GAS FLOW CALCULATIONS AND SIZING OF PIPE

   Usually, the first step in designing a gas pipe system is to estimate, as covered in Chapter III,
the magnitude and distribution of the design loads that the system will be called upon to carry.
Then, flow calculations establish either the pipe sizes required to maintain adequate pressures
throughout a new system during peak demand or the ability of an existing system to maintain
adequate pressures. The procedures used for the common types of design situations are covered
in later chapters. This chapter, which includes an advanced supplement, is concerned with the
basic flow-calculation procedures and their use in pipe sizing.

   The pressure of gas flowing through a pipe of uniform internal diameter gradually decreases in
the direction of flow. The magnitude of the pressure decrease or drop depends on the flow rate,
the dimensions and wall roughness of the pipe, and the temperature, average pressure, and
physical properties of the flowing gas. Calculations of the pressure drop can help determine
whether a pipe layout can deliver gas at adequate pressure during periods of peak gas demand.

Gas Flow Fundamentals

Viscosity

   The property of liquids and gases that quantitatively defines their resistance to flow is called
viscosity. Molasses is a common example of a liquid with high resistance to flow, or a high
viscosity; whereas water has a low viscosity. Similar variations occur in the viscosity of gases.
The viscosity of gases at room temperature and at distribution system pressures is much less
than the viscosity of most common liquids. The viscosity of water at room temperature is
approximately 100 times that of natural gas at room temperature and at pressures less than 100
psig. A representative value for the viscosity of natural gas mixtures under these conditions is 7.1
x 10-6lbm/ft-s. Viscosity is often expressed in micropoises. Micropoises are converted to the
engineering units of lbm/ft-s, when multiplied by the factor 6.72 x 10-8.
   The viscosity of a gas mixture increases slightly with increasing temperature at pressures of
100 psig or less; but at a constant temperature, it remains nearly constant with changes in
pressure over the range 0 to 100 psig.

Flow Behavior

   Osborne Reynolds was the first investigator to study systematically the conditions
under which different types of flow are encountered for tubing. He injected dye into
water flowing through glass tubes, and observed the pattern of the dye filament
downstream from the injection point. Reynolds found that at low velocities the particles
of fluid move in a direction parallel to the tubing axis. This is termed laminar flow (Fig.
4-1a). At higher velocities, he found that the paths of the fluid particles are tortuous (Fig.
4-1b). This type of flow is called turbulent. It occurs in most gas distribution piping
during periods of peak use.
(a) (b)
Fig. 4-1. Flow Patterns for Laminar and Turbulent Flow
 
Fig. 4-2. Partially Developed Turbulence
   Subsequent investigators found that the transition from laminar to turbulent flow, as the flow
rate increases, is not abrupt nor complete. Instead, laminar flow persists to very high velocities in
a layer of fluid, called a sublayer, next to the pipe wall (Fig. 4-2). This flow is termed partially
turbulent. As the fluid velocity is increased, the laminar sublayer is assumed to become thinner
and the turbulent core larger. In pipe with rough walls, such as commercial pipe, the thinning of
the sub-layer is assumed to continue until its thickness becomes about the same as the height of
the projections of the pipe wall into the gas stream (Fig. 4-3.) When this point is reached, the
turbulence is said to be fully developed; or the flow is termed fully turbulent.
   During periods of peak demand, the flow in most pipe sections is partially turbulent. This is true
for plastic pipe in general, clean cast iron pipe in low-pressure gas distribution systems, and clean
steel pipe in distribution systems with a maximum source pressure of about 30 psig (207 kPa).
Both partially and fully turbulent flow behavior can be expected in steel pipe in gas distribution
systems operated at source pressures greater than 30 psig.
 
 
 
Fig. 4-3. Fully Developed Turbulence
 
Flow Equations
The relationship between the flow variables is expressed in a flow equation. The flow equations
that best represent partially and fully turbulent flow behavior in distribution systems are different,
however. There is the lGT distribution equation and the fully turbulent flow equation. The
difference in the two are the viscosity in the lGT distribution equation and the effective roughness
in the fully turbulent flow equation.
In addition to these two equations, other flow equations have been developed since the beginning
of the commercial transportation and distribution of gas. The early flow equations, such as
Spitzglass and Weymouth, were based on flow studies over a limited range of conditions. For this
reason, their area of application is limited. The formulas for the ones most commonly used in the
gas industry are in Table 4-1. The transmission factor included in this table is discussed in the
Supplement for the Advanced Reader, comprising the second part of this chapter.
   The fully turbulent flow equation requires a value for the effective wall roughness, k. Flow
measurements conducted on pipe in natural-gas service have indicated that the effective
roughness of clean, commercial steel pipe ranges from 0.50 to 0.75 mils (0.013 to 0.019 mm) and
is independent of the pipe diameter. A representative average value for clean steel pipe is 0.7 mil
(0.018 mm), and this is frequently assumed when information on wall roughness is not available.
The effective roughness of steel pipe stored up to two years before installation can range up to
2.0 mils (0.051 mm). Wall roughness is sometimes also expressed in micro-inches (millionths of
an inch).
   AGA's Plastic Pipe Manual for Gas Service (American Gas Association, Plastic Pipe Manual for
Gas Service, Arlington, Virginia, 1977) suggests using an effective roughness of 0.06 mil (0.0015
mm) for plastic pipe and 10 mils (0.25 mm) for cast-iron pipe.
 
Table 4-1. Formulas and Transmission Factors for Commonly Used Flow Equations
EQUATION FORMULA TRANSMISSION FACTOR
     

   

Fully 4log(3.7d/k)
Turbulent
     

   

   

lGT 4.619(Re)0.100
Distribution
     

   

Mueller 3.35(Re)0.130
Equation
     

Panhandle A  
Equation
6.872(Re)0.730
     

Spitzglass
(High
Pressure)
     

Spitzglass
(Low
Pressure)
     
   

Weymouth 11.19D1/6
The units of the quantities in all of these equations are:
D= in. P1, p2, Pb = psia TfTb = ºR b
The constant 2.450 includes:

hw = in. wc Q = Mcf/hr G = specific gravity    m=7.0x10-6lbm/ft.sec.

L . ft m = lbm/ft.sec. Z = compressibility factor c

The constant 3.415 and 3.550 include:

   pb = 14.7 psia

   Tb = 520ºR

   Tf = 522.6ºR
 
 
 
Terms used in the flow equations
Qb = gas flow rate at base conditions, SCF/h*
D = inside pipe diameter, in.
P1 = upstream pressure, psia
P2 = downstream pressure, psia
L = distance between upstream and downstream points at which P 1 and P2 are
measured, ft
Tb = base temperature, ºR
Pb  = base pressure, psia
G = specific gravity of gas with respect to air, dimensionless
Tavg = average absolute temperature of flowing gas, ºR
Zavg = average compressibility factor of gas at flowing conditions, dimensionless
(in most distribution system flow calculations, the value of this factor can be
taken as 1.0)
f  = friction factor, dimensionless
m = gas viscosity, lbm/ft-5
k = effective roughness of the pipe wall (the average depth of its surface
depressions), mils (thousandths of an inch)
 
Low-Pressure Equations

   High-pressure flow equations are seldom used directly in flow calculations for piping operated
at low pressures. They do describe gas flow in a low-pressure system, but simpler forms of these
equations can be used without loss of accuracy. Thus, in the low-pressure Spitzglass equation
shown in Table 4-1, the pressure drop is expressed in inches of water column, instead of the
difference in the squares of the upstream and downstream pressure. The low-pressure form of
any equation in Table 4-1 can be obtained from the high-pressure form by substituting 1.062 h w
for the term P12-P22.

Use of Flow Equations

   Examination of the parameters in the gas flow equations shows that they can be classified into
five categories:
1. Constants

- Base pressure, Pb, the value of which is usually taken to be 14.73 psia (101.56 kP a)
- Base temperature, Tb, the value of which is usually taken to be 520ºR (288.89K)
1. The gas properties of specific gravity, G, viscosity, p, and average compressibility factor,
Zavg.
2. A pipe property - the effective wall roughness, k
3. A field variable, Tavg.
4. The basic flow variables

- The pressure drop in low-pressure systems, rP, or the difference in the squares of the upstream
and downstream absolute pressures, (P12-P22).
- The gas flow rate, Qb
- The pipe dimensions of diameter, D, and length, L.
The values of the gas properties depend primarily on its composition and normally do not vary
from one section of pipe to the next in a gas distribution system. At pressures up to 100 psig (689
kPa). the upper limit of pressure in most gas distribution systems. the value of Z avg can be taken
as 1.0. The flowing gas temperature is the ground temperature and varies with geographic
location and the season of the year. The basic flow variables are those that change from one flow
problem to the next. In most flow calculations, the values of all but one of these variables are
known and the objective of the calculation is to determine the value of the unknown.
Here are three examples of the kinds of questions answered by a simple flow calculation for a
pipe section carrying a specified gas:
1. For a specified pipe size and length, what is the capacity at a specified value of pressure
drop or rP2, (P12,(P12-P22)?
2. For a specified pipe size and length and a specified upstream pressure, P, what pressure
drop or rP2 will be experienced at a specified flow rate?
3. What pipe diameter is required for a pipe section of specified length to carry gas at a
specified flow rate without exceeding a specified pressure drop or r(P 2)?
   Before working on examples of these basic flow calculations, we need to look at a simplified
calculation procedure and some calculation aids.
Flow Calculations
   Flow can be calculated by substituting the known flow variables and gas properties into the
appropriate flow equation and solving for the unknown flow variable - commonly the gas flow rate
or the downstream pressure. These rates are usually calculated by the use of a special-purpose
flow slide rule, a programmable pocket calculator, or a digital computer. These tools will be
discussed following these examples of manual flow calculations.

Resistance Factors

   For manual flow calculations, all the constants, the gas properties, the flowing-gas temperature,
and the diameter term from the flow equation combine to obtain a convenient quantity called the
pipe section unit length resistance factor, R. The flow equation is expressed in the simplified
form:
   rP2 = RL Qbn (Eq. 4-1)
= K Qbn
where: K = pipe section resistance factor
R = resistance factor per unit length of pipe, K/L
n = exponent whose value ranges from 1.74 to 2.0, depending on the flow equation used
   The resistance-factor expressions along with the appropriate value of the exponent n are
summarized in Table 4-2 for the flow equations of Table 4-1.
   The R factor, the resistance factor per unit length, is useful because for each distribution
system, the value of R depends only on the flow equation selected and the inside pipe diameter.
The physical properties of the distributed gas are the same throughout most distribution systems.
If a table of unit length resistance factors for a flow equation is prepared for all pipe diameters
encountered in a distribution system, the resistance factor, K, for each pipe section in the system
can be determined by simply multiplying the length of the section by the appropriate unit length
factor, R.
   Tabulations of the R factors are given in Tables 4-3 and 4-4 for the Spitzglass high- and low-
pressure equations for gases with gravities of 0.60 and 0.65. Table 4-5 contains the gravity
factors for converting the resistance factors in Tables 4-3 and 4-4 to other gravity bases. For
example, the Spitzglass high-pressure equation resistance factor for 4-inch, schedule-40 steel
pipe carrying gas with a specific gravity of 0.68 may be computed as follows. From Table 4-3, the
value of the resistance factor for a 0.60-gravity gas is 9.803 X 10 -5; for a gas gravity of 0.68, Table
4-5 gives a gravity factor of 1.133. The product of these two values provides the resistance factor
for a gas gravity of 0.68:
R0.68  = 1.133 X 9.803 X 10-5
= 11.107 X 10-5
   Tables 4-3 and 4-4 are presented to illustrate resistance-factor tables and their use. Because
they are based on the Spitzglass equation does not imply that this equation should be used in
most or any distribution system flow calculations. The Spitzglass equation was selected for
illustration because it is a simple, widely used equation.
 
