
Piping
System
Design :
Appendix
APPENDIX A1 : Dimensional standards for plain steel pipes (unscrewed)
Recommended wall thicknesses: 4, 4.5, 5, 6, 8, 10, 12, 14, 16, 20, 22, 25 mm
2.5 X X
3 X X X
4 X X X
5 X X X X
6 X X X X
8 X X X X X
10 X X X X X
12 X X X X X X
14 X X X X X X
16 X X X X X X
20 X X X X X
25 X X X X
30 X X X X X
32 X X X X
38 X X X X
40 X X X X
44.5 X X X X
50 X X X X
57 X X X X
76.1 X X X X X
88.9 X X X X X X
108 X X X X X X
133 X X X X X
159 X X X X X
Sizes marked with an x are those most likely to be available, but you are advised to always consult
your suppliers' stock list.
APPENDIX A3 : Dimensions for polypropylene and high density polyethylene tubes
PP (Polypropylene)
SABS 1315 of 1981
Wall
Thlcl<ness
"""'""'
12
Tokraoc:e
.03
HomiMI
10
-
Min Wall
Thkl<-
""'
-
-
>Muf
{kg)
-
NomfNI
lO
-
Min Wall
Thkl<·
""'
- '
-
Mou/
(kg)
-
NomlMI
()
-
Min Wall
Thkl<·
""'
-
-
MM>/
(kg)
-
_,.,
I>
.
8
Min Wall
Thld<·
""'
20
,....
MM>/
(le!)
006
16 03 - - - - - - - - - 11 22 O,CfJ·
20 03 - - - - - - 15 23 0,13 - 1·f 27 015
25 03 - - - - - - 19 2,8 0,19 18 34 022
32 03 - - - 27 24 0,22 24 3,6 031 23 44 0,37
40 04 - - - 34 29 033 31 45 049 ·-::28'c · - sse- o·s8c
50 OS 45 24 035 42 37 053 38 56 076 36 68 089
63 06 57 3,0 056 53 46 0,83 48 71 1,20 45. 86 142
75 07 68 36 0,80 64 55 117 57 85 1 71 . -53 103 2 01
90 09 81 4,3 113 76 66 168 69 101 244 64 12 3 288
110 10 99 s;2 1,67 93 81 250 84 124 365 79 151 431
125 12 113 60 2,18 106 92 323 96 14,1 4 72 89 17 1 566
140 13 126 6,7 2,73 119 103 404 107 15,8 592 99 192 711
160 t:5·- 145 7,6 3,53 136 11 8 5,30 122 18 0 7,86 113 219 927
180 17 163 86 4,49 153 132 666 137 20,3 995 ·128 24 7 1175
200 18 181 95 5,51 170 14 7 824 152 225 12 3 142 274 145
225 21 203 10,7 6,97 190 165 106 171 254 156 160 308 183
250 23 226 11,9 8,59 211 184 13 2 190 28,2 192 177 342 226
280 26 253 133 10,7 237 206 16 5 213 315 240 199 38,4 284
315 29 285 15,0 13,6 266 232 209 240 35,5 304 223 43,2 359
355 32 320 16,9 17,7 300 261 264 270 40,0 386 252 486 455
400 36 361 19,0 224 338 294 335 305 451 49,1 - - -
450 41 . 406 21,4 283 381 331 425 343 507 620 - - -
500 45 451 238 35,0 423 368 524 - - - - - -
Sl units
Mass: kg
Length :m
Time: s
Temperature: K
Conversion factors
g = 9.807 m/s 2
1 US gallon= 0.8327 Imp gallon= 3.785 I
1 psi = 6.895 kPa
1 cP = 1 X 10'3 Pa.s
APPENDIX C1 : Absolute roughness for various pipe materials
I 0.050
I
HI~---~-~~- r---_
I-
- '; --' ~:- -- - li --- -~ 1
~·~~~
- §:::F:-h
-: , ' - - •
I
- ·· · - · -
1
·· ·- ·· - · · " ' -t ·1 I 11111·- ·· - - - ·- - · o.oo 2 N
;;:
~ ~
(1)
I .S 0
I t H-111---1-[~;~:~-1 ... _, __ . ·- _ 'r, o.oo1 "'"' 0
0.
