T.P.
2777
DETERMINING FRICTION FACTORS FOR MEASURING
PRODUCTIVITY OF GAS WELLS
R. V. SMITH, U. S. BUREAU OF MINES, BARTlESYlllE, OKLAHOMA
ABSTRACT
The theoretical background for calculating friction factors
for flow in gas wells by ~wo methods is presented. The first
method, requiring pressures, temperatures and specific vol-
umes. of the flowing fluids at various depths in the well bore,
shows how the mechanical-energy-balance equation for ver-
tical flow may be graphically integrated over the actual path
of the ,expansion of the fluid "in the well- Thus, assumptions
regarding the effective temperature and effe~tive compre3sib:l-
ity of the fluid in the well are avoided. The second method
pre~ents an equation derived on a basis of the assumptions that
both the temperature and .the compre3sib~l:ty are fi.;;:ed at
constant effective values throughout the flowing column of
gas. The ~econd method provides a conven:ent and practical
means of calculating friction factors for gas wells and lends
itEelf readily to the problem of calculating sub:mrface pres-
sures in a flowing gas well. The application of both methods
to actual test data taken 011 a flowing gas well is illustrated in
the paper.
INTRODUCTION
As friction factors for the producing strmg3 of flowing gas
wel13 cannot be measured directly and must be cabulated from
flow-test data, study of tlle methods of arriving at .friction
factors is a ne:::essary adjunct to understanding the charac-
teristics of flow in gas wells. There are two methods of cal-
culating friction factors for gas welh; they differ from one
anotl:er mainly in the treatment of the path of expansion of
the fluid in the weI], In the flowing well, the energy consumed
in Efting the fluid from the bottom to the top of the well,
overcoming the friction between the movIng fluid and the pipe
walls, and increasing the 'velocity of the fluid as it, flows np
the producing string is s1,lpplied by expansion of the flowing
fluid. The available energy is determined, 'by the expansion of
the fluid that follows a path determined by conditions of
temperature, compressibility and phase changes of the fluid
during the expansion; A means of evaluating the available
energy in a flowing gas well and deterniining the proportion
of the available energy used in lifting the gas, overcoming
friction, and increasing the velocity of flow is developed in
this report. Knowing how much energy is consumed in over-
c~ming friction' makes it possible to calculate friction factors
for given flow rates in given sizes of pipes in wells.
Friction coefficients, as used in tlrls report, are dimension-
Vol. 189, 1950 PETROLEUM TRANSACTIONS, AIME
less proportionality multipEers med in the flow equations to
satisfy, the equality between the terms of the equation. The
square root 'of the reciprocal of the friction coefficient is
termed the friction factor.
It has been known for many years that friction factors may
be computed directly from mathematical formulas based on
certain assumptions regarding the temperature and compres-
sibility of the moving fluid in a well. A general equation is
pre:ented in this report without assumptions for vertical flow
of fluids and a conventional-type equation is derived on the
basis of assumptions that fix the'temperature and compressi-
b:lity at constant values. Accurate pressures at the sand faces
in welh are requ:red in the method of determining· the pro-
ductivity of gas wells, as outlined by Rawlins and Schellhardt'
in Bureau of Mines Monograph 7. Where measurements are
not made w:th subsurface-pressure gauges or' static gas col-
umns are unavailable, flowing pressures customarily are cal-
culated at the sand face in the well by the use of the well-
known Weymouth formula' . Natural gas engineers have
real:Zed that errors introduced by the use of friction factors
as given by the Weymouth formula are relatively unimportant
in te3ting low-capacity gas wells; .they also know that such
factors are important considerations in testing large-capacity
gas wells. Accordingly, .pre2ent re::earch on the productivity
of gas welh at the Petroleum Experiment Station of the Bu-
reau of Mines, Bartlesv:lle, Okla., is being directed toward
measurement of the pressure loss due to friction in flowing
gas welh.
It is beyond the scope of this report to show how friction
factors vary with rate of flow and in pipes of different diam-
eters, as it is intended only to develop and illustrate the use
of mathematical, expressions for calculating friction factors
from flow-test data. The equations pre3ented apply only to
. turbulent flow in circular pipes.
ENERGY RELATIONS FOR FLOW OF FLUIDS2
The concept of conservation of energy is usually the basis
of any study of fluid flow through vertical pipes as in gas
well;, horizontal pipe lines, or orifices. In deriving equations,
the following symbols are used:
1 References given at end of the paper.
ManUr,cript received at the office of the Petroleum Branch September
20. 1949. Paper present£d "t'the Petroleum Branch meetings in Co:umbus.
Ohio, SeptEmber 26-27, and in San Antonio, Texas, October 5-7, 1!>49.
2 The explanation, in this report, of energy relations for fluid flow was
taken from a discussion of flow of fluidG by Cattell."
73
 l'~P. 2777 DETERMINING FRICTION FACTORS FOR MEASURING PRODUCTIVITY OF GAS WELLS
A = cross-sectional area of the flow channel, sq ft
U = mean linear velocity of flow parallel to the axis of