 Table 4-2. Common Flow Equation Unit Lengtha Resistance Factors
Equation Resistance Factors per Footb Flow Rate Exponent
Fully Turbulent CTfGZavg/(log 3.7D/k)2D5 2.00
      English Units: C = 3.645X10-3  Metric Units: C=1.349X104
IGT Distribution CTfG0.80m0.20/D4.80 1.80
      English Units: C=3.418X10-3 Metric Units: C=3.000X103
Mueller CTf 1.74
-3
      English Units: C=6.922X10  Metric Units: C=3.903X103
Panhandle Ac CTfG0.855/D4.856 1.855
      English Units: C=2.552X10-4 Metric Units: 9.628X102
Spitzglassd CG/D5 2.00
      High Pressure

            English Units:  C=8.575x10-2(1+3.6/d+0.03D)

            Metric Units:   C=1.7633X105(1+91.44/D+0.001 181 D)

      Low Pressuree

            English Units: C=7.935X10-2(1+3.6/D+0.03D)

            Metric Units: C=8.718X104(1.91.44/D+0.001 181 D)

      Weymouth  CTfG/D16/3

            English Units: C=4.659X10-4

            Metric Units: C=5.069X103
a
Term English Metric   
   D in. Mm   
   L Ft Meters   
   P Psia KPa   
   Qb MCF/h m3/h   
   Tf ºR ºK   
   rP in. wc mm wc   
   m lbm/ft-s Cp   
b
The following standard conditions are included in the constant::

      Tb=520ºR and Pb=14.73 psia

c
Constant includes: m  = 7.0X10-6lbm/ft-s
d
Constant includes: Pb = 14.7 psia

      Tb = 520ºR

      Tf = 522.6ºR=62.6ºF
e
General equation: Hw = RL(Qb)2
Table 4-3. Unit Length Resistance Factors for Spitzglass High-Pressure Equation
                  Resistance Factor ** 
Nominal Size Actual 
of Pipe (in.) I.D. (in.) G = 0.65 G = 0.60
1 1.049* 1.960X10-1 1.809X10-1 
2 2.067* 4.142X10-3 3.823X10-3 
3 3.068* 4.644X10-4 4.287X10-4 
4 4.026* 1.062X10-4 9.803X10-5 
6 6.065* 1.206X10-5 1.113X10-5 
8 8.125  2.656X10-6 2.452X10-6 
10 10.250  8.172X10-7 7.543X10-7 
12 12.250  3.357X10-7 3.099X10-7 
16 15.500* 1.057X10-7 9.757X10-8 
20 19.500* 3.498X10-8 3.229X10-8 
24 23.500* 1.445X10-8 1.334X10-8
*
Diameter of schedule 40 steel pipe
**L=ft
Qb = MCF/h
Diameter of pipe with 0.250 in. wall
Table 4-4. Unit Length Resistance Factors for Spitzglass Low-Pressure Equation
                  Resistance Factor * 
Nominal Size Actual 
of Pipe (in.) I.D. (in.) G = 0.65 G = 0.60
4 4.00 1.017X10-4 9.388X10-5 
6 6.04 1.140X10-5 1.052X10-5 
8 8.15 2.419X10-6 2.281X10-6 
10 10.12 8.063X10-7 7.443X10-7 
12 12.12 3.276X10-7 3.024X10-7 
16 16.16 7.992X10-8 7.377X10-8 
20 20.24 2.710X10-8 2.502X10-8 
24 24.28 1.147X10-8 1.059X10-8
*L=ft
Qb = MCF/h
 
Table 4-5. Gas Gravity Factors.
Specific
Gravity,
G  0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
0.50 0.833 0.850 0.867 0.883 0.900 0.917 0.933 0.950 0.967 0.983 
0.60 1.000 1.017 1.033 1.050 1.067 1.083 1.100 1.117 1.133 1.150 
0.70 1.167 1.183 1.200 1.217 1.233 1.250 1.267 1.283 1.300 1.317 
0.80 1.333 1.350 1.367 1.383 1.400 1.417 1.433 1.450 1.467 1.483 
0.90 1.500 1.517 1.533 1.550 1.567 1.583 1.600 1.617 1.633 1.650 
1.00 1.667 1.683 1.700 1.717 1.733 1.750 1.767 1.783 1.800 1.817
    
   Diameter data for use in flow calculations are summarized in Table 4-6 for cast-iron pipe, in
Table 4-7 for steel pipe, and in Table 4-8 for plastic pipe. Here are three examples of manual flow
calculations made by use of the data in Tables 4-2 to 4-7:
Example:
To carry gas one mile from a large-diameter utility supply main to a small city, a 6-inch steel main
is proposed. This line is to be sized for the maximum gas demand expected in the community for
five years. The design engineer estimates that this load will be 250 MCF/h, and that the pressure
at the inlet of this 6-in. main will be 60 psig. Use the Spitzglass high-pressure equation resistance
factors from Table 4-3 to calculate the pressure drop for this main, given this additional
information:
   Inside diameter (ID) of pipe = 6.065 in.
   Atmospheric Pressure = 14.7 psi
   Pb = 14.73 psia
   Tb = 520ºR Tf = 490ºR
   Specific gravity of gas 0.671
Solution:
   First, the 6-inch resistance factor for a specific gravity of 0.60 from Table 4-3 must be adjusted
to a specific gravity of 0.671. From Table 4-3, the 0.6 gravity resistance factor for a 6-in. pipe is
1.113X10-5; for a gas gravity of 0.671, Table 4-5 gives a gravity factor of 1.119 by interpolation
between the values for 0.67 and 0.68. The product of these two values provides the resistance
factor for a gas gravity of 0.671.
R0.671 = 1.113 X 10-5 X 1.119 = 1.245 X 10-5
   The basic equation used in this flow calculation is Eq. 4-1.
P12 - P22 = RLQbn
 
   On substituting the values given in the problem statement into this equation, using the flow rate
exponent from Table 4-2, it becomes
(60 + 14.73)2 - P22 = 1.245 X 10-5 (5280) (250)2
P22  = 5584.6 - 4180.5 = 1404
P2  = 37.5 psia
P2  = 37.5 - 14.73 = 22.8 psig
Example:
Calculate the capacity of the main of the previous example if the downstream pressure required is
30 psig.
Solution:
Substituting into the basic equation
(60 + 14.73)2 - (30 + 14.73)2  =  1.245 X 10-5 (5280) Qb2
Qb2  =  (5584.6 - 2000.8)/1.245 X 105 (5280)
    = (3583.8)/0.06574
Qb2 = 54,514.8
Qb  = 233.5 MCF/h
   In pounds-pressure flow calculations, the relation between upstream pressure and downstream
pressure is not linear. For example, if the inlet pressure for a pipeline, operating with an inlet
pressure of 50 psig and an outlet pressure of 30 psig is raised by 5 psi to 55 psig, the
downstream pressure will increase by 7 psi to 37 psig.
Example:
Calculate the minimum diameter required for the main of the two previous examples if the
downstream pressure required is 30 psig and the estimated load is 300 MCF/h.
Solution:
The basic equation will be solved for the unit length resistance factor. Then, the minimum
diameter required is the smallest diameter with a unit length resistance factor less than this
calculated value.
Substituting into the basic equation:
3583.8 = R(5280) (300)2
R = 3583.8/5280 (90,000)
= 7.542 x 10-6
 
   As indicated in Table 4-3, the resistance factor for an 8-in. steel pipe, 2.656 x 10-6, is less than
the resistance factor required. Because the resistance factor for a 6-in. pipe is greater, the
minimum diameter required is 8-in.
    
Table 4-6. Dimensions of Cast-Iron Pipe
Pipe Diameter(Inches)    
Actual O.D. (inches) Actual I.D. (Inches)
2 - -
2 1/4 - -
4 4.80 4.00
6 6.90 6.04
8 9.05 8.15
10 11.10 10.12
12 13.20 12.12
16 17.40 16.16
20 21.60 20.24
24 25.80 24.28
30 31.74 30.04
36 37.96 36.06
 42 44.20 42.06
48 50.50 47.98
 
SOURCE: Cast Iron Pipe Research Association, "Handbook of Cast Iron Pipe." 2nd Ed.
Chicago: The Association, 1952.
 
 
 
Table 4-7. Dimensions of Steel Pipe
Pipe Diameter  Schedule  Wall Thickness  Actual O.D.  Actual l.D. 
(inches) Number (inches) (inches) (inches)
¼ 40 0.088 0.540 0.364
½ 40 0.109 0.840 0.622
¾ 40 0.113 1.050 0.824
1 40 0.133 1.315 1.049
1¼ 40 0.140 1.660 1.380
1½ 40 0.145 1.900 1.610
2 40 0.154 2.375 2.067
4 -- 0.188 4.500 4.124
-- 40 0.237 4.500 4.026
6 -- 0.188 6.625 6.249
-- -- 0.250 6.625 6.125
-- 40 0.280 6.625 6.065
8 -- 0.219 8.625 8.187
-- 20 0.250 8.625 8.125
10 -- 0.219 10.750 10.312
-- 20 0.250 10.750 10.250
12 20 0.250 12.750 12.250
16 10 0.250 16.0 15.500
20 10 0.250 20.0 19.500
24 10 0.250 24.0 23.500
-- 15 0.312 24.0 23.376
SOURCE: Perry, J.H., "Chemical Engineers' Handbook." 3 Ed. New York: McGraw-Hill
rd

Book Co., Inc., 1950.

 
Table 4-8. Dimensions of Thermoplastic Pipe: Minimum Wall Thickness
Nominal Outside SDR = 32.5          
Pipe Size Diameter SDR = 26 SDR = 23 SDR = 17 SDR = 13.5 SDR = 11
(inches) (inches)
1/2 0.840       0.062 0.062 0.062 0.076
3/4 1.050       0.090 0.090 0.090 0.095
1 1.315       0.090 0.090 0.097 0.119
1 1/4 1.660       0.090 0.098 0.123 0.151
1 1/2 1.900       0.090 0.112 0.141 0.173
2 2.375       0.113 0.140 0.176 0.216
3 3.500       0.167 0.206 0.259 0.318
4 4.500       0.214 0.264 0.333 0.409
6 6.625 0.204 0.255 0.316 0.390 0.491 0.603
8 8.625 0.652 0.332 0.410 0.508 0.639 0.785
10 10.750 0.331 0.143 0.511 0.633 0.797 0.978
12 12.750 0.392 0.490 0.608 0.750 0.945 1.160
SOURCE: American Gas Association, "Plastic Pipe Manual for Gas Service," Arlington,
Virginia. The Association, 1977.

Equivalent Length

   The relative gas-carrying capacities of pipe sections are sometimes expressed in terms of the
equivalent lengths of pipe of a particular diameter. This is the length of a pipe of the specified
diameter that would have the same resistance factor, K, as a given pipe section of a different
diameter. Equivalent lengths can be calculated by use of the following _expression:
ReLe  = Kp
   Le  = Kp/Re (Eq. 4-2)
Where Kp = resistance factor for any pipe section
Re =  resistance factor per unit length of pipe of base or reference diameter
Le =  equivalent length of pipe of reference diameter
When using this equation, one must be sure that the units of all terms used to calculate K p and Re
are consistent.
Example:
Several pipe sections connected in series have a Spitzglass low-pressure equation effective
resistance factor of 0.03 for gas with a specific gravity of 0.65. Calculate the length of 6-inch (152-
mm) cast-iron pipe equivalent to these sections.
 
Solution:
From Table 4-4, Re = 1.140 X 10-5. Substitution of this value into Equation 4-2 gives:
Le =0.03/1.140 X 10-5 =2632 ft
   The principal use of the equivalent-length concept is to express the flow resistance of fittings in
terms of the resistance of an equivalent length of straight pipe of the same size as the fitting. This
length is greater than the length of the actual flow path through the fitting because changes in
direction in fittings like elbows, and constrictions in cross-sectional areas in fittings like valves
generate additional turbulence beyond that present in a straight section of pipe. The effective
length of a section of pipe that includes fittings is the length of straight pipe plus the sum of the
equivalent lengths of all the fittings.
   General tables and monographs of equivalent lengths have been developed by manufacturers
of pipe fittings. Equivalent lengths of special-purpose fittings, such as excess-flow valves, are
usually available from their manufacturer.

Computer Flow Programs

   Programmable pocket calculators and all types of digital computers. main frame or
minicomputers accessed via terminals and microcomputers. are now used extensively to make
flow calculations. Digital computers can be programmed to permit the selection of a flow equation
from a menu of equations. They can be programmed to be solved for any one of the basic
variables . Qb, P1, P2, D, or L . if values of the other flow variables and the flow parameters of p,
G, and Tf are provided in the input. These programs can be written so users simply follow the
instructions and responds to requests for input. They can be programmed to permit specified
variables to be changed over a range of values to determine the effect of the changes on a
particular variable of interest.

