Fluid Flow Design Calculation

Equation 29

Metric :
 
 d 5 P12 - P22 
Qg  4.367  10 -3  
 Z TfL S 

Customary :

 d 5 P12 - P22 
Qg  0.199 


 Z TfL S 

9. As was the case for liquid flow, in order to determine a pipe diameter for a given flow
rate and pressure drop, it is first necessary to estimate the diameter and then to
compute a Reynolds number to determine the friction factor. Once the friction factor is
known, a pipe diameter is calculated and compared against the assumed number. If
the two are not sufficiently close, the process is iterated until they converge.

3.2.2. Small Pressure Drops

For small pressure drops, an approximation can be calculated. The following formula can
be derived from Equation 28 if P1 - P2 < 10 percent of P1 and it is assumed that

P12 - P22  2P1 P 

Equation 30

Metric :
S Qg  Z T f L
2

P  26,215
P1 d 5

Customary :
S Qg  Z T f L
2

P  12.6
P1 d 5

40 of 148
 Fluid Flow Design Calculation

3.2.3. Weymouth Equation

1. This equation is used for short lengths of pipe where high-pressure drops are likely
(turbulent flow). It is based on measurements of compressed air flowing in pipes with
inner diameters ranging from 20 to 200 mm (0.8 to 11.8 in), in the range of the Moody
diagram where the /d curves are horizontal (i.e., high Reynolds number). In this range
the Moody friction factor is independent of the Reynolds number and dependent upon
the relative roughness.

For a given absolute roughness, , the friction factor is a function of diameter only. For
steel pipe the Weymouth data indicate:

Equation 31

Metric :
0.0941
f  1
d3

Customary :
0.032
f  1
d3

2. Substituting this into Equation 29, the Weymouth equation is:

41 of 148
 Fluid Flow Design Calculation

Equation 32

Metric :
1
 P 2 - P22  2
Qg  1.42  10 - 2 d 2.67  1 
 LSZT 

Customary :
1

2.67  P12 - P22  2

Qg  1.11 d  
 LSZT 

3. Assuming a temperature of 290°K (520°R), a compressibility of 1.0, and a specific

gravity of 0.6, the Weymouth equation can also be written (This is the form of the
equation that is given in the GPSA Engineering Data Book):

Equation 33

Metric :
0.5
 T   P12 - P22 
-5
Q  3.415  10  b  E   d 2.667
 Pb   S Lm Tavg Z avg 

Customary :
0.5
 Tb   P12 - P22 
Q  433.5   E   d 2.667
 Pb   S Lm Tavg Z avg 

where :
Tb  base absolute temperature, K  R
o

Pb  base absolute pressure, kPa  psia 

E  pipeline efficiency factor
Tavg  average absolute temperature, K  R
o

Z avg  average compressibility

S  specific gravity
d  pipeline ID, mm in 

42 of 148
 Fluid Flow Design Calculation

4. It is important to know what the equation is based on and when it is appropriate to use
it. To reiterate, short lengths of pipe with high-pressure drops are likely to be in
turbulent flow, and thus the assumptions made by Weymouth are appropriate. Industry
experience indicates that Weymouth's equation is suitable for most gas piping within
the production facility. However, the friction factor used by Weymouth is generally too
low for large diameter or low velocity lines, where the flow regime is more properly
characterized by the sloped portion of the Moody diagram.

3.2.4. Panhandle Equation

1. This equation is often used for long, larger diameter pipelines. It assumes that
the friction factor can be represented by a straight line of constant negative slope
in the moderate Reynolds number region of the Moody diagram.