channel at a given cross section, ft per sec
g = acceleration of gravity, ft per sec per sec
E = total internal energy or intrinsic energy associated with
one pound of fluid, ft-Th
p = pressure, lb per sq ft abs
P = pressure, lb per sq in. abs (psia)
v ;= specific volum~ of fluid, cu ft per Ib
L = distance of a cross section above an assumed datum
plane, ft
q = heat, in equivalent it-Ib, imparted to I Ib of fluid from
outside the fluid body
F= mechanical energy, ft-Ib per lb of fluid, transformed
into heat by frictional resistances.
Subscripts 1 and 2 applied to any of the symbols denote
conditions at an upstream cross section Al and a downstream
cross section A,; respectively. In the example considered in
this report, the upstream cross section Al is lower in .elevation
, than the downstream cross section A,.
Consider a fluid flowing upward from section Al to section
A. through a circular vertical pipe in which Al equals A,
in area. As the fluid under consideration is elastic, it fills the
pipe completely, and no external forces except gravity act
upon it between the two sections.
In considering a hypothetical well the following assumptions
are made:
1. The fluid is homogeneous.
2. The flow is steady with respect to time; that is, in a unit
of time the same weight of fluid flows through every cross
section.
3. The fluid flows with mean linear velocity U I parallel to
the axis of the pipe and at right angles to the plane of
the cross'section, AI' The mean velocitY'may be defined as
Nv
U=T
where
N =
weight rate of flow, lb per sec
v =specific volume of the fluid at the point under consid-
, eration, cu ft per Th
A =' area of cross section under consideration, sq ft
As the flow is considered steady, 1 Th of the fluid passes
through sections Al and A. in the same interval of time. When
1 lb of fluid passes the lower section Al it carries with it the
following energy: >
E" its internal energy, resulting from its temperature and
the position of its molecules,
U'
__1_, its kinetic energy, attributable to its upward velocity,
2g
P1V" which represents the mechanical work performed in
forcing the pound of fluid across section A"
L .. which is the potential energy (referred to the assumed
datum plane) of the pound of fluid resulting from the
action of gravity upon it.
2
· h y dra ulics,. Tg'
AppIying terms' used In U. p,v. and L1 may
be called, respectively, the velocity head, pressure head, and
potential head of 1 lb of fluid at section A"
74. PETROLEUM TRANSACTIONS, AIME
Under the conditions stipulated, the pound of fluid receives
no mechanical energy from external sources while passing
from At to A" but it may receive heat q from sources outside
the fluid. If the fluid loses heat, q has a negati'ye value.
At Section A:, 1 lb of fluid has the internal energy E" the
velocity head ~: ' the pressure head P.V. and the potential
head L,.
Balancing all forms of energy and expressing quantities of
heat in units of equivalent work, it is apparent that:
. The internal energy, velocity head. pressure head, and
potential head of I lb of fluid at section A" plus the heat
per Th of fluid imparted from external sources to the
fluid between sections Al and A" equals the internal
energy, velocity head, perssure head, and potential head
of 1 Ib of fluid at section A., or,
U.'
E. + - - + PiV. + L. + q
2g
=
Between sections Al and A. each pound of fluid does the
mechanical work f~dv upon the surrounding fluid; thus
'" .
each pound of fluid must receive that amount of work from the
rest of the fluid between the two sections. Such work can be
done only by converting heat energy of the fluid into mechan-
ical energy. As the work and heat mJ.lSt be equivalent,
pdv is the measure per pound of flowing fluid, not only of the
work received but also of the heat (expressed in equivalent
work units) that was transformed into that work. It may,
1;herefore, be considered that each pound of fluid does the
work f v'pdv upon itself at the expense of its own heat energy.
VI
For each pound of fluid flowing from section,At to section
A. the mechanical work F done agaInst frictional resistance
is converted into heat F (expressed in work .units) absorbed
by the fluid ..
With these conceptions, and measuring heat in equivalent
work units, a balance of heat (or nonmechanical) energy may
be expressed as follows:
The internal energy of 1 Th of the fluid at section A" minus
the heat energy per pound of fluid transformed into mechani- ,
cal work by expansion of the' fluid between the two sections,
plus the heat per pomid of fluid into which mechanical work
is transformed in overcoming frictional resistance between
the two sections, plus the heat per pound of fluid imparted
from external sources to the fluid between the two sections, is
equal to the internal energy of 1 lb of fluid at section A., or,
E, - IV.
VI
pdv + F
"'
+ q = E, (2)
Subtractipn of equation (2) from equation (1) gives equa-
tion (3)3
3 Equation (3) may be developed also by ba'ancing mechanical t'nergy.
The velocity head. pressure head. and poten'ial head of 1 pound of fluid at
section A" plus the work per pound of fluid (derived from heat energy)
done upon the fluid between the two sections. minus the mechanical ...nergy
per pound of fluid converted into heat by frictional resistanCES between
the two sections. is equal to the velocity head. pressure head. and potential
head of 1 pound of fluid at section A:!.
Vol. 189, 1950
 R. V. SMITH T.P. 2777