Sizing of Services and Individual Mains

   The simplest application of flow calculations is the sizing of service lines and individual mains.
This can be done by the straightforward use of a flow equation.

   Services

   In addition to the straight run of piping from the nearest main to the structure to be served,
several fittings are usually required at each end of a service. Typical service connections to mains
are shown in Fig. 4-4. The resistance of these fittings to gas flow is usually expressed in terms of
an equivalent length of straight pipe of the same size and material as the fitting.
   A brief tabulation of approximate equivalent lengths for the fittings commonly used in services
connected to cast iron mains is given in Table 4-9. The values for the street tees will differ for
services connected to steel mains; values for the rest of the fittings remain the same. For pipe
sizing purposes, the total length of the service is taken as the sum of the lengths of the straight
runs of pipe comprising the service, plus the sum of the equivalent lengths for each of the service
fittings.
Example:
A 1 1¼-inch steel service is to be laid 40 feet from a cast-iron main to an outside gas meter set.
The standard main fittings to be used are an insulating street tee and a street elI. No stopcock is
to be used near the main because local codes allow the use of a shutoff valve which is
incorporated in the meter loop. Determine the effective length of the service for use in a pipe-
sizing calculation.
Solution:
Equivalent lengths from Table 4-9
 
 
  EQUIVALENT
FITTING LENGTH, FT.
   
Street Tee 10.5

Street ElI 7.5

Outlet fittings 8.0

Total equivalent length of fittings 26.0

Length of straight run of service pipe 40.0

Total equivalent length of service for pipe sizing 66.0

calculations
 
 
   An equivalent fitting allowance table for plastic services in high pressure distribution systems is
given in Table 4-10. In this table, the values in the tee size columns under the plastic heading
apply for plastic mains with plastic fusion tees through which a shell cutter was inserted to make
an opening into the main. The much higher values under the steel heading are for steel mains
fitted with self-tapping metal tees with a higher flow resistance than plastic tees.
   More extensive equivalent length data are given in Appendix I of the "Distribution Design" book
of the AGA GEOP series (American Gas Association, "Book D-1: System Design," Gas
Engineering and Operating Practices Series, Arlington, Va., 1990).
 
Table 4-9. Flow Resistance Cf Service Fittings for Steel and Copper Tubing Services
Connected to a Cast Iron Main
Length of Straight Pipe With Equivalent Flow Resistance, ft.
   Copper Tubing Steel Pipe
   1-in. 1 ¼ in. 1 ¼-in 1 ½-in.
Street Tee 5.5* 9.5** 10.5 15.0**
Street Elbow 5.0* 4.0 7.5 7.5
Curb Cock 3.5 3.5 13.5 12.0
Outlet Fittings  6.0 4.5 8.0 22.0
(Tee and Service Cock)
  
* 1 1¼-in. fittings

** 1 ½-in. tee attached to sleeve around main, 1 ¼-in. hole in main.

1 ½-in. elbow.
Source:  Menegakis, D., and E. H. Luntey, "Experimental Investigation of Flow
Characteristics of Low Pressure Service." AGA Operating Section Proceedings 1961:
Paper DMC-61-15.
 
Fig. 4-4. Service Connections to Mains
 
Table 4-10.  
Equivalent Fitting Allowances for Plastic Pipe 
High-Pressure Mains (15-80.96 psig)
Pipe I.D. ½" Tee 1" Tee 2" Tee 3" Tee 4" Tee Steel
½. CTS 0.445 12 11          13
1. IPS 1.033 164 58 33       112
1 ½. IPS 1.304       29       117
2. IPS 1.865       54       96
3. IPS 2.747          80    128
4. IPS 3.536             100 160
NOTES: The equivalent footage includes a curb valve, service tee, 450 elbow, service valve,
tap at main, and, for medium and high pressure mains, a regulator connection.

Because fitting allowances for plastic vary widely, manufacturer data should be consulted or
tests conducted.
 
Source: American Gas Association, "Book D-1: System Design," Gas Engineering and
Operating Practices Series, Arlington, Va., 1990
    
Capacity tables, Tables 4-11 and 4-13, are widely used to simplify service sizing. They show the
flow rates corresponding to a specified value of maximum allowable pressure drop for services of
the sizes and lengths most commonly used. The values in Table 4-11 were calculated from the
Mueller Equation, recommended for service flow calculations, for the gas properties, and flow
conditions specified in the table.
Example:
What size steel service should be used to serve a maximum load of 2500 CF/h. The service is 50
feet Iong and the maximum allowable pressure drop is a DP of 40. What if the service is 100 feet
long?
Solution:
At a service length of 50 feet, Table 4-11 shows that a ¾-in. service is too small as its capacity is
1440 CF/h. A 1-in. service is adequate (2780 CF/h. capacity). A 1-in. service is too small at a
service length of 100 feet because its capacity is only 2100 CF/h; therefore, a 1 ¼-in. service
must be used.
Table 4-12 provides gas service capacities for plastic services off high-pressure plastic mains. It
is used in the same manner as Table 4-11 as illustrated in the example above.
 

Stub Mains

   A stub main is any length of main that is not part of a closed loop of piping. In the network
diagram shown in Figure 4-5, Pipe Sections 1-2, 12-13, and 12-14 are stub mains. Stub mains
feed gas only to the services connected to them. Thus, the gas flow rate in a stub main is not
constant, but decreases at each service off-take as shown in Figure 4-6. As with services, the
resistance of valves and main fittings to gas flow is expressed in terms of an equivalent length of
straight pipe of the same size as the fitting.
Fig. 4-5. Network Diagram With Stub Mains
 
 
Fig. 4-6. Gas Flow Rates in a Stub Main
 
   The most straightforward way to calculate the pressure drop along a stub main is to treat the
main as a series of short pipe sections between services. The calculation of the pressure loss in
each of these pipe sections is performed easily by the straight forward use of a flow equation;
however, the loads on a stub main can often be concentrated without any significant loss of
accuracy. Thus, for a stub main that feeds equally spaced loads of equal volume, a close
approximation to the total pressure loss is obtained if the entire distributed load is concentrated at
a point one-third of the main length from the input end.
Table 4-11. Capacity of Steel Services in a Medium-Pressure Distribution System*  
Service Capacity, CF/h
Service Pipe Size, In.
Length, Feet** ¾ 1 1¼ 1½ 2
10 2,140 4,140 8,730 13,300 26,300
20 1,890 3,640 7,690 11,700 23,100
30 1,700 3,280 6,930 10,500 20,800
40 1,560 3,010 6,350 9,650 19,100
50 1,440 2,780 5,860 8,920 17,600
60 1,350 2,600 5,490 8,350 16,500
70 1,270 2,450 5,160 7,840 15,500
80 1,200 2,320 4,900 7,450 14,700
90 1,140 2,210 4,660 7,090 14,000
100 1,090 2,100 4,430 6,740 13,300
120 1,000 1,935 4,080 6,210 12,300
140 940 1,800 3,810 5,790 11,400
160 870 1,690 3,560 5,420 10,700
* Calculated with the Mueller Equation for a flowing gas temperature of 400F and for a A P2 of 40,
which corresponds to a pressure drop of 1 psig at a pressure level of 5 psig. The gas properties are:

G = 0.60 m = 6.72 X 10-6 lbm/ft s

** Thirty feet was added to the length of each service as the equivalent length of the fittings.

      Schedule-40
 
 
Table 4-12. Gas Service Capacity for Plastic (6)
Cubic Feet per Hour (0.6 Specific Gravity gas)
High Pressure - 2.0 psig Pressure Drop
Service Size: ½" CTS 1" PS 1~A' IPS 2" IPS 3" IPS 4" IPS

Total Actual (0.445") (1.033") (1.304") (1.865) (2.747") (3.536)

Length, ft ½" Tee 1" Tee 2" Tee 2" Tee 3" Tee 4" Tee
Plastic Service Off Plastic Main
£30 760 4,410 10,500 23,600 70,000 102,100
31 to 60 620 4,030 9,200 21,300 60,000 95,900
61 to 100 470 3,490 7,600 18,200 54,000 86,000
101 to 150 360 2,910 6,090 15,000 47,000 74,400
151 to 200 300 2,540 5,180 13,000 41,700 66,300
201 to 250 260 2,270 4,570 11,600 37,100 60,200
Do not add fitting allowance to the above lengths; the capacities include an allowance for fittings.
Example:
Twenty residential gas customers are served by a 4-in. (102-mm) stub main 500 ft (152 m) long in
a low-pressure distribution system. Design load studies have indicated that the average load per
customer at the time of the peak usage rate for the system is 100 CF/h. Estimate the pressure
drop for this main by use of the Spitzglass low-pressure flow equation.
Total load = 20 X 100 = 2000 CF/h or 2 MCF/h
hw = KQb2
From Table 4-4, R4 = 1.017 X 10-4
Solution:
Assume that the load per customer and the distance between services are approximately the
same; then, one-third of the length of the line, rather than the full length, should be used to
calculate the flow equation resistance factor. Thus:
K  = RL = 1.017X10-4 X 500/3
= 0.0167
hw  = 0.0167 (2)2
= 0.07 in. wc
 

Undersized Mains in a Network.

   When undersized mains are present in a distribution system, large pressure losses may be
incurred in mains only one block long and fed from both ends.
   When gas loads in a pipe network are concentrated at the end of each block-. as is commonly
done in developing pipe-system loads for network analysis. the pressure losses in such mains are
not computed when the flow distribution in the network is calculated. Thus, separate calculations
must be made of the pressure losses along smaller mains fed from both ends, or from both ends
and the middle, to be sure that excessive pressure losses are not incurred along them. Values of
the flow rate and main length that should be used in these calculations for two uniform flow
configurations are given in Figure 4-7. The tabulated lengths are one-third of the distance from
input point to no-flow point.
 
 
Fig. 4-7. Flow Rates and Main Lengths for  
Two Uniform Flow Configurations
Example:
A four-inch (102-mm) low-pressure main, 1200 ft (365.8 m) long, that serves a uniformly
distributed residential load of 6 MCF/h is fed from both ends. Estimate the pressure drop for this
main for gas with a gravity of 0.65; use the Spitzglass low-pressure equation.
From Table 4-4, the low-pressure Spitzglass equation resistance factor for 4-in. pipe carrying a
0.65 gravity gas is 1.017 X 10-4. From Figure 4-7, one-sixth of the length of the line should be
used to calculate the flow equation resistance factor, and the flow rate should be taken as one-
half the total distributed load. Thus:
K  = RL = 1.017X10-4 X 1200/6
= 0.0203
hw  = 0.0203 (6/2)2 = 0.0203(9)
= 0.183 in. wc
Example:
Estimate the pressure drop for the main of the previous example if it were fed from the middle as
well as from both ends.
When a feed in the middle is added, Figure 4-7 indicates that one-twelfth of the length of the line
should be used to calculate the flow equation resistance factor, and the flow rate should be taken
as one-fourth the total distributed load. Thus:
K  = 0.02034/2
K  = 0.01017
hw  = 0.01017 (6/4)2 = 0.01017 (9/4)
K  =  0.023 in. wc
 

SUPPLEMENT FOR THE ADVANCED

READER
 
Fundamentals
   Because all basic textbooks on fluid mechanics published in the last three decades provide
thorough coverage of gas flow behavior, it. s important to review these essential fundamentals.

Frictional Effects

   As a fluid flows through a conduit, the flow is retarded by friction at the wall surface and by the
internal friction of the fluid particles moving past each other. As a result of the friction at the wall
surface, the average velocity component in the direction of flow of a fluid particle at the wall is
zero. This average velocity component increases with distance from the wall to a maximum in the
center of the conduit as shown in Figure 4-8. As noted in the basic chapter text, viscosity is the
gas property that determines its resistance to flow and the shape of its flow velocity profile.
Fig. 4-8. Velocity Profile of Fluid in Laminar Flow Through a Pipe
 
 
   Metric units of viscosity and the factors for converting them to English units are Table 4-13.
Example:
To convert a gas viscosity of 105 m-poises to lbm/ft s, enter Table 4-13 on the line for 1 m-poise
and read the multiplier conversion factor from the lbm/ft s column:
105 m-poises X (6.72 X 10-8)
= 7.056 X 10-6 lbm/ft s
 
Table 4-13. Conversion Factors for Viscosity Units
       
lbm/ft s lbf s/ft2 Centi-poise   m-poise 
gm/cm s X 10-2 gm/cm s X 1006
1 32.16 1.488X103 1.488X107
3.108X10-2 1 46.27 4.627X10-5
6.72X10-4 2.16X10-4 1 1X104
6.72X10-8 2.16X10-6 1X10-4 1
 
Source: Hyman, S. I., M. A. Stoner, and M. A. Karnitz. "Gas Flow Formulas . An
Evaluation." Pipeline and Gas J. 202 (December 1975): 34-44; ibid, Part 2, 203 (January
1976): 28-34.