2. A straight line on the Moody diagram would be expressed:

Equation 34

log f  - n log Re  log N

or

Equation 35

N
f 
 
Re n

3. Using this assumption and assuming a constant viscosity for the gas, Equation
29 can be rewritten as the Panhandle equation:

43 of 148
 Fluid Flow Design Calculation

Equation 36

Metric :
0.51
-3  P12 - P22 
Qg  1.229  10 E f  0.961  d 2.53
 S ZTLm 

Customary :
0.51
 P12 - P22 
Qg  0.028E f  0.961  d 2.53
 S ZTLm 

where :
E f  efficiency factor, dimensionless
 1.0 for brand new pipe
 0.95 for good operating conditions
 0.92 for average operating conditions
 0.85 for unfavorable operating conditions

4. In practice, the Panhandle equation is commonly used for large diameter long
pipelines where the Reynolds number is on the straight-line portion of the Moody
diagram.

5. Neither the Weymouth nor the Panhandle equation represents a "conservative"

assumption that can always be used to overstate pressure drop. If the
Weymouth formula is used and the flow is in a moderate Reynolds number
regime, the friction factor will in reality be higher than assumed (because the
sloped line portion is higher than the horizontal portion of the Moody curve), and
the actual pressure drop will be higher than calculated. If the Panhandle formula
is used and the flow is actually in a high Reynolds number regime, the friction
factor also will be higher than assumed (because the equation assumes the
friction factor continues to decline with increased Reynolds number beyond the
horizontal portion of the curve), and the actual pressure drop will again be higher
than calculated.

44 of 148
 Fluid Flow Design Calculation

3.2.5. Spitzglass Equation

1. The Spitzglass equation is used for near-atmospheric pressure lines. It is

derived directly from Equation 29 by making the following assumptions:

 3.6  1 
a) f  1    0.03d   
 d   100 

b)   520 o R

c) P1  15 psi

d) Z  1.0

e) P  10 percent of P1

2. With these assumptions, and expressing pressure drop in terms of inches of

water, the Spitzglass Equation can be written:

45 of 148
 Fluid Flow Design Calculation

Equation 37

Metric :
 
 hw d 5 
Qg  3.655  10 - 2  
 S L  1  9.144  1.18  10 -3 d  
  
 d 

Customary :
 
 hw d 5 
Qg  0.09  
 S L  1  3.6  0.03d  
  
 d 

where :
Qg  gas flow rate, std m 3 / hr MMSCFD 
hw  pressure loss, mm of water in of water 
d  pipe ID, mm in 
L  Length of pipe, m  ft 
S  specific gravity of gas relative to air

3.2.6. Comparison and Recommended Use of Gas Flow Equations

1. The Weymouth and Spitzglass equations both assume that the friction factor is a
function of pipe diameter only. Figure 7 compares the friction factors calculated
from these equations with the factors indicated by the horizontal line portion of
the Moody diagram for two different absolute roughnesses.

46 of 148
 Fluid Flow Design Calculation

Figure 7: Friction Factor vs. Pipe Diameter for Three Correlations

2. In the small pipe diameter range [75 to 150 mm (3 to 6 in)], all curves tend to yield
identical results. For large diameter pipe [250 mm (10 in), and above], the Spitzglass
equation becomes overly conservative. The curve is going in the wrong direction and
thus the form of the equation must be wrong. If used, its predictive results are higher
pressure drops than actually observed. The Weymouth equation tends to become
optimistic with pipe diameters greater than 500 mm (20 in). If used, its predictive
results are lower pressure drops than actually observed. Its slope is greater than the
general flow equation with  = 0.002 in. These results occur because of the ways the
Spitzglass and Weymouth equations approximate the Moody diagram.

3. The empirical gas flow equations use various coefficients and exponents to account for
efficiency and friction factor. These equations represent the flow condition at which
they were derived but may not be accurate under different conditions. Unfortunately,
these equations are often used as if they were universally applicable.