' ·c
If.'-<
o.o2o
~~f-. 0.000. 3 J1)_
'<
0.
draulically smooth pipes I ~t:::---r-r- o.ooo 6 0 n:;·
Hlll-·----1--1-ii-1-1111- :-1-11-·-·- ' ---- ~§;'±:: - --- --. -- --"-..--- . --- "'ill
~
0.000 4
0.015
Ii f'i:: I t--._ ' , 0.000 2
3
1~1 +H-------- --
Turbulent flow
1-- _, -~ I ' l-==t-- ' t--. ' O.OCJ 1
II
~:~~~J
0.010 rr~l-nlllll1-
1=.--:- ~H-- ~.-TITIIIIII- .l !,t~~~lb<' ~~~"
~~---. I~ltl o.oo~oors~ J -1-1-1-H-1-8 '·"'"
--.
dD
o.oo·Jo1
6 3 2 3 4 56 8 2 3 4 56 8 2 3 4 56 8 2 3 4 56 8 2 3 4 56 8
103 104 1Q5 106 10 7 1QB
!
I Reynolds number Re
L ·---- ----------------.
APPENDIX C3 : Equivalent lengths for various components
Description Lr/D
Globe Stern With no obstruction in flat, bevel or plug type seat Fully open 340
valves perpendicular With wing or pin guided disc Fully open 450
to run
Angle valves With no obstruction in flat, bevel or plug type seat Fully open 145
With wing or pin guided disc Fully open 200
Cocks Straight- Rectangular plug port area equal to 100% of pipe area Fully open 18
through
Three-way Rectangular plug port area equal to 80% Flow straight through 44
of pipe area (fully open) Flow through branch 144
§ Minimum calculated pressure drop (psi) across valve to provide sufficient flow to lift disk fully
APPENDIX C4 : Resistance coefficient data for piping components
PIPE ENTRANCES
L Kr= 0.5
L Kr = 0.04
PIPE EXITS
Kr = 1.0
STRAINERS
Nomd Kr Nomd Kr
mm inch mm inch
: ... : .... : ......... ! ..•...... ; .... -·····. ) •••. : ..•. 1.• '/.'; •• _l .... :. _:,.
I
--:~-~-
'\:./·
..:~--~
.
... ! __ _
.-
'' '.-· ''--l---!-.6,-3-f---'--+--'--i---:--T.-::--ct-:--',;:i-:\' '--,---;---c--f--,--.."".,--j---,-7."
... -.·.i:
Kr-DRTA VIIi?. TsTUKK£
.,t~f--~-j~~~-~~-~--~1--·~f§·-~·tf-~-~~~~~
·· .... C-f ·/ -f-f .
J..
.. ------- __ ~--------~..-,~~~--.I(E---
•• L..~~-~---~--,~---~~~_i~-L-~---~'~-~-·~_J
o-1 o·l o.l o• oS o-4 o.ll o9 .,;'1 t-o
00
f'Low .u.,,., o.,j;. .3
o.l !+·!+H-'
..,
o.ol:...lJU.Ll...l..L.i_.:...l_!__L:..lJU.Ll-'-.L.i_~-'--L-'-''--L-'-'-L..L--J
o.o o,l o.l. o.J, o.+ o-S o . ..:. o.; o-8 o-~ l.o
fl.ow RA.r,o 0.1 /Gtl
d. • d. ' d,
(1.9
0·8
0·7
~
'< o.o.
"'.,
0
o,s '
It~
\.{ o,q.