Types of Flow Behavior

   As defined in the basic chapter, the two basic types of flow behavior that occur in tubing,
laminar and turbulent, were defined by the classic experiments performed by Osborne Reynolds.
They are diagrammed in Figure 4-1. The results of Reynolds' experiments, published in 1883,
established the Reynolds number as an index of the type of flow to be expected for given flow
conditions in a pipe:
 
   Re = DU r/m (Eq .4-3)
where: 
Re  =  Reynolds number
D =  internal pipe diameter
U =  average velocity of the fluid stream
r =  fluid density
m =  fluid viscosity
 
   The dimensionless character of the Reynolds number is apparent when its terms are expressed
in consistent units. When D is expressed in ft, r in lbm/ft 3, U in ft/s, and m in lbm/ft s:
DU r/m = ft (ft/s) (lbm/ft3) (ft s/Ibm)
and all dimensions cancel.
   At Reynolds numbers up to approximately 2000, flow through a cylindrical conduit will always
be laminar. At Reynolds numbers above 2000, flow becomes unstable. The transition to turbulent
flow occurs in the Reynolds number range from 2000 to 13,000, commonly near the lower end of
this range; however, when precautions have been taken to eliminate disturbances in the fluid
stream and isolate it from vibrations, laminar flow has been maintained up to Reynolds numbers
as high as 26,000.
   From 1900 to 1930, Prandtl and his co-workers studied turbulent flow behavior and developed a
general theory still used in solving turbulent-flow problems. To briefly review this theory already
discussed in the first part of this chapter, the transition from laminar to turbulent flow is not abrupt
and complete. Instead, laminar flow persists in a sublayer of fluid next to the pipe wall well into
the turbulent flow range, as illustrated in Figure 4-2. This flow is termed "partially turbulent." As
the fluid velocity, and consequently the Reynolds number, increases, the laminar sublayer is
assumed to become thinner and the turbulent core larger in diameter. In pipe with rough walls,
such as commercial pipe, the thinning of the sublayer is assumed to continue until its thickness
becomes of the same order of magnitude as the height of the projections of the pipe wall into the
gas stream, as shown in Figure 4-3. When this point is reached, the turbulence is said to be fully
developed, and the flow is termed "fully turbulent."
   These assumed flow patterns may not exactly represent actual behavior; however, extensive
tests conducted in the laboratory and in the field have firmly established the existence of the two
distinct types of turbulent-flow behavior. The data obtained in these tests have also been closely
described by the flow relationships derived from Prandtl's assumed flow patterns.

Evaluation of Frictional Energy Loss

   When gas flows through a pipe at a steady rate in laminar flow, its velocity is not the same at all
points across the cross section of the pipe. The classical parabolic velocity profile for laminar flow
was shown in Figure 4-8. The gradual change in the fluid velocity from the wall to the center of
the pipe results from the friction between adjacent layers of the flowing gas.
   As gas flows through a pipe, the combined effect of friction at the wall and fluid friction results in
the conversion of compression energy first into kinetic energy and then into heat energy. In the
latter form, energy is lost by the gas stream to the earth through the pipe wall. The amount of fluid
friction depends on the viscosity of the fluid.
   Developing a flow equation that relates the pressure drop to the flow variables and the gas and
pipe properties required a way to express the frictional energy loss. Inductive logic supported by
early flow experiments indicated that this energy loss in steady-state flow was directly
proportional to the kinetic energy of the flowing fluid and the wetted perimeter surface area of the
conduit; it was inversely proportional to the cross-sectional area of the conduit. The mathematical
_expression of this statement is as follows:
 
 FaU2Aw/2gcA
Where:
F  =  frictional energy loss per unit mass of fluid
U2/2gc  = average kinetic energy of the fluid
Aw = wetted perimeter surface area of the conduit
A  = cross-sectional area of the conduit
For a cylindrical conduit such as pipe:
Aw  =  pDL
A  =  pD2/4
F a2LU2/gcD
   To replace the proportional sign with an equals sign, a constant of proportionality, the friction
factor, f, must be introduced:
F  =  2fLU2/gcD
Two friction factors are used widely: the Fanning friction factor, f, and the Darcy-Weisbach friction
factor, fD = 4f. The friction factor is an engineering factor; its value must be adjusted according to
flow conditions to make the equation valid. The value of the friction factor is not constant; it must
be evaluated empirically through correlation of the results of extensive flow tests both in the
laboratory and in the field.
The General Flow Equation
   In the development of flow equations for use in pipe design calculations, the rate of mass flow is
assumed to be constant. Strictly speaking, this assumption is not valid in gas distribution systems
because the loads, and thus the flow rates, are continuously changing. In sizing gas piping,
however, it is only the maximum flow rate or design load that is of any practical concern. In typical
distribution systems, this load is apparently of sufficient duration or the line pack in small-diameter
 distribution systems sufficiently small to validate the assumption of a steady flow rate. Results of
flow calculations based on this assumption have agreed fairly well with actual performance.

Development

   Once steady-state flow is assumed, it is possible to set up a simple differential equation to show
the relationships of the flow variables. This is accomplished by writing a Bernoulli energy balance
over a small element of fluid in flow. To simplify the solution of this differential equation, the
following assumptions are made:
 The temperature of the gas is constant.
 Changes in the kinetic energy of the gas between the inlet and outlet are negligible.  
 There are no elevation changes.
 The gas does not pass through a compressor or expansion engine.
 The deviation from ideal gas behavior is essentially constant over the length of the line.

   All of these assumptions are valid For most distribution-system flow calculations, all of these
assumptions are valid, except for the third. no elevation changes. Methods of correcting for
elevation effects are outlined later in this supplement.
The solution of the simplified energy balance equation, known as the general flow equation, is:
   Qb = C  (Eq. 4-4)
Where:
Qb  = gas flow rate (CF/h*, m3/h)
C  = a constant whose value depends on the units of the other terms in the equation; C = 117.3
for the English and 0.23944 for the metric units specified
Tb  = absolute base temperature (ºR = temperature in ºF + 460º; ºK = ºC + 273º)
Pb  =  base pressure (psia; kPa)
P1  =  upstream pressure (psia; kPa)
P2  =  downstream pressure (psia; kPa)
D  =  inside pipe diameter (in.; cm)
G  =  specific gravity with respect to air (dimensionless)
Tf  =  average absolute temperature of flowing fluid (ºR, ºK)
Zavg = average compressibility factor of fluid at flowing conditions (dimensionless). At operating
pressures of 100 psia (689 kPa) and less, the value of this factor for natural gas can be taken as
1.0.
L  =  distance between upstream and downstream points at which P 1 and P2 are measured (ft,
meters)
f  =  Fanning friction factor (dimensionless)
   The constant, C, includes several unit-conversion factors, so that a specific value applies only
for the set of units specified. If any of the quantities to be substituted into a flow equation are in
units other than those specified for the quantity, either its value must be converted into the units
specified or the value of the constant must be changed. Values of the constant, C, for several
common sets of flow-variable units are given in Table 4-14.
 
Table 4-14. Values of Flow Equation Constant, C, for Various Units
UNITS SPECIFIED C
            With Darcy-
          With Fanning Weisbach Friction
Pressure Temp. Diameter Length Flow Friction Factor Factor
psia ºR Inches Miles CF/day 38.77 77.54
psia ºR Inches Miles CF/h 1.590 3.180
psia ºR Inches Feet CF/day 2817 5634
psia ºR Inches Feet CF/h 117.4 234.8
kPa* ºK Meters Meters m3/day 574.7X103 114.9X104
kPa* ºK Meters Meters m3/h 239.4X102 478.8X102
kPa* ºK cm Meters m3/h 0.23944 0.47888
in. wc ºR Inches Feet CF/day 101.7** 203.4
in. wc ºR Inches Feet CF/h 4.238** 8.476
in. wc ºR Inches Miles CF/day 1.400** 2.800
in. wc ºR Inches Miles CF/h 0.0583** 0.1166
* kilo Pascal = 1000 Newtons per square meter (1 psi = 6.8948 kPa).
**Using (P12-P22) in in. wc at 60ºF and assuming the flowing pressure = P b.
Source:  Hyman, S. I., M. A. Stoner, and M. A. Karnitz. "Gas Flow Formulas . An
Evaluation." Pipeline and Gas J. 202 (December 1975): 34-44; ibid, Part 2, 203 (January
1976): 28-34.
 
   The only practical problem in the general flow equation is in evaluating the friction factor f. As
the general flow equation is usually written, the friction factor appears in the form . This quantity is
called the transmission factor. For a specific gas, section of pipe, and set of operating conditions,
the gas flow rate is proportional to the value of the transmission factor, whose behavior and
evaluation now need to be described.

Evaluation of the Friction/Transmission Factor

   The practical flow equations used in the gas industry all have the same form as the general flow
equation, although some of the older equations use some of the terms into the constant C: base
pressure, Pb; temperature, Tb; and flowing gas temperature, Tf. The principal difference between
the equations lies in the evaluation of the friction or transmission factor.
 
   In the older equations, such as those of Pole, Spitzglass, and Weymouth, the friction or
transmission factor was evaluated empirically on the basis of gas flow tests in pipe over a limited
range of conditions. Today, knowledge of the behavior of the friction factor rests on very solid
technology. Von Karmen used Prandtl's turbulent-flow model to derive the form of the friction-
factor expressions for the two types of turbulent flow . partially turbulent (the smooth-pipe flow
law) and fully turbulent (the rough-pipe flow law). These expressions were confirmed first in
laboratory flow tests conducted by Nikuradse in 1933, then by flow tests conducted by the U.S.
Bureau of Mines on 3-8 in. (76-203 mm) commercial pipe, and finally by field flow tests conducted
in the 1950s on large diameter operating pipelines supervised by the Institute of Gas Technology.

Moody Diagram

   Textbook presentations on fluid mechanics summarize the behavior of the friction factor in
terms of a Moody diagram, shown in Figure 4-9, which plots the Darcy-Weisbach friction factor,
fD, against Reynolds numbers. The short, straight-line segment at the left represents laminar flow.
The curving line labeled "Smooth Pipes" represents the friction factor for partially turbulent flow at
flow rates below the level at which the wall roughness has an effect and in conduit with no
additional turbulence-inducing factors such as weld beads or bends. It defines the lower limit of
the friction factor for each Reynolds number.
   The horizontal-line segments to the right of the curving dashed line represent the friction factor
for fully turbulent flow. These friction factors are independent of the Reynolds number and
depend only on the relative roughness of the pipe:
e = k/D
where:
e  = relative roughness (dimensionless)
k  =  effective roughness of the pipe wall [commonly measured in mils (10 -3 in.) or m inches (10-6
in.)]
D  =  inside pipe diameter (in.)
   The effective roughness can be visualized as the depth of the surface depressions or the
average height of the projections of the wall surface into the flow stream in a straight section of
pipe with a uniform wall roughness. It would have the same flow behavior as an actual pipe
section with its varying wall roughness, weld beads, and bends. The units of k and D must be the
same. The direct measurement of surface roughness is difficult because the instruments for doing
so sample only a very small area.
   The upward curving extensions of the horizontal-line segments to the left of the curving dashed
line in Figure 4-9 are plots of the Colebrook-White friction-factor _expression, which represents
an effort to summarize, in a single _expression, the complete range of turbulent flow friction
factors. At relatively low Reynolds numbers for each pipe diameter, the _expression closely
approaches partially turbulent flow behavior; at high Reynolds numbers, it closely represents fully
turbulent flow behavior. The smooth- and rough-pipe law limits are based on flow test data. The
curving Colebrook-White lines result simply from the mathematical form of the _expression. The
actual behavior of the friction factor in the transition zone shown in Figure 4-9 does not follow a
consistent pattern and cannot be defined precisely, just as the transition from laminar to turbulent
flow does not always occur at the same Reynolds number.
Fig. 4-9. Moody Diagram of Friction Factor
Source:  Moody, L. F. "Friction Factors for Pipe Flow." Transactions ASME 66
(November 1944): 671-84; Chem. Eng. 64 (November 1944): 671-84.
   If the transmission factor is plotted as a function of Reynolds number on semi-log paper, the
diagram is inverted and the plot for partially turbulent flow is nearly a straight line, as shown in
Figure 4-10. This diagram does not show the Colebrook-White curves. Instead, the horizontal
 lines representing fully turbulent flow simply are extended to intersect the plot for partially
turbulent flow.
 