4. The following guidelines are recommended in the use of the gas flow equations:

47 of 148
 Fluid Flow Design Calculation

a) Use the general gas flow equation for most general usage. If it is
inconvenient to use the iterative procedure of the general equation but
high accuracy is required, compute the results using both the Weymouth
and Panhandle Equations and use the higher calculated pressure drop.

b) Use the Weymouth Equation only for small diameter, short run pipe within
the production facility where the Reynolds number is expected to be high.
The use of the Weymouth equation for pipe greater than 500 mm (20 in)
in diameter or in excess of 4,600 meters (15,000 ft) long is not
recommended.

c) Use the Panhandle Equation only for large diameter, long run pipelines
where the Reynolds number is expected to be moderate.

d) Use the Spitzglass equation for low pressure vent lines less than 300 mm
(12 in) in diameter.

e) When using gas flow equations for old pipe, attempt to derive the proper
efficiency factor through field tests. Buildup of scale, corrosion, liquids,
paraffin, etc., can have a large effect on gas flow efficiency.

3.3. Two-Phase Flow

3.3.1. General

1. In some single-phase flow conditions, a small volume of gas may be entrained in

liquid flow (such as a liquid dump line from a separator), or a small amount of
liquid may be carried in the pipe in gas flow (such as gas off a separator). These
small amounts usually have a negligible effect on pressure loss and are not
considered in single phase flow calculations. However, there are certain flow
conditions where sufficient volumes of a second gas or liquid phase exist to
produce an appreciable effect on pressure loss. The pressure drop in such lines
shall be considered using techniques for two-phase flow.

48 of 148
 Fluid Flow Design Calculation

2. Examples of two-phase flow situations include:

a) Fluid coming out of the well bore prior to liquid separation

b) Gas and oil that have been metered and then recombined for flow in a common
line to a central facility

3. Using the best correlations available for pressure drop and liquid hold up,
predictions may be in error by  20 percent for horizontal flow and  50 percent
for flow which is slightly inclined.

3.3.2. Flow Regimes

1. When a gas-liquid mixture enters a horizontal pipeline, the two phases tend to
separate with the heavier liquid settling to the bottom. The type of flow pattern
depends primarily on the gas and liquid flow rates. Figure 8 shows typical flow
patterns in horizontal two-phase pipe flow.

49 of 148
 Fluid Flow Design Calculation

Figure 8: Two-phase Flow Patterns in Horizontal Flow (Source: P. Griffith,

"Multiphase Flow in Pipes," JPT, March 1984, pp. 363-367)

2. Horizontal flow regimes can be described as follows:

a) Bubble Very low gas-liquid ratios. Gas bubbles rise to the top.

b) Elongated Bubble With increasing gas-liquid ratios, bubbles become larger and
form gas plugs.

c) Stratified Further increases in gas-liquid ratios make the plugs become longer
until the gas and liquid are in separate layers.

50 of 148
 Fluid Flow Design Calculation

d) Wavy As the gas rate increases, the flowing gas causes waves in the flowing
liquid.

e) Slug At even higher gas rates, the waves touch the top of the pipe, trapping gas
slugs between wave crests. The length of these slugs can be several hundred
feet long in some cases.

f) Annular Mist At extremely high gas-liquid ratios, the liquid is dispersed into the
flowing gas stream.

3. Figure 9 can be used to approximate the type of flow regime expected for any flow
condition. In most two-phase lines in the field, slug flow is predominant in level and
uphill lines. In downhill lines, stratified flow is predominant.

51 of 148
 Fluid Flow Design Calculation

Figure 9: Horizontal Multi-phase Flow Map (Source: P. Griffith, "Multiphase Flow in

Pipes," JPT, March 1984, pp. 363-367)

4. Two-phase flow patterns in vertical flow are somewhat different than those occurring in
horizontal flow. Different flow regimes may occur at different segments of pipe, such as
flow in a well tubing where pressure loss causes gas to come out of solution as the fluid
moves up the well. Figure 10 shows typical flow regimes in vertical two-phase flow.