U
%
l'
+ P"V, + L. +
JV'v, pdv-F =
U.' .
2~ + P.v. + L., • (3)
suhstiuting the value of the term F given in equation (8)
into equation (4)
Equation (3) may be written in the form: -
PIV1+ y
' =
!V2pdv-p.v. (L.-L.)+
U; - U l ' - + _4f fL!l if dL
U' U· 2g 2gD L,.
f
V- 1
PlVl + pdv- p,v, = (L.,-Ll) + • - , + F . (4) (9)
VI 2g
Equation (4) may be considered a general one for the flow The expression at the left of the equality sign (as may
of fluirls in pipes. . be shown analytically or graphically), equals

f
P:>
- vdp
PI
EQUATIONS FOR FLOW OF FLUIDS
IN VERTICAL PIPES
Methcd I (graphic integration)
jP1 vdp.
1'2
If the pressures are expressed in psi instead of lh per sq ft,
The application of equation (4) to the evaluation of friction
the expression equals
factors for gas wells requires further treatment of term F
Pl
or the mechanical energy per pound of fluid transformed into
heat by frictional resistance. The work - done in overcoming
144 LP2 vdP.
frictional resistance through distance dL equals frictional As equation (9) contains a velocity term U which is not com-
resistance times the distance through which the resistance is monly used in the gas industry, velocity in ft per sec may be
overcome. For turbulent flow the frictional resistance is pro- converted to cu ft per 24 hours and specific volume expressed
portional to the surface in contact with the fluid, to the density in cu ft per lb by:
of the fluid, and approximately proportional to the square of Nv
U=
the velocity. If the resistance is proportional to the square A
of the velocity it is also proportional to the first power of the- where
U = velocity, ft per sec
kinetic energy U· of I lh of fluid. Expressed hy symbols the
2g • N = rate of flow of fluid, lb per sec
frictional resistance is proportional to: v = specific volume, cu ft per lh
A = cross-sectional area of pipe, sq ft.
(dL) (Per) (~) (5) When the volume rate of flow is measured in cu ft of gas per
24 hours, the term. N becomes:
where QG
dL = length of pipe, ft N =------
Per =
perimeter of pipe, ft 86,400 x 13.364
U = velocity, ft per sec where
5 = density, lh per cu ft Q = volume rate of flow, cu ft per- 24 hours
or (absolute pressure base = 14.4 psi and temperature
base = 60°F)
frictional resistance = (f) (dL) (Per) ( ; : ) (5) . (5) G = specific gravity of fluid, air = 1.0
The term f then hecomes a proportionality factor to satisfy 86,400 = numher of seconds in 24 hours
the equality; it is commonly called the coefficient of friction. 13.364 = volume of one pound of dry air at an absolute
pressure of 14.4 psi and at a temperature of
Multiplying the frictional resistance as expressed in equa-
60°F, cu ft.
tion (5) by the length of pipe dL to obtain the work in
Then the velocity term U becomes:
overcoming the resistance and dividing by the pounds of fluid
contained in the length of pipe dL, the work expended per QGv
U= or
pound of fluid in overcoming frictional resistance becomes: 1,154,600 A

(f) (dL) (Per) (~~) (5) (dL) U' =

Q'G'v'
1.3331 x 10" A'
dF = CA) (dL) (5)
(6)
By making the indicated substitutions and changing the
where subscripts I and 2 so that they indicate conditions at a depth
A = cross-sectional area, sq ft. s and at the wellhead (L = 0) respectively, taking the value
For a circular pipe the perimeter divided by the area equals of g, the acceleration of gravity, to be 32.17 ft per sec per sec,
4 - and changing elevations to depths, equation (9) becomes:
D where D is the internal diameter in ft. Equation (6)
2
becomes: 144 jP· vdP
= L. + Q·G·(v,/-v.") + fQ'G
12
Pw 85.772 X 10 A2 21.443 X 1012 DN
dF= 4fU'dL (7) rL
2gD v"dL . • . . . . • • . . . • • . (10)
o
or for a finite length of pipe
Changing A which is in units of sq ft into terms of inside
4f (L,
pipe _diameter, in in., and D, in ft, to in., equation (10)
F = Lt U' dL (8)
2gD becomes:

Vol. 189, 1950 PETROLEUM TRANSACTIONS, AIME 75

T.~. 2777· DETERMINING FRICTION FACTORS FOR MEASURING PRODUCTIVITY OF GAS WEllS

144 jP':;dP = L. + Q'G'(vw'-vs') + fQ'G' Nv Ndv

U = - - , and dU = - - .
rw 2.»51» x lu" d< 53.157 x 10" dO A ·A
~L v'dL (11) AI:o, pr.essure may be expressed in terms of specific volume
by the relationship:
where
ZTb
P = pressure, psia pv = ZTb, p -
v = specific volume, cu ft per lb v
L. = energy required to lift 1 Ib of the fluid from depth L Nv Ndv ZTb
ill the well, ft-Ib (as 1 lb of fluid is under considera- Substituting - - , --.-, and - - for U, dU, and p, re-
·A A v
tion L. is numerically equal to L)
L = depth of flowing column of fluid, ft spectively, in equation (13),
Q = volume rate of flow, cu ft per 24 hours (absolute dv N° 4fN'
pressure base = 14.4 psi and temperature base ZTb - = dL + - vdv + - - - v"dL (14)
v gA" 2gDA"
= 60 F) D

G = specific gravity of fluid, air = l.0 Rearranging and collecting terms,

d = internal pipe diameter, in. N' vdv
dL = ZTb dv
f = friction coefficient, dimensionless
.
v
(4fN' v'
+1 )
gA2 (4fN' v' + 1) (15)
The subscripts sand w refer to conditions at depth Land
2gDN 2gDN
at the wellhead respectively.
As equation (11) does not contain assumptions which de- Integrating between the limits, L = 0 and v = v, at the lower,
fine the path of the expansion of the fluid, it may be consid- or upstream, end of the length of flow string under considera-
ered general for the flow of fluids in wells. tion and L = L and v = v, at the wellhead, the result is,

Method II (mathematical integration) 2

lL 12 log --::::-::::--v"--=--] V
I
Starting with equation (9) L
J0
=ZTb
rl 4fN' v'
.---+1
+ - fLO
U?- U l '
P,Vl + J V'

Vl
pdv- P.v. = (L. - L , ) + 2
g
4f
2.151) LI
U'dL
2gDN v,

(9)
and making the following assumptions: N"-
1. The flow of fluid in a gas well can be considered to he gAO
isothermal at an '~effective''> temperature.
2. The deviation of the fluid from the ideal gas laws (16)
throughout the flow string of the weJI can be considered or
constant and can be expressed as an "effective" compres-
sibility factor_
2L
Then the following relationship exists:
ZTb
P,Y! = P,V. = ZTb
where
e = (:::)
Z = effective compressibility factor, dimensionless
(17)
T = effective absolute temperature, OF + 459.6
b = gas constant in appropriate units for 1 lb of gas. substituting pressure for specific volume according to the fol-
lowing relationships:
In accordance with the assumptions, equation (9) become3:
v.
pdv = (L.-L,) + -
u.? - U,' + -4f- L" U'dL j
(12)
PlV, = P,V. = ZbT
Z'b'T' Z<b2 T'
/
~ ~ ~D ~ v/' = -p,--.-, 2
VZ '=
P?
or in differential form' it is:
eciuation (17) becomes:
UdU 4£
pdv = dL + - - + - - U'dL (13)
g 2gD
As shown in a previous section of this report, velocity equals 2L
rate of flow (weight basis) times specific volume divided by
the cross-sectional area.
e ZTb