Fig. 4-10. Plot of Transmission Factor Vs. Reynolds Number

Transmission Factor for Laminar Flow

   From Newton's law of viscosity and the definition of the friction factor, the following _expression
for the laminar low transmission factor, called the Hagen-Poiseuille relationship, can be derived:
=  =
where: 
f =  16/Re or fD = 64/Re
(Also see definitions for Equation 4-3.)

Transmission Factors for Turbulent Flow

   The form of the transmission-factor expressions for the two types of turbulent flow were derived
by Prandtl and Von Karman based on Prandtl's models of turbulent flow behavior. The unknown
constants that appeared in these relationships were originally quantified from the flow data of
Nikuradse and from data obtained by the U. S. Bureau of Mines (the values used here) to obtain
the following flow expressions:
   For partially turbulent flow (the smooth-pipe flow law) .
   = 4 log Re -0.6 (Eq. 4-5)
   For fully turbulent flow (the rough-pipe flow law) -
   = 4 log (3.7 D/k) (Eq. 4-6)
   The smooth-pipe and rough-pipe flow laws closely approximate the flow behavior in gas pipe
systems over the complete range of conditions encountered. They are consistent with flow data
obtained from extensive field tests on 12- to 36-inch (305- to 914-mm) gas transmission pipelines.
These data were obtained by IGT during a research program conducted from 1954 to 1960 that
are still valid today. In these tests, deviations between predicted and actual smooth-pipe behavior
of about five percent were observed. The deviations are believed to have been caused by the
extra turbulence generated at bends, fittings, and weld beads.
   The smooth-pipe flow law is a limiting law for turbulent flow in pipelines. It gives the maximum
possible value of the transmission factor at each Reynolds number. Because of the turbulent
effect of fittings and weld beads in most distribution-system piping, actual transmission factors for
partially turbulent flow can be expected to be a few percent less than the values computed by the
smooth-pipe flow law.
   At high flow rates or Reynolds numbers, flow behavior in rough pipe is described accurately by
the rough-pipe flow law. This law is represented by the horizontal lines in Figure 4-9, with each
line corresponding to a specific value of the ratio k/D=e, the relative roughness of the pipe.
To illustrate the relationships between the rough-pipe and smooth-pipe flow laws (see Figure 4-
10), consider a pipeline with a relative roughness value, e, of 1.17 X 10-4. This pipeline could be
 10 inches (254 mm) in diameter with an average wall roughness of 1.17 mils (0.03 mm)
or
 4 inches (102 mm) in diameter with an average wall roughness of 0.47 mil (0.012 mm) or
 any other combination of diameter and roughness that corresponds to a relative
roughness of 1.17 X 10-4. 

   From Reynolds numbers of about 4000 to 800,000, where the smooth-pipe flow law intersects
the rough-pipe flow law at a relative roughness of 1.17 X 10 -4, flow behavior in the pipeline will be
approximated closely by the smooth-pipe flow law. At Reynolds numbers greater than 800,000,
flow behavior in the pipeline will follow the rough-pipe flow law for a relative roughness of 1.17
X 10-4, and the transmission factor will assume the constant value 18.0.

Field Measurement

   For simple pipelines such as utility transmission lines and longer feeder mains with metered
input and deliveries, the actual friction factor under flowing conditions can be calculated from
measurements of the other flow variables. Calibrations of all meters and the use of sensitive,
calibrated pressure gages (such as dead-weight gages) are necessary to obtain accurate results.
Best results are obtained in situations with high pressure drops, minimum branch line deliveries,
and steady flow conditions.
The flow-rate measurements needed for field evaluation of friction factors are rarely available in
gas distribution systems because it is seldom possible to control the flow rates at a constant level
for a long enough time to get accurate results.
Practical Flow Equations
   Since the beginning of the gas industry early in the 19th century, gas engineers have developed
and used practical equations to express the relationships between flow variables and pipe
dimensions. The early flow equations were based on flow studies over a limited range of
conditions, and, because of this, their application was limited. Many of these early equations are
no longer used.
The equations commonly used today for making distribution-flow calculations are listed in Table
4-1. All of these equations have a similar form because all can be obtained from the general flow
equation by substituting into it the transmission-factor _expression in the third column.

Selection of a Flow Equation

   Studies of flow rates and Reynolds numbers obtained from network analyses of distribution pipe
systems have established that turbulent flow predominates at design-load conditions. Thus, the
flow equations used should closely represent one of the two types of turbulent flow behavior. This
can be done most directly by using either the rough-pipe or the smooth-pipe flow law as
appropriate to quantify the transmission factor in the general flow equation. It can also be done by
using a flow equation based on a transmission factor _expression that approximates either the
rough-pipe or smooth-pipe flow law.

Equations for Fully Turbulent Flow

   When the rough-pipe flow law is used to evaluate the transmission factor in the general flow
equation, the fully turbulent flow equation results. This is the first equation in Table 4-1. Its use
requires the effective wall roughness of the pipe, which is discussed in the basic chapter.
   To represent turbulent flow fully and effectively, an equation must be based on a transmission-
factor _expression that is not a function of the flow rate or Reynolds number. The type of
_expression that is commonly used yields a constant value of transmission factor for each
diameter or effective roughness. In addition to the "Fully Turbulent" equation, other equations of
this type in Table 4-1 are the Spitzglass and the Weymouth equations.
   The Spitzglass Equation is based on the right form of transmission-factor _expression to
represent fully turbulent flow behavior; but it is almost never used for that purpose. It is used
primarily to make flow calculations in cast-iron low-pressure distribution systems where partially
turbulent flow predominates. (Its transmission factor _expression was developed in 1912 from
flow tests on large-diameter cast-iron pipes operated at inch-wc pressure and at flow conditions
that probably resulted in partially turbulent flow.)
   The Weymouth Equation is used extensively to represent fully turbulent flow behavior; however,
it is necessary to use an empirically determined correction factor to get a good simulation. The
form of the factor normally used is an efficiency factor, E, that is applied as a multiplier to the flow
rate. The proper value of this factor is about 1.0 for 36-in. (914-mm) diameter pipe. For 30-in.
(762-mm) and smaller pipe, values of the efficiency factor range upward from 1.10 to over 2.00.
Thus, the uncorrected equation predicts a lower flow rate for specified flow conditions than is
actually experienced. Its primary application is for flow calculations for large diameter, high-
pressure pipelines on which field flow tests can be readily performed to evaluate the efficiency
factor and monitor line performance.
   The Weymouth equation has the wrong form to represent partially turbulent flow behavior, and
it requires the application of an efficiency factor of 1.1 to 1.2 for the 2 to 6-inch pipe sizes
commonly used in gas distribution systems. Because efficiency factors cannot normally be
developed for sections of distribution pipe, use of this equation for distribution flow calculations is
not recommended.

Equations for Partially Turbulent Flow

   Because the smooth-pipe flow law is an implicit equation for the transmission factor (the friction
factor appears on both sides of the equation), it cannot be substituted into the general flow
equation to obtain an explicit flow equation that is easy to work with. Instead, it has been found
more convenient to approximate the smooth-pipe flow law with expressions of the form:
 =  aReb
where 
a and b  = empirically determined constants selected to give a good fit of the smooth-pipe flow
law over the range of Reynolds numbers of interest
   The three equations recommended to use for represent partially turbulent flow behavior in
distribution flow calculations are the Mueller equation, the lGT Distribution Equation, and the
Panhandle A Equation (Table 4-1). The transmission factors used in these equations provide,
between them, a fit of the smooth-pipe flow law within 11 percent over a range of Reynolds
numbers from 3000 to about 50,000,000 as shown in Figure 4-11. This more than covers the
complete range of interest for distribution flow calculations. The area of application of each of
these flow equations is summarized in Table 4-15.
 
Fig. 4-11. Comparison of Mueller, lGT Distribution, and Panhandle A Equations'
Transmission Factors With Smooth-Pipe Flow Law
   These guidelines apply for all plastic pipe and clean cast-iron and steel pipe of average wall
roughness, about 0.7 mil (0.018 mm) for steel pipe. Their use should result in a representation of
the smooth-pipe flow law within ±2 percent. More accurate representation of field flow behavior
requires an efficiency or modification factor to drop the transmission factors for these equations a
few percent below those of the smooth-pipe flow law, which is discussed later.
   As noted previously, the Spitzglass low-pressure equation has been used extensively for flow
calculations in low-pressure cast-iron pipe systems, despite the fact that it uses a diameter-
dependent transmission factor _expression. It has provided results that check reasonably well
with experience. Comparison of the transmission factor of the Spitzglass equation with the
smooth-pipe flow law show: that for pipe diameters of 2-12 inches, the Spitzglass equation
closely approximates results obtained through use of the smooth-pipe flow law for pipe sections
that are working the hardest and have the highest pressure drops. At lower flow rates, the
Spitzglass equation is optimistic, because its transmission factors exceed the maximum possible
values of the smooth-pipe flow law.
   The results obtained from use of the Spitzglass equation for low-pressure flow calculations can
be expected to be optimistic. the equation will yield calculated pressures slightly higher than
those that actually will be experienced at specified flow rates. This optimism is usually not
excessive. For 16-in. (406-mm) and larger pipe, however, the results obtained with the Spitzglass
equation become unacceptably conservative, and it should not be used for low-pressure flow
calculations for 16-in (406-mm) and larger pipe. Instead, the IGT Distribution Equation should be
used.
 
Table 4-15. Areas of Application of the Mueller, lGT Distribution, and Panhandle A
Equations for Distribution Flow Calculations
      Range of Applicability for Partially Turbulent
Flow Calculation
Equation Range of Re for Fit of Diameter, in. Pressure
SPL, 1000s
Mueller 1%, 3-80 3/8-6 in. wc

2%, 2-125 3/8-2 2-20 psig

3/8-1 ½ 20-100 psig

IGT Distribution 1%. 25=17 3-30 in. wc

2%. 17=3000 1 ½-20 2-20 psig

¾-20 20-100 psig

 Panhandle A 1%, 3,300-50,000 16+in. 20+psig

2%, 1,300-75,000
    

Determination of Type of Turbulent Flow Behavior

   The key consideration for selecting a flow equation for calculating distribution flow is the type of
turbulent flow behavior expected for the flow conditions specified. Available information on pipe
wall roughness, data on Reynolds numbers from network analysis of distribution pipe systems,
and comparisons of measured pressures and pressures calculated by the practical flow equations
have established the following guidelines:
 Plastic pipe: It is probable that the type of flow is always partially turbulent over the
complete range of distribution flow conditions.
 Clean cast iron pipe: It is likely that partially turbulent flow is nearly always present in low-
pressure gas distribution systems because of the relatively low-pressure gradient (rh/L) that can
be tolerated.
 Clean steel: Partially turbulent flow will predominate in pipe [average wall roughness no
greater than 0.7 mil (0.018 mm)] in distribution systems with source pressures of up to 30 psig
(207 kPa).
 Steel: Both partially turbulent and fully turbulent flow behavior can be expected in
distribution systems operated at source pressures greater than 30 psig, either because of the
relatively high gradients and flow rates or because of a rougher than normal pipe wall roughness.
Fully turbulent flow will be present in the pipe sections that are working the hardest as indicated
by above-average gradients, rP2/L, or relatively high pressure drops, rP2.
 Cast iron or steel: Accurate representation of the flow behavior in dirty sections requires
empirical adjustment of the flow calculations.