52 of 148
 Fluid Flow Design Calculation

Figure 10: Two-phase Flow Patterns in Vertical Flow (Source: J.P. Brill, "Multiphase
Flow in Wells," JPT, January 1987, pp. 15-21)

5. Vertical flow regimes can be described as follows:

a) Bubble

Small gas-liquid ratio with gas present in small, randomly distributed bubbles. The liquid
moves up at a uniform velocity. Gas phase has little effect on pressure gradient.

b) Slug Flow

The gas phase is more pronounced. Although the liquid phase is still continuous, the gas
bubbles coalesce into stable bubbles of the same size and shape, which are nearly the
diameter of the pipe. These bubbles are separated by slugs of liquid. Both phases have a
significant effect on the pressure gradient.

c) Transition Flow or Churn Flow

53 of 148
 Fluid Flow Design Calculation

The change from a continuous liquid phase to a continuous gas phase occurs in this
region. The gas phase is predominant and the liquid becomes entrained in the gas. The
effects of the liquid are still significant.

d) Annular-Mist Flow

The gas phase is continuous and the bulk of the liquid is entrained in and carried by the
gas. A film of liquid wets the pipe wall and its effects are secondary. The gas phase is the
controlling factor.

6. Normally, flow in oil wells is in the slug or transition flow regime. Flow in gas wells can
be in mist flow. Figure 11 can be used to determine the type of regime to be expected.

54 of 148
 Fluid Flow Design Calculation

Figure 11: Vertical Multiphase Flow Map (Source: Yaitel, Y., Barhea, D., and
Duckler, A.E., "Modeling Flow Pattern Transitions for Steady Upward Gas-Liquid
Flow in Vertical Tubes," AIChE J., May 1980, pp. 345-354.)

7. In two-phase piping, pressure drop is caused by the friction developed due to the
energy transfer between the two phases as well as that between each phase and the
pipe wall. Pressure drop calculations shall take into account the additional friction loss
due to the energy transfer between phases.

8. The detailed calculation of pressure drops in two-phase pipelines requires an evaluation

of phase changes due to pressure and temperature changes, evaluation of liquid
holdup using empirical formulas, and evaluation of energy transfer between the phases.
These are addressed in the many proprietary computer programs available. It is
beyond the scope of this manual to evaluate all the equations which have been
proposed in the literature or to develop a new computer algorithm.

55 of 148
 Fluid Flow Design Calculation

9. It shall be kept in mind that even under the best conditions small changes from
horizontal in piping systems can lead to large errors in calculating pressure drops.
Table 2 shows that, although the different correlations analyzed against field data on
the average give reasonable results, the standard deviation is large; any one calculation
could be as much as 20 to 50 percent in error.

Table 2: Two Phase Flow Correlations, AGA Multiphase Pipeline Data Bank for Gas-
Condensate Lines (From: Battarra, Mariana, Gentilini and Giaccheta, Oil and Gas
Journal, Dec. 30, 1985)

Flow Holdup Down-Hill Friction Mean Error, Standard

Pattern Recovery Percent Deviation,
Percent

BEG2 BEG2 NOCO BEGO -10.8 40.5

BEG2 BEG2 NOCO BEGC 28.1 62.7

MAN1 BEG2 GARE BEGC 29.5 63.3

MAN1 EATO FLAN DUKO 33.7 47.9

MAN1 EATO FLAN DUKC 92.5 89.8

MAN1 EATO FLAN OLIE 0.5 30.9

BEG2 = Revised Beggs and Brill

MAN1 = Mandane

EATO = Eaton

NOCO = No correction to institute density

56 of 148
 Fluid Flow Design Calculation

Flow Holdup Down-Hill Friction Mean Error, Standard

Pattern Recovery Percent Deviation,
Percent

GARE = Gas recovery only

FLAN = Flanigan

BEGO = Original Beggs and Brill

BEGC = Beggs and Brill with Colbrook

DUKO = Original Dukler

DUKC = Dukler with Colbrook

OLIE = Oliemans

10. The following four correlations have been found to give reasonable results when used
within the limitations inherent in their derivation.

