Chapter 4

Multiple-Rate Testing

4.1 Introduction 162.6 Bit " ' kh

The drawdown testing and analysis methods in Chapter 3 and intercept
b’ = mTlog/.—-----------) - (4.2)
require a constant flow rate; however, it is often impractical or
3.2275 +
impossible to maintain a constant rate long enough to complete a 0.86859rl .
drawdown test. In such a situation, multiple- (variable) rate testing
and analysis techniques are applica ble. A multiple-rate test may L W,-vv J
range from one with an uncontrolled, variable rate*1'* to one with a .............................................................(4.3)
series of constant rates,3,4 to testing at constant bottom-hole
pressure with a continuously changing flow rate.5 Pressure-buildup Multiple-rate transient test data should appear as a straight line
testing1 (Chapter 5) is a special kind of multiple-rate well test. when plotted as
Almost any flow-rate change can be analyzed as a well test by
using the concepts presented in this chapter. P'JLPxL vs V rfezfltll logo -
Accurate flow rate and pressure measurements are essen tial <7JV L <?v
for the successful analysis of any transient well test. Rate j-1
measurements are much more critical in multiple-rate testing, To make that plot correctly, it is important to understand that the
however, than in conventional, constant-rate well tests. Without rate corresponding to each plotted pressure point is qN — tlte last
good flow-rate data, good analysis of multiple-rate tests is rate that can affect that pressure. As time increases, the number of
impossible. rates may increase and the last rate may change; but each
Multiple-rate testing has the advantage of providing transient pressure point is identified with the rate occurring when that
test data while production continues. It tends to minimize changes pressure was measured. There may be several pressure points
in wellbore storage coefficient and phase segregation (humping) associated with a given rate.
effects and, thus, may provide good results when drawdown or
buildup testing would not.

4.2 A General Multiple-Rate Test Analysis Technique

Fig. 4.1 schematically shows a variable production-rate
schedule. Although flow rate may change continuously, it is treated
as a series of discrete constant rates for analysis purposes. The
step-wise approximation improves as the time intervals become
smaller. Section B.7 presents the derivation of a general equation
for pressure behavior caused by a variable flow rate. The approach
presented here requires that the log approximation to the line
source (Eq. 2.5b) applies. Then,

ELT-OSL - m' V
1
ffa -ft-i) iog(, - , )] + y. pf L <?.v
J
.............................. ..................................... (4.1)
Eq. 4,1 is the equation of a straight line with slope
Fig. 4.! Schematic representation of a variable production-rate schedule.
 32 ADVANCES IN WELL TEST ANALYSIS

Examples 4.1 and 4.3 illustrate how the summation term in this 0.4. We do not recommend using this analysis technique for
plotting technique is calculated. variable-rate tests unless the variable rate results only from
Once the data plot is made, the straight-line slope and intercept wellbore storage, in which case the surface rate is constant.
are measured. Permeability and skin factor are estimated from the
slope and intercept data using Eqs. 4.2 and 4,3, rewritten as . Example 4.1 Multiple-Rate Drawdown Test Analysis Production
162.6Bfi rate during a 48-hour drawdown test declined from 1,580 to 983
m 'h STB/D (Fig. 4,2). Rale and pressure data appear in Table 4.1.
and • (4.4)
Reservoir data are
j= 1.1513
Pi = 2,906 psi D = 1.27 RB/STB
108
....................................................................
+ 3.2275J .
Fig. 4.3 shows flowing bottom-hole pressure vs log of flow
time. That plot, which is normally used for a constant- rate
drawdown test, neglects rate variations. Nevertheless, we use the
straight line from 3 to 12 hours to estimate permeability. Using an
(4.5)
average rate of 1,450 STB/D
p = 0.6 cp h «
The analysis procedure is direct and simple, but the computations for the first 12 hours of the
40 ft.
required to make the data plot can be tedious. The analysis has the test, a slope m = —160
disadvantage that the initial reservoir pres sure, Pf, and the entire psi/cycle, and Eq. 3.9,
flow-rate history must be known; frequently, they are not. As
( .-™»)=28,lrad,
discussed in Section 4.5, the analysis technique may be modified ( —160)(40)
in some situations so that Pi is not used. If the pressure is constant
Fig. 4.4 isa plot of (Pi - p,rt)lqx vs
during a test and the rate declines, Eqs. 4.1 through 4.5 generally
are not used; instead, the techniques in Section 4.6 are preferred.
When flow-rate variation is a result of wellbore storage, a iTHr
simplified plotting method, which does not require use of (is for this test, using the rate breakdown shown in Table 4. 1.