 
   It is not possible to choose the proper flow equation unless the type of turbulent flow that will
occur has been determined. A good, practical criterion for this purpose is the critical Reynolds
number, i.e., the Reynolds number at which the rough-pipe and smooth pipe flow laws intersect,
as was shown in Figure 4-10. Figure 4-12 is a chart from which the critical Reynolds number can
be read as a function of internal wall roughness and nominal pipe diameter.
Values of the critical Reynolds number for common sizes of steel pipe with an average wall
roughness of 0.7 mil (0.018 mm) and at the maximum wall roughness normally encountered in
commercial steel pipe. 2.0 mils (0.051 mm) . are in Tables 4-17 and 4-18. At Reynolds numbers
below the critical values, partially turbulent flow can be expected; at Reynolds numbers greater
than the critical values, fully turbulent flow can be expected.
Table 4-16. Values of the Critical Reynolds Number and Corresponding Transmission
Factor for Steel Pipe of Average Wall Roughness (k = 0.7 mil)
Nom. Pipe       Re at Intersection
Diameter (inches) Schedule Number Actual I.D. (inches) at Intersection (thousands)
1 40 1.049 14.98 117
1¼ 40 1.380 15.45 159
1½ 40 1.610 15.77 189
2 40 2.067 16.15 249
3 40 3.068 16.84 385
4 40 4.026 17.31 520
6 40 6.065 18.02 816
8 40 7.981 18.50 1,100
10 20 10.250 18.94 1,450
12 20 12.250 19.25 1,760
16 20 15.376 19.64 2,250
20 20 19.250 20.03 2,880
24 20 23.250 20.36 3,530
30 20 29.000 20.74 4,490
Source: Wilson, G. G. and R. T. Ellington. "Selection of Flow Equations for Use in
Distribution System Network Calculations." AGA Operating Section Proceedings 1958:
D231-42.
 
Table 4-17. Values of the Critical Reynolds Number and Corresponding Transmission
Factor for Rougher Than-Normal Steel Pipe (k = 2.0 mil)
Nom. Pipe        
Diameter (Inches) Schedule Number Actual I.D. (inches) 1/f at Intersection Re at Intersection
(thousands)
1 40 1.049 13.15 36.1
1¼ 40 1.380 13.63 49.1
1½ 40 1.610 13.90 58.5
2 40 2.067 14.33 77.4
3 40 3.068 15.02 120.4
4 40 4.026 15.49 163
6 40 6.065 16.20 257
8 40 7.981 16.68 348
10 40 10.028 17.07 447
12 20 12.250 17.42 557
16 20 15.376 17.80 708
20 20 19.250 18.21 914
24 20 23.250 18.53 1127
30 20 29.000 18.92 1430
Source: Wilson, G. G. and R. T. Ellington. "Selection of Flow Equations for Use in
Distribution System Network Calculations." AGA Operating Section Proceedings 1958:
D231-42.
   Use of this critical-Reynolds-number criterion ignores departure of partially turbulent flow
transmission factors from the smooth-pipe flow law in the field. It also represents flow behavior in
the transition region by the rough-pipe flow law. Because no consistent pattern has been
established for the behavior of the friction factor in the transition region and because the
departure from the rough-pipe flow law is small, this is a reasonable assumption. Indeed, the
simple and practical critical-Reynolds-number criterion for type of turbulent flow is sufficiently
accurate that further refinements are probably not warranted.
 
   Use of the critical-Reynolds-number criterion for type of flow requires evaluation of the
Reynolds number. This is facilitated by expressing it in terms of the flow variables commonly
used:
   Re = (0.011459 QbGPb)/(mDTb) (Eq. 4-7)
   All of the terms have the same meaning and units as those used in the general flow equation
(Equation 4-4). The following is an example of the use of the critical-Reynolds-number criterion.
 
Example:
A nominal 6-in. (152-mm) steel main is proposed to carry gas a distance of one mile from a large-
diameter utility supply main to a small city. This line is to be sized for the maximum gas demand
expected in the community in five years. The design engineer has estimated that this load, Q b will
be 250 MCF/h and that the pressure at the inlet of this 6-in. main will be 50 psig. Determine which
flow equation should be used to calculate the pressure drop for this main given the following
additional information:
Inside diameter (ID) of pipe = 6.125 in.
Atmospheric pressure = 14.7 psi
Pb  =  14.73 psia
Tb  =  520ºR
Tf = 490ºR
Specific gravity of gas, G = 0.671
Viscosity of gas, m = 7.23 X 10-6 Ibm/ft s
   Assuming that the main will be clean commercial steel pipe with a wall roughness of 0.7 mil,
entering Figure 4-12 at a wall roughness of 0.7 mil, moving vertically up to a pipe diameter of 6
inches, and then moving horizontally to the left, one finds the critical-Reynolds-number
corresponding to the breakpoint between partially turbulent and fully turbulent flow to be about
750,000.
   The Reynolds number is calculated as follows:
Re  =  0.011459 Qb GPb/mDTb
= 
= 1,230,000
   Because this Reynolds number exceeds the breakpoint value of 750,000, fully turbulent flow
behavior can be anticipated. Therefore, the Fully Turbulent flow equation in Table 4-1 will best
represent flow behavior.
 
 
Fig. 4-12. Plot of Critical Reynolds Number as a Function of Nominal Pipe and Internal Wall
Roughness
Source:  UhI, A. E., et al., "Steady Flow in Gas Pipelines," Technical Report No. 10.
Institute of Gas Technology, 1965.
   If many flow calculations are to be made for the same gas, as in network analysis, it is
convenient to express the critical Reynolds number for each pipe diameter in terms of a critical
flow rate (Qcrit) This is done easily by substituting the gas properties into Equation 4-7. For each
diameter and associated critical Reynolds number, the corresponding critical flow rate can be
calculated. In this manner, a table of critical flow rates can be prepared for each diameter of
interest. To determine whether flow behavior is partially turbulent or fully turbulent for specified
flow conditions, it is necessary only to compare the flow rate against the critical flow rate for the
specified diameter. If it is less, flow is partially turbulent; if it is greater, flow is fully turbulent. The
equation used for the flow calculation should be chosen accordingly.
Example:
Gas is fed into a 1-mile (1.6-km) section of pipe. Values of the base conditions and gas properties
are as follows:
Pb = 14.73 psia
Tb = 520ºR
G  = 0.60
m  = 7.23 X 10-6 Ibm/ft s
Determine the type of flow to be expected for clean steel pipe for the following conditions:
    
Pipe Diameter (in.)   
NPS ID Qb (MCF/h)
4 4.124 65
6 6.125 200
8 8.125 350
First, the Reynolds number equation is expressed as a function of flow rate and diameter, as
follows:
Re = 0.011459 QbGPb/mDTb
= (0.011459 X 0.6 X 14.73)Qb
(7.23 X 10-6)(520) (D)
= 26.934 Qb/D
Next, the critical Reynolds numbers for 4-, 6-, and 8-in. (102-, 152-, and 203-mm) pipe can be
read from Figure 4-12, assuming an average wall roughness of 0.7 mil, or the following more
accurate values can be obtained from Table 4-16:
    
 D (in.) Crit Re
4 520,000
6 816,000
8 1,100,000
If the Reynolds number-_expression is solved for the flow rate, it becomes:
   Qb = D Re/26.934
By substituting the appropriate diameters and critical Reynolds number values into this equation,
the critical flow rates now can be determined as follows:
 
    
  Q Crit (MCF/h)
D (in.)
4 79.62
6 185.56
8 331.83
    
Comparison of the specified flow rates with the critical flow rates indicates that the flow in the 4-
inch main would be partially turbulent and the flow in the 6- and 8-inch mains would be fully
turbulent. Thus, the IGT Distribution Equation should be used for the 4-inch calculation and the
Fully Turbulent equation for the 6- and 8-inch calculations.
Calculation Procedures
   The use of resistance factors and equivalent lengths has been described in the basic chapter.
All of the examples given there are based on the use of the Spitzglass high and low pressure
equations which have limited application in distribution system flow calculations. The same
procedure can be used with the other flow equations given in Table 4-1 and resistance factor
expressions for them are given in Table 4-2. Data on inside pipe diameters are also given in
Tables 4-6 to 4-8 in the basic chapter.

Equivalent Resistance Factors

   Formulas can be readily developed to express the equivalent resistance factor of several pipe
sections connected in series or in parallel as a function of the resistance factors for the individual
pipe sections.
   For mains connected in series, as shown in Figure 4-13:
Pi2 - P12  =  K1 Qtn
P12 - P22  =  K2 Qtn
P22- P02 = K3Qtn
Pi2 - Po2  =  (K1+K2+K3)Qtn=KeQn
\ Ke  =  K1 + K2 + K3
Fig. 4-13. Pipe Sections in Series
 
   For mains connected in parallel, as shown in Figure 4-14, it can be shown that:
Qt  = Q1+Q2
(1/Ke)1/n = (1/K1)1/n + (1/K2)1/n
When this equation is solved for Ke, it becomes:
   Ke = K1 K2/(K1 + K2) (Eq. 4-8)
   Using equivalent resistance factors obtained in this manner simplifies the solution of more
complicated flow problems by reducing the number of separate pipe sections that must be
considered.
Fig. 4-14. Pipe Sections in Parallel
Example:
Calculate the equivalent resistance factors and equivalent lengths for the Spitzglass low-pressure
equation for each of the cast-iron piping layouts diagrammed in Figure 4-15. Use 6-inch pipe as
the base; specific gravity of the gas = 0.65.
 
Fig. 4-15. Cast-Iron Pipe Layouts
 
    
   Determining the equivalent resistance factors requires calculating the resistance for each
individual section of pipe. This is done easily by use of the resistance factors given in Table 4-4:
    
    
Pipe Diam. (in.) R L (It) K
4 1.017X10-4 309 0.0314
6 1.140X10-5 50 5.7X10-4
8 2.419X10-6 309 7.47X10-4
12 3.276X10-7 374 1.23X10-4
12 3.276X10-7 500 1.64X10-4
    
   The equivalent resistance for the first layout is the sum of the resistance factors for each
section of pipe:
Ke  =  8.57 X 10-4
The equivalent length of 6-inch pipe is obtained as follows:
L6  =  8.57 X 10-4/1.140 X 10-5 = 75.2 ft
The equivalent resistance for the second layout is obtained by substitution into Equation 4-8:
Ke  = (3.14 X 10-2)(7.47 x 10-4)/[(3.14 x 10-2)½ + (7.47 X 10-4)½]2
= 2.35 X 10-5/(0.1773 + 0.0274)2 = 5.62 X 10-4
L6 = 5.62 X 10-4/1,140 X 10-5
= 49.2 ft
    
   Data on the capacity of various pipe sizes help provide designers with an idea for the size
diameter required to carry specified flow rates. Table 4-18 shows the capacities of steel pipe for a
range of diameters and gradients. It may seem plausible that pipe sizes add capacity linearly, i.e.,
two 4-inch pipes will have the same capacity as one 8-inch. Examination of the values in TABLE
4-18 shows that this is not true. At a gradient of 1000 rp 2/mi, two, parallel 4-inch mains would
have a capacity of 2 X 63.467 = 126.934 MCF/h whereas one 8-inch has a capacity of 366.963
MCF/h.
 