0.3
0·<
0 0.1 o.z 0-.l 04 c.s o.~ 0,] 0.3 0,9 1,0
PLOW <RTIO oi0_J CP. Gyc-'
·~tO
0$
d. -ct.. ~d.3 -·1
·o.:r
0.7
~
',/ -· 0..0
"',, .. o.s
0
~
~
'
.It~
'X or;
...... 0.3
K, 1
K= + Kjl +-d)
Re d = inner diameter of attached pipe in inches
Fitting type K, K.
Note: Use RiD= 1.5 values for RiD= 5 pipe bends, 45° to 180°.Use appropriate tee values for flow through
crosses.
SPECIAL CASES :
K, K.
'
t
60 JO
60 -
,- i- 0,3
1_1_
' 40
50 l
I
70
0,3 ----1
~
><
~
~
}- 0,2 0,5
80
2 -+- 100
'
''l
90 -t
3 ---,-
0.4
90 200 0,5
0,5 J
j -
10 _;....
-l-
500
1---i
--'
_)
j
r- 1,5
150-
20 -J -'_,
30-+ 2000
1,5 'l 3
40__s-
50-:-
.r
3000 ,-.::~
"
_,
4
J -(_ 4000 I
2
~
5
200 ~-- 200 '
i 100
:f-
-+-
5000
'
" J-
c
3-i
250 -!
J }- 10000
l
' 3i
200--f 4 10
:~
~
-(_
~ 300 --!...
_J-
5
t
i
400 20000
300-;- 300 15
500 30000
40000 -j 20
350 --:::.._ 50000
j 10
400 400
1000
1 "
3
m /mio
100
10 JO
2000_3:- 40
500 500
3000 _i 200
15 20 50
==t 300 20
1he nomogram is based on the Prantl-(oalbrook formula using a k factor of: k = opo7 rrm
Factors "ffO)Iicable to other flow formulae ore:
Hazen Williams ...... _........ _.. c ='1 50
h\anning ..................... . n = 0,010
Darcy roughness factor ......... :o--= o,rxn rrm
APPENDIX 01 : Calculation of APr.K for isothermal compressible flow
Known : Pp T, M, 11, W, D, L, e
= MP 1
P!
RT I
w
111
p!A
I
I
p1u 1D
Re =--
11
Since flow is turbulent for gas flow, solve Colebrook equation for j',
j'=0.02 can be used as first estimate :
e!D + 2.51 )
3.7 Re/1
Solve isothermal model for P2 , the In-term can be neglected for first estimate:
p2 = , p; - ( :r (: ) (f'j; ;J ) + 21n (
APPENDIX 02: Calculation of Wfor isothermal compressible flow
Find first estimate for f' using the von Karman equation for fully rough flow:
-
1
- ~ 2 log ( _!}__) + 1.74
IJ: 2s
A ~ "-D2
4
;J )
w~A
\ ( ::) ( ~ + 2ln (
w
u, --
I p,A
I
I
p u 1D
Re ~ -1 -
I fl I
Since flow is turbulent for gas flow, solve Colebrook equation for ],
]=0.02 can be used as first estimate:
s/D + 2.51 )
3.7 Re/1
MP 1
---
Pt
RT1
I
Ill --
w
p 1A
I
Since flow is usually turbulent for gas flow, solve Colebrook equation for ],
j'=0.02 can be used as first estimate :
1 s!D + 2.51 )
-2 log (
11 3.7 Re/1
= MP 1
p, RT
I
Find first estimate for f' using the von Karman equation for fully rough flow:
J1 = 2 log ( ~) + 1.74
1
]L
D
= ~
k
lnp'p + ( 1_(P,)') (k-1 p1 2k
+ P,p,(~)')
W
1
(pw)
A 1
2
+ k-1
2k(P~ 1
) k (
+ 2_.
k-1
r,)
Pz
Since flow is usually turbulent for gas flow, solve Colebrook equation for ],
f=0.02 can be used as first estimate :
e!D 2.51 )
3.7 + Re/1
1000
8 Reynolds No. Turb-Turb Vise-Turb Turb-Vise Vise-Vise ~
"tl
6 "tl
m
ReL >2000 <1000 >2000 <1000 2
4 0
3 ReG >2000 >2000 <1000 <1000 x
"T1
2 ::0
CD
Turb- turb !!!.