• Using the transformation of the terms to the left of the equality sign
in equation (9), which is explained in the previous section of the text,
(18)
an alternate differential form is obtained.
where
vdp + dL + UdU
g
+~
2gD
U"dL = 0 • _ . (13a).
e = natural logarithm base = 2.71828+
The solution of equation (13a) yiel<'s equ"tion (18). Identical. solutions L = length of pipe, ft
for equations (13) and (13a) are made possible by the given assumptions
regarding temperature and compressibility. Z - effective compressibility factor, dimensionless

76 PETROLEUM TRANSACTIONS, AIME Vol. 189, 1950

 R. V. SMITH. T.P. 2777

T = effective ahsolute temperature, of + 459.6 Equation (20) may he simplified further for use in the
. 1545.4 natural-gas industry by corrverting the variables to terms that
h = gas constant = - - - , eu ft, Ih per sq ft, per degree are easily measured. It is customary to express the quantity
M
that will flow in a well through a flow string of given dimen-
F ahs ·per Ih, M = molecular weight. sions, under va~ious pressure conditions, in terms of cu It
PI = pressure at lower, or upstream, end of se::tion und\lr per 24 hours at specific pressure and temperature hases. To
consideration, Ih per sq It ahs convert volume NVl in equation (20) from cu ft per second -at
p, = pressure at wellhead, lb per sq It abs temperature T and pressure plto cu ft per 24 hours Q at a
I = friction coefficient, dimensionless hase temperature of T • and hase pressure of Po it is necessary
N = rate of flow of gas, Ih per sec to multiply by
g = acceleration due to gravity, ft per· sec per sec
D = inside pipe diameter; ft 24 x 3600 P,T.
A = cross-sectional area of pipe, sq ft. PoZT
Further simplification of equation (18)5 may be accom~ Also, it is customary to express pressure in units of Ih per
sq in. ahs (psia), pipe diameter in in., and area in sq ft in
plished by considering the exponent ( 1 + 2f~Tb) and hy terms of diameter in in. Conversions in these units are ob-
tained by substituting in equation (20) the following:
suhstituting the values of the symhols shown later under
Method I of this report. The inside diameter of the flow string p = 144 P, where p is in Ib per sq ft and P in psia,
was 5.012 in. and the average value of the friction coefficient d
D = - , where D is in ft and d in in., and
was 0.003664 (see Tahle II) for the three depths considered. 12
The effective. ~ompressibility factor may be assumed to he ?rd'
0.964 and the effective temperature to he 540.9°F ahJ. The A=
4xl44
vaJue of b, the gas constant for 1 Ih of gas, may he computed
i'rom its specific gravity (0.708) by: 53.327
Also; gas constant b was shown to equal~. Suhstitut-
53.327
h=-- ing this quantity for b in the exponent of e, the exponent
G
GL
where 53.327 is the gas constant for 1 lb of dry air, and G is becomes 0.037504 - - -
. TZ
the specific gravity of the gas referred to air.
It is ctmvenient to use the symbol s for the' exponent of e in
The value of the exponent (1 + 2fZTb
D ) becomes equation (20).
Equation (20) now becomes:
5.012
12
P,T. .,rd' . r (32.17) (d) (P."-eSp w ' ) 1%
1 + ------------------------- = 1.00145 Q=24x3600poZTX4x144 ~. (2) (P,') (12) (f)' ( eS - 1 ) ,
2 x .003664 x .964 x 540.9 x 53.327
-_ . .' (21)
.708
The exponent 1.0014') mean<; that the quantity in hrackets,

"f
equation (18), is to he multipl:ed hy it'5 690th root. Although

~ (
P~-e'P~)
the exact implications of the fractional part of the exponent .Q = 545.58 To( _1_) .1 !._ d5 ' S w

Po ZT e" -1
are difficult to descrihefor all gas-well condit!ons, it is he- (22)
Eeved that the fractional portion of the exponent can be neg-
lected for rates of flow u:;ed in testing' gas welh. Aho, the If To = 60 .+ 459.6 = 519.6 OF ahs and Po = 14.4 psia
fractional portion of the exponent is derived from the kinetic equation (22) becomes

( ZT1 ) .(P=_e"-lsp . )
energy term of equation (14) and usually is neglected in hori-
~--r ~ d s I< w

-r
5
zontal pipe line formulas. Q = 19,686
Assuming that the exponent is unity, equation (18) hecomes: , (23)
2L 2L
where
.• ,ZTb. (2fN'Z'T'h') ( LoT;'.) Q = volume rate of flow, cu ft per 24 hours (absolute:
P( - e Po- = /!DA' e . -1 . (19)
pressure, base :::;; 14.4 psi and temperature hase
Rearranging and dividing hy P,' = 60°F)
% Z = effe::tive compre'3sihility of gas, dimensionless (see

NZTh _ _ ·r (Dg) (p,,_e::

P.·)
b text for explanation)
T = effe::tive temperature, O'F ahs (see text for explan-
-p,- - NV1 - A

l
(2f')
P1_ ( 2L

e ZTb
-
1 )1, (20) ation)· .
f = friction coefficient, dimensionless
where NV1 = volume rate of Howat a pressure of p, and a
temperature of T, cu ft per sec. ~ ~ = friction factor, dimensionless
• This "implificatIon of equation (18) was f'u«gested by M. H. Cullender ' d = inside diameter of pipe, in.
of the Phillips Petroleum Co., Bartlesville, Okla.