Table 4-18. Gas Flow Rate as a Function of Diameter and Gradient *
Gradient, GAS FLOW RATE, MCF/H
r p2/ Mile 1.25. 2.125. 3.218. 4.218. 6.250. 8.25. 10.25. 12.25. 15.50.
300 1.264 5.225 15.800 32.512 92.774 194.512 346.998 558.154 1,030.941f
400 1.489 6.130 18.539 38.147 108.853 228.225 407.139 647.031f 1,190.429
500 1.686 6.939 20.985 43.182 123.221 258.348 455.699f 723.402 1,330.940
600 1.866 7.679 23.223 47.785 136.357 284.248f 499.193 792.448 1,457.971
700 2.032 9.366 25.299 52.058 148.550 307.023 439.190 855.941 1,574.780
800 2.189 9.010 27.248 56.067 159.618f 328.222 576.419 915.040 1,683.520
900 2.337 9.620 29.090 59.859 169.301 348.132 611.384 970.546 1,785.643
1,000 2.478 10.199 30.844 63.467 178.459 366.963 644.456 1,023.046 1,882.233
1,100 2.613 10.754 32.521 66.919 187.169 384.874 675.971 1,072.979 1,974.102
1,200 2.742 11.287 34.132 70.330f 195.491 401.988 705.966 1,120.690 2,061.883
1,300 2.867 11.800 35.684 73.202 203.474 418.402 734.793 1,166.451 2,146.076
1,400 2.987 12.296 37.184 75.965 211.155 434.197 762.530 1,210.484 2,227.088
1,500 f
3.104 12.777 38.874 78.631 218.566 449.436 789.294 1,252.970 2,305.255
1,600 3.217 13.243 40.149 81.210 225.734 464.176 815.179 1,294.062 2,380.857
1,700 3.327 13.697 41.384 83.709 232.681 478.461 840.268 1,333.888 2,454.131
1,800 3.435 14.139 42.584 86.136 239.427 492.333 864.628 1,372.560 2,525.280
1,900 3.539 14.570f 43.751 88.496 245.988 505.824 888.321 1,410.171 2,594.479
2,000 3.642 14.991 44.887 90.795 252.378 518.964 911.398 1,446.805 2,661.879
3,000 4.562 18.634 54.976 111.201 309.099 635.599 1,116.230 1,771.967 3,260.123
4,000 5.378f 21.517 63.480 128.404 356.917 733.926 1,288.912 2,046.091 3,764.466
5,000 6.013 24.056 70.973 143.560 399.045 820.554 1,441.047 2,287.599 4,208.801
*Gas Properties Base conditions Tf = 510ºR Flow Formulas

G = 0.60 Pb = 14.73 psia k = 0.85 mil IGT Distribution

m = 8 x 10-6lbm/ft.s Tb = 520ºR Fully Turbulent

f Gradient level at which fully turbulent flow begins
Source: Mefford, J. J., Oklahoma Natural Gas Company, Tulsa, Oklahoma. Private
communication, 1985.
    

Modification and Efficiency Factors

   Weld beads, bends, and fittings in field pipe systems generate additional turbulence beyond
that resulting from the wall roughness. This increases frictional energy losses in partially turbulent
flow beyond those predicted by the smooth-pipe flow law. Empirical corrections based on field
flow data have long been introduced into gas flow equations by pipeline companies to improve
the match of their flow calculations with field flow behavior. The most common form of correction
factor is a multiplier called the efficiency factor, E, which is introduced into the right side of a
flow equation in which the flow rate is the only term on the left side. Values of this factor are well
established for flow behavior in large-diameter steel transmission pipe.
   Experience in comparing calculated and measured pressures in distribution pipe systems
indicates that distribution-system transmission factors can be expected to deviate by 3-6 percent
below those calculated by the smooth-pipe flow law. This corresponds to efficiency factors of
0.94-0.97 for the general flow equation.
   Because the practical flow equations for approximating the smooth pipe flow law already have
some built-in deviation below the smooth pipe flow law, somewhat lower efficiency factors are
appropriate for these equations. Suggested efficiency factors for the Mueller and IGT distribution
equations are in Table 4-19.
Table 4-19. Suggested Efficiency and Modification Factors for Mueller  
and IGT Distribution Equations
    
Max. Deviation Below Efficiency Factor, E Modification Factor, M
SPL (%)
3 0.980 1.042
4 0.970 1.064
5 0.960 1.086
6 0.950 1.109
7 0.939 1.133
8 0.929 1.158
9 0.919 1.184
10 0.909 1.210
    
   For distribution-system flow calculations, it is more convenient to apply the correction factor as
a multiplier of the resistance factor. It can be shown that the resistance modification factor, M, is
(1/E)2. Suggested resistance modification factors for the Mueller and IGT distribution equations
are included in Table 4-19. Through use of these factors, a specified departure from the smooth-
pipe flow law can be introduced into these equations.
Example:
To represent the flow behavior in a distribution system more closely, it is desirable to modify the
lGT equation so that it deviates a maximum of 4 percent below the smooth-pipe flow law.
Determine the unit length resistance factor that should be used.
The desired unit length resistance factor is obtained by multiplying the IGT equation resistance
factor by the 4-percent maximum deviation modification factor from the third column of Table 4-
19:
R = 1.064 (3.418) X 10-3 TfG0.80 m 0.20/D4.80
= 3.637 X 10-3 TfG0.80 m 0.20/D4.80
This _expression is used to calculate resistance factors for the simplified form of the IGT
equation:
rP2 = RL Qb1.80

Modifications for Dirty and Rough Pipe

   Flow behavior can be affected by rougher-than-normal pipe or by deposits that increase the
roughness of the internal pipe wall and/or reduce the internal diameter. Rougher-than-normal
pipe can result from sub-standard fabrication procedures at the pipe mill, internal corrosion during
storage of steel or cast-iron pipe, or dust deposits. Rougher than normal pipe creates fully
turbulent flow behavior at Reynolds numbers or flow rates lower than the critical values for pipe of
average wall roughness of 0.7 mil (0.018 mm).
   The only way to allow for the effect of rougher-than-normal pipe on flow behavior is to compare
calculated values with measured values of the flow variables such as flow rates and pressures.
This is difficult to do in a distribution system because flow-rate measurements are usually limited
to city gate stations. Thus, flow-rate data are usually calculated values based on load estimates
derived from billing data, as described in Chapter 3. Pressure data can be obtained from
recording pressure gages set temporarily at selected locations, preferably along the routes of
feeder mains and at low-pressure points in the system.
   If the estimated system loads are accurate and the calculated pressures obtained by the
procedures outlined previously in this chapter exceed the measured pressures at a number of
locations by 1 to 2 psi (6.9 to 13.8 kPa) or more, a rougher-than-normal surface in some of the
pipe sections in the system is indicated. Fully turbulent flow is probably in pipe sections where it
was not predicted by the critical-Reynolds-number or flow-rate criterion. Calculation of the flow
rates and pressure drops in the system under study should be repeated with the Fully Turbulent
equation with a 0.7 mil effective roughness for all feeder mains for which the calculated pressure
drops were too low. The calculated pressure still should be significantly higher than the measured
pressures, because abnormally high wall roughness is considered to cause the fully turbulent
flow. The flow distribution of the pipe sections in the network, however, will be a better estimate of
the actual values than the flow distribution from the initial calculation.
   These calculated flow rates, together with the measured pressures along trunk and feeder
mains, can be used to adjust pipe-section resistance factors. The adjusted resistance factors then
should be used in all subsequent flow calculations for the system. They can also be used to
calculate the corresponding effective roughness. The values obtained are a useful guide for
estimating resistance factors for dirty or abnormally rough pipe.
    Sometimes an inspection reveals that the deposits in a main have significantly reduced its
effective diameter. Such a main will have an abnormally high resistance to flow even though the
flow remains partially turbulent. If the flow rate and pressure drop across the main are measured,
the main's effective resistance and diameter can be calculated directly from a flow equation that
closely represents partially turbulent flow behavior. usually the standard or modified lGT equation.
   If a partially plugged main is part of a network system, its flow rate cannot be measured directly.
Instead, the flow rate must be estimated by a trial-and-error procedure. First the network is solved
by use of an equation that closely approximates smooth-pipe behavior and assumes that the dirty
section of pipe is clean. The calculated flow rate is then used to estimate the effective resistance
of the partially plugged main. Finally, the network is analyzed again after the clean-pipe
resistance factor has been replaced with the calculated effective resistance factor. This process is
repeated until the calculated flow rate remains close to the value used to obtain the effective
resistance factor.
   Sometimes the deposits in a main not only decrease the cross sectional area of the pipe but
also roughen the pipe wall. In such cases, the two effects can be separated only by estimating
the effective diameter of the pipe on the basis of inspection at cuts in the line during repair or
construction work. When several feeder mains with enough deposits to affect flow behavior are in
the same network, their resistance factors are adjusted simultaneously by the trial-and-error
procedure for partially plugged mains. Such adjustment of resistance factors requires
considerable engineering judgment based on experience with the system.
Flow Calculation Applications
   Principal applications of flow calculations in distribution-system design:
 Capacity design of distribution pipe systems.
 Reinforcement studies.
 Determination of severity of weather conditions beyond which a line can not be taken out
of service for repair or replacement or to accommodate construction of sewers, highways, etc.
 Sizing of services.
 Sizing of fuel runs.
 Sizing of a temporary long by-pass.

   There are a few circumstances in which the steady-state flow equations is not applicable. In the
derivation of the general flow equation, a number of assumptions are made as outlined in The
General Flow Equation section of this supplement. Three of them. constant gas temperature,
absence of compressors or expansion engines, and constant deviation from ideal gas behavior.
are almost universally valid for gas flow in distribution systems. Elevation effects are readily
corrected for as outlined in this supplement's subsequent section on Elevation Corrections.
However, significant changes in kinetic energy between the inlet and outlet flow conditions make
standard flow calculation procedures using flow equations inapplicable for use in sizing
calculations that involve the following elements:
 Short by-passes
 Relief vents
 Regulator station outlets
 Meter and regulator piping.
   As outlined in the basic chapter, most flow calculations today are performed by pocket
calculator or computer. Examples of two common types of flow calculations are given below to
illustrate the process for the two most widely applicable flow equations. the IGT Distribution
Equation and the fully turbulent equation.

Calculation of Downstream Pressure

Example 1:
Statement of Problem: The atmospheric pressure, base conditions, flowing gas temperature, and
the properties of gas fed into the feeder system of a gas distribution company are tabulated
below:
Pb  = 14.73 psia
Atm. P = 14.7 psia
Tb  = 520ºR
Tf  = 30ºF
G  = 0.60
m  = 7.23 X 10-6 lbm/ft s
Gas is fed into a mile-long (1.61 km-long) section of 4-in. (102-mm) schedule-40 steel pipe at a
pressure of 20 psig (138 kPa), and a flow rate of 30 MCF/h. Compute the downstream pressure
by use of an appropriate equation.
Solution:
To select an appropriate flow equation, the Reynolds number for the flow conditions needs to be
computed and compared to the critical Reynolds number for 4-inch (102-mm) steel pipe. From
equation 4-7:
Re  = (0.011459 Qb GPb)/(mDTb)
= (0.011459 X 0.6 X 14.73) Qb/(7.23 x 10-6) (520)D
= 26.934 Qb/D
The inside diameter of the pipe can be obtained from Table 4-7 as 4.026 inches (102.26 mm).
This value is substituted in the above equation to give:
Re = 26.934 X 30,000/4.026 = 200,700
 
   From Table 4-15, the critical-Reynolds-number for 4-inch (102 mm) pipe of average roughness,
0.7 mil, is 520,000. Therefore, partially turbulent flow can be expected. Table 4-17 indicates that
the lGT distribution equation is a good choice to represent flow behavior at the conditions
specified.
   From Table 4-2, the unit length resistance factor _expression for the lGT distribution equation
is:
R  = 3.418 X 10-3 TfG0.80 m0.20/D4.80
= 3.418 X 10-3(490)(0.60)0.80(7.23 x 10-6)0.20/(4.026)4.80
= (1.777)(0.6645)(0.09372)/800.6
= 0.13025 X 10-3
   The downstream pressure can now be calculated from the simplified form of the equation, for
which the exponent of 1.80 can be obtained from Table 4-3:
rp2 = RLQb1.80
 P12-P22 = (34.7)2 - P22 = 0.13025 X 10-~ (5280)(30)1.80
P22 = 1204 - 0.6877 (455.9)
P22  = 1204 - 313.5 = 890.5
P2  = 29.8 psia
= 15.1 psig
   The pressure drop is 20 - 15.1 = 4.9 psi.

Calculation of Upstream Pressure

   The previous example was solved for the downstream pressure when values of all of the other
flow equation variables were known. Sometimes, the downstream pressure is specified at some
minimum required value, and the upstream pressure needed to maintain this minimum
downstream value is to be calculated. This calculation is essentially the same as that above
except that P2 is known and the upstream pressure, P1, must be solved.