Vise- turb a·
100 <t>L ::J
"
€<
....
6 Vise- vise i
CD
::J
4 N.B. If X 2: 1.0 use <I>L ~
~ 3 m
€< If X<1.0 use <t>G ::J
a.
2 'El<
0'
~
Turb- turb
!'
"0
::J"
Vise- turb m
4
3
<t>G~ Turb- vise
Vise- vise
-~"'
CD
2
t
1
1
2
1 H3JHjr~-;c.-;--;}~
46 8 0.012 6 8 0.1 2
4
~I
6 81.0 2 6 81 0 2 4
Iii
4
1 J1iiJqil1
4 6 8100
X= V -~PL/-~PG
APPENDIX G1 :L1P 1"' and L1P 100ft for mild steel systems
• LIQUIDS
FLOWRATE
LIP,"' (kPa) L1P1ooft (psi)
m3/h lgpm
< 25 < 100 0.35- 1.35 1.5-6.0
25- 125 100- 500 0.25-0.90 1.0-4.0
125- 1250 500- 5 000 0.10-0.50 0.5- 2.0
> 1250 > 5000 0.04-0.25 0.2- 1.0
N~ s otids I, 52 Steel
With solids 7,5 2,29 Monel 0' oicl<el
( 6 !-lin.- IS Mox.l {1,83-I.,S7)
{Min. -!-lox.!
Steam
Saturated: to- 30 ::>Sigl 66,67- 20,32- Steel
{0- 206,84 KPog} 100. 30,4S.
Saturated or $LSf?Nheo\e{( 100 30,1.8 Steel
{30-\50 ;JS19]
(20C.,8t. i03t.,22 kPng]
-
166.67.
-
SO, 60.
Superhec1ed, (ISJ DS><J uoi \08.3 3 . 33,02. Steel
(103<., 12 kPo9 u;.J 250. 76,20.
Shor1 l1nes 2$0 ,'-.lox. 76,20 Mo;o;.
Sulluric oc.id
1,22 SS-316 ieod
88-93%
93·100% '' I, 22 Cost oron i iteel, Scto.80
S<.Jllur dioxide 66.b7 20,32 Steel
Styrene 6 I ,&3 Steel
Tr ichlor e tnyle n e 6 1,83 Steel
'/my\ ~htor1de 6 t,eJ Steel
'linylident;> c.n!orioe 6 '\,83 Steel
Wolt;>r
Pump suction U•t:>s 3 -8 0,914-2.1.4 Steel
Averc9e SN"icc 3 -e 0,9\t. -2,44 Sto:e I
{0119. SJ \Oil'}, \,'OJ)
Mo.~. ec.o,..om-cc.~ (v~ol} 7- \0 2.13- J,OS
SE:o cr>d brc<..ki~n w:ru. S-8 152-2,1.£.
tJ ~-"J i0.91 '-l:r.)
~- 12 I,S2- 3,66.
{J !-lin) tO.St ~1inJ
NOTE: The V~!-\ocal~s e1re suqqeshve o<'~ly ond ctre to Pe u~>e<i t¢ appto)fimote tine ~iH
= storlm<j point lor pas~rc- Qrop c.pfcvkttions. fhe linn! tin<! size. ~r.ould oe 5vch
9i11>? o."\1"1 ecomtnlC-O\ boSon(::e oetwe<'n pressure drop on.o re(l<;."On<fbte ve!oc:ty.
0'
o~
0
to
APPENDIX H : C., values
0.75 20 5.4 8
1 25 9 12
1.25 32 14 18
1.5 40 21 28
2 50 36 48
2.5 65 54 72
3 80 75 110
4 100 124 195
6 150 270 450
8 200 480 750
10 250 750 1160
12 300 1080 1620
14 350 1470 2000
16 400 1920 2560