Vol. 189, 1950 PETROLEUM TRANSACTIONS, AIME 77

 T.P. 2777 DETERMINING FRICTION FACTORS FOR MEASURING PRODUCTIVITY OF GAS WELLS

P 8 = subsurface pressure at depth L in the well, psia For gas wells producing hydrocarbon liquid and water, the
p... - wellhead pressure, .psia hydrocarbon liquid: gas and the water: gas ratios must be
e = natu~allogarithm base = 2.71828+ measured so that a combined specific gravity may be computed
s = exponent of e in equations (22) and (23) = for the flow equation.
GL The specific volume of the fluid at various points in the
0.037504 TZ
flow string may be determined in the laboratory by pressure-
G = specific gravity of gas, air = 1.0 v\>lume-temperature studies of the fluid or it may be calcu-
L = length of pipe, ft. lated from published sOurces of information on the compres-
If the' quantity to the right of the equality sign in equation sibility of natural gases. The calculation of specific volumes
(23) is multiplied by and at the. same time divided by the from compressibility factors may introduce errors unless the
square root of s, the terms may be rearranged into the follow- factors are known to apply to the fluid.
ing equatio~: . Evaluation by mathematical means of -the terms containing
the integral signs in equation (ll) involves assumption which
Q = 101,650 ~ G~~ ~ ~ ~(P.--~"Pw') (es ~ J) define the path of the integral. It is thought that the study of
. . .. . (24) the actual process of expansion il1- a gas well is not advanced
Equations (23) and (24) are identical. However, equation to the point where the exact type of flow, with regard to heat
(24) approaches the appearance of the equations for flow in transfer between the fluid and its surroundings, can be defined
horizontal pipe lines. Equation (24) contains terms, to cor- for all conditions. In a previous section of this report an equa-
rect for the weight of the flowing column of gas, that ar.e tion was developed which is based on the assumption' that
identical with those derived by Fergpson.s flow in gas wells is isothermal.
The accuracy of equation (23) depends entirely upon the Provided pressures, temperatures and specific volumes are
validity of the three assumptions that were made in the devel- at hand for a given flow rate in a gas well, equation (ll) may
opment. Assuming that the temperature and. compressibility be solved for the friction coefficient by graphic integration of
of the gas .in the well are constant at "effective" values defines
the path of the expansion in the well. Actually the ·tempera-
j,P. vdP and tL ,v"d.L. Lin~ integral signs are
used because
P-... 0
ture and compressibility of the flowing gas vary with depth in the values of specific volume required for the integration are
the well. If the actual path of the expansion (adiabatic, poly- a function of the temperature as w~ll as of the pressure and
tropic or isothermal) were known for all conditions; the the relation between temperature and pressure is not specified
development of equation (23) could be .modified so that by the terms of the integral. Solutions for similar equations
. the assumptions concerning temperature and compressibility applied to flow in oil wells ~ave been given by CattelP and
would be eliminated. The third assumption, that the kinetic Eilerts, Smith and Cook4 • The graphic integration may be
energy change is negligible, is justifiable for rates of flow done with a planimeter or other suitable means.
used in testing gas wells. However, the exact implications of
- The data necessary for evaluating the friction coefficient' for
this assumption may be found by solving equation (18) for
a given flow rate in a gas well by means of equation (ll) are
a specific well problem.
illustrated in Table I. .
Equations (23) and (24) are more covement than equation
(ll) fo~ calculating subsurface pressures in a flowing gas
well but the value of the friction factor must be known before Table I
this may be accomplished. Data Required for Graphic Integration (Method I)
If the rate of flow Q is zero as in a static column of gas, to Evaluate Friction Coefficients and Friction Factors
equation (23) and (24) reduce to the familiar equation
For a Flowing Gas Well in the Panhandle Field, Texas
P: = e"P w Z (25)
for computing pressures resulting from the weight of a static
column of gas. Thus, the equations for flowing columns of Depth to Temperature Pressure Compressi- Specific Specific
FJowRate. Pressure at Pressure at the In- bility Volume, Volume
gas meet the necessary conditions that they must describe M ou ft per D~tum Datum dicated Factor at eu rt Squared,
24 Hours Point, Point, of Deptb,psia Dat'lm perlb v'
static columns of gas when .there is no flow. ft Point
5,904 0 14.9 244.6 0.966 1.1011 1.2168
APPLICATION OF EQUATIONS TO A ~Il\ 5,904 50 75.2 245.0 .965 1.10\8 1.2140
~! ~:~g~
1,000 80.7 253.9 .964, 1.0729 1.1511
. SPECIFIC WELL PROBLEM 2,000 85.7 263.4 .963 1.0427 1.0872
. t 5,904 2,250 87.0 265.8 .963 1.0359 l.07a!
Method I (graphic integration) I! 15,904 2,500 87.6 267.5 .963 1.0305 1.0619
Equation (ll) is particularly usefJil in evaluating the fric-
d. inside diameter of flow string, io.=5.012
tion coefficient f for gas wells. Pressures and temperatures G, specific gravity of gas (air= L 0) =0.708
must be measured in the flowing column of gas at several
depths in the well and the flow rate must be measured by suit- The flow rate was measured with an orifice meter and com-
able means to permit the application of equation (ll). As puted to a pressure base of 14.4 psia and 60°F in accordance
the foregoing equations show, the flow mechanism in a gas with Gas Measurement Committee Report No. 2 issued by
well is essentially a weight-lifting process with frictional ef- the American Gas Association. Pressures at the various depths
fects. Therefore, the specific-gravity term G must convert the in the flowing column of gas were measured with an Amerada-
volume of gas into the weight of the total fluid lifted in the type subsurface-pressure' gage, and a recording-subsurface-
flow. Consequently, the specific' gravity term must contain all temperature gage was run simultaneously with the pressure
the hydrocarbon liquid and water produced during the flow. gage. The pressure-gage. chart was interpreted according to

78 PETROLEUM TRANSACTIONS, AIME Vol. 189, 1950

 R. V. SMITH Y.P. 2777

the methods proposed by Smith and Dewees.s Wellhead pres- L

sures were constant throughout the test and were determined to ~dL.