Calculation of Pipe Diameter or Flow Rate

   If the pipe diameter or flow rate are the unknowns to be solved, a somewhat different procedure
must be used because the Reynolds number cannot be calculated first to determine the type of
flow behavior to be expected. In distribution systems, partially turbulent flow is most prevalent.
Therefore, if the flow rate or diameter required is to be determined, partially turbulent flow should
be assumed initially. The guidelines in Table 4-15 can be used to select an appropriate flow
equation for the flow conditions specified. Usually, it will be the lGT distribution
equation. Although for small diameter pipe ≤2 inch (51-mm) for pounds pressure systems, it might
be the Mueller equation; for large diameter pipe, ≥16 inch (404-mm) , the Panhandle A equation
may be chosen.
The following procedure is used:
 The flow equation selected is used to solve for the flow rate or diameter.
 The resulting value computes the Reynolds number for comparison against the critical
Reynolds number. If the calculated value is smaller than the critical Reynolds number, the
solution is valid.
 If the calculated value is greater than the critical Reynolds number, the calculation is
repeated using the fully turbulent flow equation. This obtains a better estimate of the unknown
flow rate or diameter. For calculating diameter, the next largest actual diameter for the type of
pipe involved is used.

Example 2:
Statement of Problem: For the system described in the previous example, compute the capacity
for a one-mile (1.61-km) section of 4-in (102-mm) thin wall [0.188-in. (4.77-mm) wall thickness]
feeder main with an inlet pressure of 40 psig and a minimum allowable downstream pressure of
20 psig.
Solution:
The inside diameter of the pipe can be obtained from Table 4-7. 4.124 inches (104.7 mm).
Assuming partially turbulent flow, the appropriate flow equation to use is again the lGT distribution
equation.
Because the unit length resistance factor for 4-inch, schedule 40 pipe has already been
calculated for the conditions specified, it can be used to calculate the unit length resistance factor
for 4-inch thin-wall pipe by the ratio process outlined below:
RSch.40  = 3.418 X 10-3 TfG0.80 m0.20/(4.026)4.80
Rt  = 3.418 X 10-3 TfG0.80 m0.20/(4.124)4.80
Rt/RSch.40  =  (4.026/4.124)4.80
Where 
Rt  = resistance factor for 4-inch (102-mm), thin-wall pipe
= 0.13025 X 10-3 (0.9762)4.80
= 0.13025 X 10-3 (0.8908)
= 0.11603 X 10-3
   The known values now can be substituted into the simplified form of the lGT equation:
rP2  =  RL Qb1.80
(54.7)2-(34.7)2 =  0.11603 X 10-3 (5280)(Qb)1.80 (2992-1204)/0.61264 = Qb1.80
   This equation now can be solved for Qb:
2918.5 = Qb1.80
Qb  = (2918.5)1/1.80
= (2918.5)0.555
= 84.2 MCF/h
The Reynolds number can now be calculated from the Reynolds number _expression from the
first exercise and checked against the critical Reynolds number:
Re  =  26.934 Qb/D
= 26.934 (84,200)/4.124
= 549,913
   From Table 4-15, the critical Reynolds number for 4-inch (102-mm) pipe of average wall
roughness, 0.7 mil, is 520,000. Because the calculated Reynolds number exceeds the critical
Reynolds number, the calculation will be repeated using the fully turbulent flow equation. The
_expression for the unit length resistance factor can be obtained from Table 4-2.
Rt  =  3.645 X 10-3 TfG Zavg./(log 3.7 D/k)2D5
= 3.645 X 10-3 (490)(0.60)(1.0)/(log 3.7 x 4.124/0.7 X 10-3)(4.124)5
= 1.07163/(log 21.798 x 103)2 (1.193) X 103
Rt =  1.07163 X 10-3/(4.33842)2 (1.193)
= 1.07163 X 10-3/(18.822)(1.193)
= 4.772 X 10-5
   The known values can be substituted into the simplified form of the fully turbulent flow equation:
P2  = RL Qb2
1788  =  4.772 X 10-5 (5280)(Qb)2
When this equation is solved for Qb2, it becomes:
Qb2  = (1788/0.25196)
= 7096
Qb  =  84.2 MCF/h
   This result is the same as that for the lGT distribution equation because the Reynolds number is
close to the critical-Reynolds-number and the transmission factor for the lGT distribution equation
is about 2% below the Smooth Pipe Law at this Reynold. s number.

Elevation Corrections

   All of the flow equations that have been discussed assume a horizontal pipeline. Significant
elevation changes that occur along the route of a pipeline must be taken into account by using a
different flow equation or by adding a correction term to the results obtained from a flow equation
designed for horizontal piping.

   High-Pressure Systems

   In medium- and high-pressure distribution systems, the effect of an elevation change on a
pressure drop is less than five percent of the frictional loss until the elevation change per mile
exceeds several hundred feet. Corrections for an elevation change are seldom necessary in
these systems; since elevation changes of several hundred feet per mile are seldom experienced
in distribution systems.
   At the higher pressures sometimes carried in utility supply mains, the effect of elevation change
becomes more significant and must be included in flow calculations. Because flow problems of
this type are infrequent and exceptions in distribution design, the equations and calculation
procedures are not addressed here.

   Low-Pressure Systems

   In low-pressure gas distribution systems, the gage pressure is the pressure that causes gas to
flow out of the mains and through gas appliances. Because gage pressure is the difference in
pressure between the gas in the pipe and the air surrounding the pipe, it is affected by the
changes in the atmospheric pressure with elevation as well as by changes in the gas pressure
with elevation. Because natural gas is lighter than air, the change in atmospheric pressure
outside a gas main is greater for a given change in elevation than the change in the total gas
pressure inside the main. As a result, the gage pressure at the top of an inclined section of pipe
may be greater than the gage pressure at the bottom if frictional effects are not too great. This is
illustrated in Figure 4-16, in which the diagram at the top shows an elevation view of a gas main
that is laid uphill; downstream elevation, upstream elevation, and the difference in elevation are
noted. The lower diagram is a plot of absolute pressure with distance along the main. The terms
in this diagram are defined as follows:
Pa1  = absolute atmospheric pressure at lower upstream location, assuming the direction of flow is
uphill (in. wc)
Pa2  = absolute atmospheric pressure at higher downstream location (in. wc)
Pg1  = absolute upstream pressure of the gas (in. wc)
Pg2  = absolute downstream pressure of the gas (in. wc)
hw1  = Pg1 . Pa1 = gage pressure of the gas at the upstream location (in. wc)
Fig. 4-16. Diagram of Pressure vs. Distance Along Low-Pressure a Gas Main Laid Up-Hill
 
 hw2  = Pg2 - Pa2 = gage pressure of the gas at higher downstream location
hw2 > hw1
   As the diagram shows, the gage pressure of the gas at the higher downstream location (2) is
greater than the gage pressure of the gas at the lower upstream location (1) ,even though the
absolute pressure of the gas at the higher location is lower than that at the upstream location
because of the frictional energy loss resulting from the uphill flow.
   The magnitude of the pressure change associated with elevation change can be estimated from
the following equation:
   P2 = P1 + 0.015 (1-G) rH (Eq. 4-9)
   where:
P2  = downstream gas pressure (in. wc gage)
P1  = upstream gas pressure (in. wc gage) G = specific gravity of gas
rH  = difference in elevation between downstream and upstream locations at which P 2 and P1 are
measured (ft)
   To obtain the difference in the gage pressure between the downstream and upstream locations
along the main, one subtracts the pressure drop caused by friction, as calculated from any
suitable flow equation, from the elevation correction given by Equation 4-9.
Example:
A 4-inch (100-mm) low-pressure gas main 1000 ft (305 m) long runs up a hill with an elevation
change of 300 ft (91 m). This main delivers 0.65 gravity gas to customers whose maximum rate of
gas use is 1 MCF/h. Calculate the difference in the gage pressure of the gas in the main between
the bottom and top of the hill during a period of maximum gas use. Use the Spitzglass low-
pressure equation.
Calculation of the pressure drop caused by friction:
hw  = KQb2
From Table 4-4
R4  =  1.017 X 10-4
\K  = RL = 1.017 X 10-4 X 1000 = 0.1017
hw  = 0.102 (1)2 = 0.10 in. wc
The calculation of the elevation correction is as follows:
P2 = P1+0.015(1 - 0.65) 300 = P1 + 1.58 in. wc
   The net difference in gage pressure between the top and the bottom of the hill is 1.58 - 0.10 =
1.48 in. wc. That is, the gage pressure at the top of the hill is 1.48 in. wc greater than the gage
pressure at the bottom. If the direction of gas flow was downhill, the elevation correction would
have a negative sign because rH = (H2 - H1) would be negative.

Pressure Setting Caution

   Because of the elevation effect, if a regulator that is located at the bottom of a hill is set to the
maximum allowable operating pressure to meet peak demand, unacceptably high pressures will
result at the top of the hill during periods of low demand.
   Similarly, significant elevation effects are present in the fuel-line risers for tall buildings. For
each 20 feet (6.1 m) rise, there is a pressure gain of about 0.1 inch (2.5 mm) wc to offset the
friction loss in the riser. Again, the designer must take precautions: the outlet pressure setting of
the building regulator must be set low enough so unacceptably high pressures will not be
experienced at the top of the building during low demand.
 
 

APPENDIX
_Expression of Gas Flow Rates at Standard Conditions
   In the General Flow Equation, the gas flow rate is expressed in terms of SCF/unit of time, the
flow rate unit usually used in the gas industry. The actual volumetric flow rate at any specified
location along a section of pipe differs numerically from the flow rate expressed in SCF because
the temperature and pressure of the flowing gas are not at standard conditions. Both ways of
expressing flow rate, however, refer to the same mass rate of flow. If an ideal gas behavior is
assumed, a relationship between these two methods of expressing volumetric flow rates can be
established by using the Ideal Gas Law to relate the volumetric flow rates to the mass rate of
flow:
   PbQb = NRTb or (PbQb)/Tb = NR (Eq. 4-10)
   and
   PfQf = NRTf or (PfQf)/Tf = NR (Eq. 4-11)
Where
Q  =  gas flow rate, SCF/h
N  =  gas flow rate, moles/h
subscript b = base conditions
subscript f = flowing gas conditions
   Because the mass flow rate at a given point is the same whether expressed by Equation 4-10
or 4-11:
   (PbQb)/Tb = (PfQf)/Tf (Eq. 4-12)
   If both sides of this equation are divided by Pf and multiplied by Tf, it becomes:
   Qf = (Pb/Pf) (Tf/Tb)Qb (Eq. 4-13)
   This equation relates the actual volumetric flow rate in the pipe to the volumetric flow rate at
base conditions.
   Because the pressure of the gas flowing through a pipe gradually decreases in the direction of
the flow, the actual volumetric flow rate and gas velocity increase with distance down the line.
The gas velocities are sometimes of interest in the design and operation of gas distribution
systems discussed in the next chapter.
Example:
Gas is flowing through a 4-mile section of 12-in. pipe (inside diameter = 12.25 in.) at a steady rate
of 850,000 SCF/h (Pb = 15.0 psia; Tb = 520ºR), and at a constant temperature of 60ºF. At this flow
rate, the gas pressure at the outlet is 26.6 psig (atmospheric pressure = 14.7 psi) when the
pressure at the pipe inlet is 50 psig. Determine the volumetric flow rate at flow conditions, and the
velocity of the gas, at the inlet and outlet of the pipe section.
Solution:
At the pipe inlet:
Gage pressure  = 50.0 psi
Atmospheric pressure = 14.7 psi
Absolute pressure  = 64.7 psi
From Eq. 4-13:
Qf = (15.0/64.7) (520/520) 850,000 = 197,000 CF/h
If A is the cross-sectional area of the pipe,
A = (PD2)/4 =P(12.25)2/4 = 117.9 in.2
- 117.9/144 = 0.819 ft.2
If V is the velocity of the gas at the inlet,
V = Qf/A = CF/h/ft2 = 197,000/0.819
= 241,000 fl/h or 45.6 miles/h
At the pipe outlet:
Gage pressure  = 26.6 psi
Atmospheric pressure  = 14.7 psi
Absolute pressure  = 41.3 psi
From Eq. 4.13:
Qf  =  (15.0/41.3) (520/520) 850,000 = 390,000 CF/h
V  =  309,000/0.819 = 377,000 ft/h or 71.4 miles/h

*
 The conversion of flow rates in cubic feet at flowing conditions to standard cubic feet at base
conditions is discussed in the Appendix of this chapter.

*
The conversion of flow rates in standard cubic feet at base conditions to cubic feet at flowing
conditions is discussed in the appendix to this chapter.