3.3.3. API RP 14E

1. The following formula, presented in the American Petroleum Institute's Recommended

Practice API RP 14E is derived from the general equation for isothermal flow assuming
that the pressure drop is less than 10 percent of the inlet pressure:

57 of 148
 Fluid Flow Design Calculation

Equation 38

Metric :
2
fLWh 
P  62,561
md 5

Customary :
2
3.36  10 -6 fLWh 
P 
md 5

where :
P  pressure drop, kPa  psi 
L  length of pipe, m  ft 
Wh  flow rate of liquid and vapor, kg / hr lb / hr 
 m  mixture density, kg / m 3 lb / ft 3 
d  pipe ID, mm in 
f  Moody friction factor

2. This equation assumes that there is no energy interchange between the phases,
that bubble or mist flow exists so that the fluid can be described by an average
mixture density, and that there are no elevation changes.

3. The flow rate of the mixture to use in this equation can be calculated as follows:

58 of 148
 Fluid Flow Design Calculation

Equation 39

Metric :
Wh  1.21 Qg S  999.7 Ql SG 

Customary :
Wh  3,180 Qg S  14.6 Ql SG 

where :
Wh  flow rate of liquid and vapor, kg / hr lb / hr 
Qg  gas flow rate, std m 3 / hr MMSCFD 
Ql  liquid flow rate, m 3 / hr BPD 
S  specific gravity of gas relative to air
SG   specic gravity of liquid relative to water

4. The density of the mixture to use in Equation 38 is given by:

59 of 148
 Fluid Flow Design Calculation

Equation 40

Metric :
28,814SG  P  34.81R S P
m 
28.82P  10.0 R T Z

Customary :
12,409 SG  P  2.7R S P
m 
198.7 P  R T Z

where :
 m  mixture density, kg / m 3 lb / ft 3 
P  pressure, kPa  psia 
SG   specific gravity of the liquid relative to water
use the average gravity for the hydrocarbon and water mixture 
S  specific gravity of gas relative to air

R  gas / liquid ratio, std m 3 / m 3 std ft 3 / bbl 
T  temperature, K  R
o

Z  gas compressibility factor

3.3.4. AGA Equation

The American Gas Association method uses a frictional pressure drop calculation originally
developed by Dukler and an elevation pressure drop calculation originally developed by
Flanigan. This seventeen-step method is an iterative procedure described in the Fluid
Flow and Piping Sections of the GPSA Engineering Data Book.

3.3.5. Beggs and Brill Equation

1. This correlation was developed by two University of Tulsa students, Dale Beggs
and James Brill. Their original procedure first appeared in the May 1973 issue of
the Journal of Petroleum Technology. Almost all correlations prior to the Beggs
and Brill method could predict pressure drop in two-phase flow for vertical or
horizontal flow only. Prior to 1973, however, no correlation existed for predicting

60 of 148
 Fluid Flow Design Calculation

the pressure drop in two-phase flow at any angle of inclination. Beggs and Brill
therefore set out to develop such a correlation.

2. The original article presents a specific description of the experimental procedure

used to develop this method. Such a description is beyond the scope of this
Tutorial. The Beggs and Brill method was correlated using small diameter pipe
and generally shall be applied to pipelines 200 mm (8 in) or less in diameter.

3.4. Head Loss in Valves and Pipe Fittings

In many piping situations, including those in most production facilities where space is
limited, the pressure drop through valves, pipe fittings, and enlargements and contractions
is a significant portion of the overall pressure drop in the pipe segment. A pipe flow
restriction that changes velocity or direction of the flow stream causes pressure drops
greater than that which would normally occur in a straight piece of pipe of the same length.
The three most common ways of calculating these pressure drops are by using resistance
coefficients for fittings, flow coefficients for valves and equivalent lengths for both valves
and fittings.

3.4.1.

Resistance Coefficients for Fittings

1. The Darcy-Weisbach equation, Equation 17, can be

rewritten as:

61 of 148