with a calibrated piston gage while the subsurface-pressure
gage was at each indicated depth in the well. Compressibility
factors at each pressure and corresponding temperature in
the well were computed from the measured specific gravity of
the gas using unpublished data in the files of ihe Bureau of
Miries_ Specific volumes and the squares of the ~pecific vol-
umes were computed from the compressibility factors.
The specific-volume data for a flow rate of 5904 Mcf per
24 hours (14.4 psia, 60°F) of Tahle I were plotted against
pressure as is illustrated in Fig. 1. At a depth of 2500 ft, the
specific volume was 1.0305 cu ft per Ib, while at the wellhead
the specific volume was 1.1031 cu ft per Ih. At the same time,
the temperature of the gas stream decreased from 87.6°F at
a depth of 2500 ft to 74.9°F at the wellhead. The area on
Fig. 1 bounded by the pressure ordinates between the well-
head pressure of 244.6 psia and the pressure of 267.5 psia
at a depth of 2500 ft, the. dashed lines to the curve and the
curve itself, -when mUltiplied by 144 is the energy in foot-
pounds expended by 1 Ib of gas in moving from a depth of
2500 ft to the wellhead or
267•S
Energy, ft-lh per Ib of gas . t
= 144'244.6
The area between the indicated pressure limits may be meas·
ured with a planimeter or by any other method of approxi-
mation suitable for a finite area. For example, the area
within the limits outlined above as determined with a plani-
meter is equivalent to 3515.8 ft-lb of energy which is used in
the flowing progress to overcome frictional effects, to increase
the kinetic energy. of the gas as represented by a velocity
change, and to lift the lb of gas from a depth of 2500 ft to
the wellhead.
Equation (11) requires the evaluation of another quantity:
By plottiIig specific volume squared against depth as illus-
trated in Fig. 2, the quantity may be found by measuring the
area between the curve and the depth orrunate within the
proper depth limits. The actual measurem~nt of the area is
similar to that described in the preceding paragraph. The area
on Fig. 2 bounded by tb,e wellhead depth (L 0) and a depth =
of 2500 ft was found by measurement with a planimeter to
be 2839.3 units whose dimensions are volume squared times
length divided by weight. Multiplication of these units by the
proper quantities as indicated in equation (11) converts them
to work in ft-lb.
The evaluation of the remaining terms of equation (11),
the work required to lift 1 Ih of gas to the wellhead and tlle
work converted to kinetic energy requires the substitution of
the proper quantities into the equation. The work necessary
to lift lIb of gas through' a vertical distance L is L ft-lh. The
value of the kinetic energy term of equation (11) is found
by substituting the values for the rate of flow, wellhead spe-
cific volume and specific volume at the given depth, and' the
inside diameter of the pipe.
The complete solution of equation (11) for the friction
'vdP coefficient covering three successive depths for a flowing gas
well (data presented in Table I) is given in Table II. The
energy terms apply to 1 Ih of gas flowing in the well. Item 1
shows the ,!vailable ene;rgy for the flow process from the indi-
cated depths to the wellhead. The energy required to lift the'
Ib of gas to the surface and the energy used to increase the
velocity of the gas are given opposite items 2 and 3, respec~
tively. The excess of the available energy over that required
to lift the gas and increase its velocity is transforin~d into
heat by frictional effects (item 4). The remainder of the
table illustrates the method of computing friction coefficients
and friction factors from data on the amount of energy ex-
pended to overcome friction.

f-- -- --- 2.500

I
.... =\ 2.~50 /
f-- -- --\ '.Ql>O
. /
'60 ..... 1.000

..
\ /
w
~
!
• ~ ~
\
~ .

[\
1.000 %
lL
Z
I
2,000 -- f - - f - - - - /

• ,....
\ 50
o
-- f - - 1 - - - -
-- f - - f-. -I
.-0
o Loo L05 I JO
SPEC1FIC VOLUME., CU. FT. PER LB. (Y J
U$
1.00 IJO 1.20 1.30
(SPEe:IFle: YOLUIIE.GU.FT. ,~ft LIM" (."1
FIG. l-GRAPHICAL EVALUATION OF ENERGY REQUIRED FOR MOVING
ONE LB OF FLUID TO THE WELLHEAD FOR A FLOWING GAS WEI.L,
FIG. 2 - GRAPHICAL EVALUATION OF THE INTEGRAL fL v'dL
Ps
vdP. fP.. FOR A FLOWING GAS WELL
o

Vol. 189, 1950 PETROLEUM TRANSACTIONS, AIME 79

T.~. 2777 DETERMINiNG FRICTION FACTORS FOR MEASURING PRODUCTIVITY OF GAS WELLS

Table II Table III

Evaluation by Graphic Integmtion (Method I) of . Distribution of Available Energy for a Flowing Gas
Frictbnal Effects, FrIction Coefficients and FriCtion Well in the Panhandle Field, Texas
Factors for a Flowing Gas Well in the
Panhandle Field, Texas
Pcr Cent for Flow from Indicated Depths, ft·
Depth considered, ft 209 0 2250 2500 Distribution of Energy
2,000 2,250 2,50J
(I) Available energy, Lifting .....•.........•..........•...... 68.87 68.96 71.11
Khetic ....•...•.......•.•............. .05 .05 .05
Ps 28.84
144 f
Pw vdP, ft-Ib 2903.8 3,262.9
,
3515.8
Frictional dI,cts .•................•.•..
Total or available energy .........•...•..
31.08
100.00
30.99
100.00 100.00