Fig. 4.2 Multiple-rate drawdown test rate history and its approximation. Data for Example 4.1.

Fig. 4,3 Semilog plot of bottom-hole pressure for the multiple-rate drawdown test of Example 4.1.
superposition, may be used. 6'8 In this case, one plots (pt — pUf)/qSf Table 4.1 summarizes the calculations of the quantities plotted in
vs
Fig. 4.4.
log /.The result should be a straight tine with slopem' given by
f To illustrate the method of computing the time summa tion, we
Eq. 4.2 and interceptb given by Eq. 4.3. Permeability is estimated
calculate it at 6.05 and 12.0 hours. At 6.05 hours, q — 1,440
from Eq. 4.4 and skin factor is estimated from Eq. 4,5, Ramey 8
STB/D is the third rate observed (although the point is the ninth
points out that the skin factor so calculated may be low by about
MULTIPLE-RATE TESTING 33

pressure point), so Af = 3. Computing the summation term,

N
<3, - ?I-T) i°g<f - o-i >

■ {[(1,580 - 0) log(6.05 - 0)],„,

+ [(1,490 - 1,580) log(6.05 - 2.40)],.,

+ [(1,440 - 1,490) log(6.05 - 4.80}],.,}

=
f^40 {[l'580lOg(6'°5j]'=1 + [-901og(3.65)]J=2 + [-50Iog(1.25)]J4

= J— {1,235.17- 50.61 - 4.85} = 0.819.

1,440 L 1

Thus, the point for 6.05 hours plots at coordinates (0.819, 0.738) in
Fig. 4.4. At 12.0 hours, q = 1,300 STB/D and N = 5. Thus,
34 ADVANCES IN WELL TEST ANALYSIS

N 4.3 Two-Rate Testing

— V* % ~ %-i) log(t - fj_|)
<IK t-J
j*l
= _J [(1,580 - 0) log(12.0 - 0)
1,300 L
+ (1,490- 1,580) log( 12.0-2.40)
+ (1,440- 1,490) log(12.0-4.80)
+ (1,370- 1,440) log(12.0 — 7.20)
+ (1,300- 1,370)log(12.0-9.60)]
= —L_ [1,705.11 - 88.40 - 42.87 -47.69 1,300
When a multiple-rate test consists of only two flow rates, both
testing and analysis are simplified. The two-rate test provides
information about k and s while production continues. Wellbore
storage effects are often thought to be minimized or eliminated by
two-rate tests. In fact, wellbore storage effects last just about the
same amount of time in a
FLOW TIME, HR
Fig. 4,4 Correct data plot for the multiple-rate drawdown test of Example 4.1;/ is in hours.
L

-26.61]
= 1.154.
The point at 12.0 hours plots at coordinates (1.154, 0.853) in Fig.
4.4,
Two straight lines can be drawn through the data of Fig. 4.4.
The slope of the second line is greater than that of the first,
possibly indicating transition to pseudosteady state, faulting, or a
decrease in permeability away from the well (see Sections 10.2
and 10.4). The incorrect semi log data plot, Fig. 4.3, has a
reduction in slope for / > 12 hours that might be interpreted as
increasing permeability away from the well. That is an incorrect
concl usion, however, since the slope change in Fig. 4.3 is caused
by the declining production rate.
Using the slope of the first straight line in Fig. 4.4 and Eq. 4.4,

(16261027X0^ 6
J
(0.227)(4O)