(2) Energy to lift gas, L., ft-Ib 2000.0 2250.0 2500.0

(3) Energy converted to kinetic Employing the same data pre::ented previously to illustrate
Q'G' (v..,' -VB') the use of Meth<?d I, with e3timated effective temperatures and
energy = ft-Ib 1.4 1.6 1.7 compressibilities', friction factors may be computed by means
2.5515 x IO"d'
of equation (23). The required data are illustrated in Table
(4) Frictional effects = IV; which contains part of the data presented previously in
f Q'G" fLv'dL, ft-Ib
Table I. In addition, Table IV shows effective temperatures
53.157 x IOOd" 0 and compressibilities estimated for the flowing column of gas
= (I) - (2) - (3) 902.4 IOll.3 1014.1 to the indicated depths.
Q'G2 Exact methods of estimating the 'effective temperature and
(5) 103.92 103.92 103.92 compressibility to describe adequately the behavior of a col-
53.157 x IOOd"
. umn of gas are not known. If the temperatures measured at
(6) fLo v'dL' 2302.4 2572.5 2839.3 different depths ·in the flowing well (Table I) are plotted as
a function of depth on regular coordinates, a curve is obtained
_(7) Friction coefficient. which changes from a straight line near the surface to a
f = (4)/(5) x (6), curved line at the lower depths. The values of effective tem-
dimensionless 0.003772 0.003783 0.003437
peratures pre:;:ented in Table IV were obtained by determin-
(8) Friction factor,
ing the area within the limits of zero depth and the depth
under consideration between the depth-temperature curve and
. f I , dimensionless 16.28 16.26 17.06
a given reference temperature; the measured area was then
" f . divided by the spe~ific depth under cOD3ideration to obtain
an effective temperature.
Note: Well was flowin~ at the rate of 5904 Mcf per 24 hours
(14.4 psia, 60°F) through a 5.012 in. inside diameter
Table IV
flow string.
The percentage distribution of the available energy among Friction Factors and Data Required for Their Com-
friction, liftIng, and velocity increa£e is shown in Table III. putation by Mathematical Integration (Method II)
For this particular well producing gas through a 5.012-in.- fDr a Flowing Gas Well in the Panhandle Field, Texas
inside-diameter flow string at a -rate of 5904 Mcf per 24' hours
approx.imately 30 per cent of the available energy was con- Friction
Depth to Effective Effective
sumed in overcOIDlng friction; the remaining 70 per cent was· Flow Rate,
M eu ftp_r
PriSsure
Datum
PrESSure at
Indicated
Tellip rature Collipr ssibi-
to D pth, lity to Dlpth, Factor, ~+
24 Hours Point, ft D.pth, psia OF Dimensionless Dimensio'llSs
m:ed to Eft the gas to the wellhead, and 0.05 per cent of the
energy was med to increa£e the velocity of the gas. In com- 5,904 0 244.6
5,904 50 245.0
parison with frictional effects and lifting requirements "the 5,904 1,000 253.9
5,£04 2,000 263.4 80.7 0.964 16.27
loss due to velocity increa=:e was negligible. 5,904 2,250 265.8 81.3 .£64 16.25
5,904 2,500 267.5 82.0 .964 17.05
The difference behveen frictional effects as reflected' by
energy, friction coefficient or friction factor as shown in d, inside diamcter of flow string, in.=5.012
G, spzeific gravity of gas (air= 1.0)=0. 708
Table II for the same flow rate in the same well but from
different depths probabiy can' be attributed to errors in the
measurement of pressure. Other methods of arriving at an effective temperature bave
been devised but the superiority of anyone method is not
known.
Method II. (mathematical integration) The effective compressibilities given in Table IV are those
computed for the gas under pressure and temperature condi-
Equation (23), (or equation (24», which was obtained by tions at the midpoint of the flowing column of gas.
mathematical- integration of the energy-balance relationships Substitution of the data shown in the first five columns uf
is in conven:ent form for computing friction factors for flow- Table IV, with' the values of the inside pipe diameter and the
ing gas wells if the flow rate and the pressure at a given depth specific gravity of the gas, into equation (23) permits calcu-
in the well is known. lating the friction factor shown in the last column of Table

80 PETROLEUM TRANSACTIONS, AIME Vol. 189, 1950

 R. V. SMiTH T.P. 2777

IV. For example, after substituting the values given for the assumption having been made that the flow in a gas well is
depth of 2500 ft in Table IV into equation (23), the friction isothermal and at a constant compressibility, even though
factor may be compnted by completion of the indicated the temperature and compressibility of the fluid vary through
arithmetical operations in the following expressions; wide limits.
. 708 x 2500 The graphic integration of the energy-balance equation
s =
exponent of e =
0.037504 - - - - -
.964 x 541.6 (Method I) provides a means of determining the energy avail-
able for lifting the fluid from the well and overcomiRg fric-
s = 0.12714
e S
=
e·l.."714 =
1;1356 tion as the integration follows the expan~ion of the fluid _along
the path determined by temperature, compressibility and phase
changes. Thus, condensation of liquids and wide variations in
5,904,000 = 19,686 ' ( 1
temperature and 'compressibility are taken into account with-
.964 x 541.6
out the 'necessity of making assumptions that mayor may not
apply. Method I, therefore, is particularly applicable in cal-
~
267S -1.1356 x 244.6-2
5.012" ( culating friction factors for deep high-pressure and high-tem-
1.1356-1
perature gas wells.
_/l The accuracy of Method I depends upon the efficiency of
or " f = 17.05 the graphic integration. The curVes, illustrated by Fig. 1 and
2, should be plotted on suitable coordinates so that the areas
Values·of the number.e raised to powers may be taken from are large enough to permit accurate measurement, with a
mathematical tables, computed with logarithms or calculated planimeter. If a planimeter is not available, one of several
by means of the infinite series: methods of estimating areas may be used.
s' S3 s· , The mathematically integrated equation (Method II) as-
e S = 1 + s + - + - + - +'
2 6 24 sumes a definite path for the expansion of the fluid in the
Previously, friction factors for the same depths and basic well with both the temperature and the compressibility fiXed
data were computed by· Method I, involving graphic integra- at constant values. Its accuracy -depends upon the degree to
tion, and ,,,ere presented in Table II. Friction factors .com- which the effective temperature and c~:tp.pressibillty approxi-
puted by Methods I and II and shown for comparison in mate conditions in the well. Method II provides a convenient
Table V indicate that by either method virtually identical and practical method for calculating friction factors from data
results a.re obtained. Data shown in Table V indicate for the on flowing gas wells and lends, itself readily to the calculation
given eX~!llple that the estimation of effective temperatures of subsurface pressures in a flowing gas well which cannot be
and compressibilities by Method II gives accurate results. accomplished by Method I.
The friction factors, presented in Table V, agree closely
for depths of 2000 and 2250 ft but have a much higher value
for a depth of 2500 ft; the difference probably is the result ACKNOWLEDGMENTS
of error in pressure measurement at a depth of 2500 ft. Pres- The Bureau's study of the productivity of gas wells is being
ent-day knowledge of friction factors for gas wells is too carried on in cooperation with the American Gas Association.
limited to permit conclusions as to the accuracy of data shown The assistance given by E. L. Rawlins, chairman" Ga:s-Well
in Table V, which are presented only to show the close agree- Deliveries Subcommittee of the Natural Gas Department,
ment of friction factors calculated by two different methods. American Gas Association, is gratefully acknowledged. .
The investigations are un<!er the general supervision of
Table V R. A. Cattell, chief, Petroleum and Natural Gas Branch, Bu-
reau of Mines, Washington, D. C., and H. C. Fowler, super-
Friction Factors Computed by Method I (Graphjc
vising engineer, and C. J. Will1elm, engineer in charge of
Integration) for a Flowing Gas Well in the,Panhandle
production research, Petroleum Experiment Station, Bartles-
Field, Texas
ville, Okla.
Acknowledgment is due the management of Phillips Petro-
Flow Rate,
M cuftper
Depth
Considered,
Friction Factors leum Co. for ~anting the writer permission to publish certain
24 Hours ft Method I Method II data contained in the report. C. B. Carpenter, R. H. Espach,
5,904 2,000
2,250
16.28
16.26
16.27
16.25
J. W. Ferguson, R. V. Higgins and H. C. Miller, of the Bureau
5,904
5,904 2,500 17.06 17.05 of Mines, offered helpful suggestions in the preparation of the
report. The comments of C. W. Binckley, of Phillips Petroleum
Co., who reviewed the manuscript, are appreciated. S. E. Dan-
SUMMARY iel prepared the illustrations. F. G. Archer assisted in the
Two methods of calculating friction factors from data taken field testing.
on flowing gas wclls are presented in this report. Method I,
which requires flowing pressures and temperatures at several
REFERENCES
points in the flowing gas stream, provides a solution for' the
energy-balance equation by graphic integration over the actual 1. Rawlins, E.' L., and Schellhardt, M. A., Back-Pressure Data
path of the expansion of the fluid in the well. In' Method II, and Their Application to Production Practices, Bureau of
the energy-balance equation is integrated mathematically, the Mines Monograph 7, 1936, 210 pp.

Vol. 189, 1950 PETROLEUM TRANSACTIONS, AIME 8l

T.P. 2777 DETERMINING FRICTION FACTORS FOR MEASURING PRODUCTIVITY OF GAS WElLS
2. Cattell, R. A., General Equations for Flow of Fluids, ap-
pended to: Lindsly, B. E., A Bureau of Mines Study of a
"Bottom-Hole" Sample from the Crescent Pool, Petroleum
Engineer, vol. 7, part 3, April 1936, pp. 38-42.
3. FergUson, J.' William, "Calculation of Back-Pressure Tests
on Natural Gas Wells," Oil and Gas Jour., vol. 37, Jan. 19,
1939, pp. 47-53. ' effective compressibility and temperature found at the mid-
4. 'Eilerts, K., Smith, R. V., and Cook, A. B., Properties of a
Petroleum-Reservoir Liquid and Its Residua with Applica-
tions of the Data to Production Problems, Bureau of Mines
Report of Investigations 3474, 1939, 32 pp.
5. Smith, R. V., and Dewees, E. J., "Sources of Error in Sub-
surface-Pressure Gage Calibration and Usage," Oil and Gas
Jour., vol. 47, Dec. 9, 1948, pp. 85-98.
DISCUSSION
By C. W. Binckley, Phillips Petroleum Co., Bartlesville, Okla.:
The vertical flow equation developed by Smith in his paper
and described as equation (24) is the equation generally used
to compute bottom-hole flowing pressures in gas and conden-
sate wells. In the use of the equation,' it is necessary to first
obtain friction factors, an effective temperature factor of the
gas stream and an effective compressibility factor. The 'accu-
:racy of the computed bottom-hole working pressures is pri-
mariJy dependent upon the accuracy of the friction factors
and the validity of the methods used to determine the effective
temperature and effective compressibility factors:
Several engineers. have recently attempted to develop a
generkI vertical flow equation for gas in which the temperature
and compressibility are included as functions of depth and
pressur~. Such an equation-could be used to compute bottom-
hole flowing pressures by application of suitable friction fac-
PETROLEUM TRANSACTIONS, AIME
tors. An equation of this nature has n~t been published and
I do not believe one has been developed in a usable form.
Smith gives us a different approach to the problem. ~e
describes a graphic integration method for determining fric-
tion factors by use of well test data. He then proceeds to sh~w
that by use of the general vertical flow equation, assuming au
point of the flowing gas column, that a, very close check is
obtained for friction factors determined by the equation and
by the graphic integration method. The check on friction
factors was obtained 'by using data taken on a well with a
relatively low pressure and low temperature flowing gas
column.
The application of the procedures described in the paper to
the determination of effective compressibility for high temper-
atU1:e and pressure gas columns of wells is not discussed by'
the author. This is an important point in improving the results
of computed bottom-hole flowing pressures. For example, the
graphic integration method can be used to establish friction
factors. The friction factors can be used in connection with
the general vertical flow equation and temperature gradient
data to establish the effective compressibility. If the nature
of the eff~ctive com.~ressibility can be ascertained, we will
have a better approach to the computation of reliable bottom;
hole flowing pressures. In fact computed pressures should be
as reliable as those measured with the bottom-hole pressure
gage.
Furthermore, by proper correlation of experimentally de-
termined friction factors with Reynolds numbers or other in-
dexes and by use of a method for determining effective com-
pressibility, it is probable that accurate bottom-hole flowing
pressures' can be established for all wells in a field by obtain-
ing basic test data on a few representative wells. * * *
Vol. 189; 1950