Thus, the permeability computed from Fig. 4.3 is about 107 percent
two-rate test as in a normal r,, at the start of the test. If,
too high.
TABLE 4.1—VARIABLE FLOW RATE DRAWDOWN DATA FOR buildup, drawdown, falloff, etc., instead, the well was stabilized
EXAMPLE 4.1. test. However, a two-rate test at rate */,. then Eq. 4.6 is still a
Time Rate, Pi ~Ptrf
, often can be used to prevent a good approximation if r, is
t 9 P P< Pi'f Qv
» wellbore storage increase, thus calculated from
(hou
rs)
(STB/
D)1.580
N (psi) (psi) (psi/STB/D
) 0.5589
I providing an analyzable test
1.00 1 2,02 883 0.00
1.50 1,580 1 3
1,96 938 0.5937 0
0.17 when one otherwise might not
1.89 1,580 1 1,94 96 0.6108 0.27 be possible (see Section 11.2),
2.40 1,580 1 — — — —
3.00 1,490 2 1,89 1.014 0.6805 0.51 The main advantage of a two-
3,45 1,490 2 1,88 1,024 0.6872 0.56 rate test over a buildup test is
3.98 1,490 2 1,87 1,033 0.6933 0.62
4.50 1,490 2 1,86 1.039 0.6973 0.67 that deferred production is
4.80 1,490 2 _ __
minimized.
5.50 1,440 3 1,85 1,053 0.7313 0.78
6.05 1,440 3 3,84 1,063 0.7382 0.81 Fig. 4.5 schematically
6.55 1.440 3 1.83 1,072 0.7444 0.84 illustrates the rate and
7.00 1,440 3 1,83 1,076 0.7472 0,87
7.20
7.50
1,440
1,370
3
4 1,82
_
1,079
_
0.7876 0.97
_
___ pressure behavior for a two-
rate flow test;3 either a
8.95 1,370 4 1,82 1,085 0.7920 1.00
9.6
10.0
1,370
1,300
4
5
—
1.81 1,091
_ —
0.8392 1.12
___ decreasing or increasing rate
sequence may be used. Eq.
12.0 1,300 5 1,79 1,109 0.8531 1.15
14.4
15.0
1,260
1,190
6
7 1,77
_ ___
1,131
___
0.9504 1.33
___ 4.1 may be modified to the
form presented by Russell 3 for
38.0 1,190 7 1.77 1,135 0.9538 1.35
19.2 1,190 7 — — — a two-rate test:
20.0 1,160 8 3,77 1,134 0.9776 1.42
21.6
24.0
28.8
1.160
1,137
1,106
8
9
1
___
1,75
_
___
1,150
—
1.0114 1.48
—

—
Arf=nii' jj°i;+ 7T log Ar
]

30.0 1,080 1 1,75 1,155 1.0694 1.60 +Pin|

-
33.6 1,080 1 ___ ___
36.0 1.000 1 — — — — ....................................................................
36.2 983 1 1.75 1,150 1.1699 1.78
48.0 983 1 1.74 1.163 1.1831 1.80 Eq. 4.6 assumes a constant
3 3 0 flow rate,*/,, from lime 0 to time
MULTIPLE-RATE TESTING
Fig. 4.6 schematically shows
where Vt> is the cumulative
such a data plot. Because of
volume produced since the last
the choice of the abscissa
rate stabilization.
variable, time increases from
Eq. 4.6 implies that a
right to left. At long times, the
graph ofpirf vs
data deviate from the straight
line as a result of boundary
[l08(,Jir“)+^logA']
should be a straight line with slope
and interference effects. At
and intercept
short times they have not yet
reached the straight line
because of rate restabilization
and wellbore storage effects.
-3.2275 + 0.86859 s\ Field test results indicate that
..............................................................................
rate restabilization is faster for
a rate reduction than for a rate
increase. 1*3
Once the slope of the
straight line is determined from
the data pfot,
permeability may be estimated
from
A= ................................................
JOT, h
The skin factor is estimated
from
r <h
--
( P„/(A/=0)-P,Kr \
l ”"
4|
t,----------F— At —
Fig. 4.5 Schematic rate and
pressure history for a two-rate
flow test, q, > q.,.
35
!
• -i....................................«■"'
The intercept of the data plot
may be used to estimate the
false pressure,3
V =p„„ - [rv,
(A/=0) -Plhr] (4.12)
</. -q-i
which is used to estimate
average reservoir pressure
using methods in Chapter 6.
Example 4.2 Two Rate
Flow Test Analysis3 Well A is
a flowing producer in a low-
permeability limestone
reservoir in the Permian Basin.
Pressure buildup tests in this
reservoir usually do not provide
interpretable data because of
long, low-rate afterflow
periods.3
A two-rate flow test was
run by stabilizing the flow rate
reservoir
at 107 STB/D for several days
and then reducing the flow rate
to 46 STB/D. The pressure
data during the second rate are
shown in Fig. 4.7. Other
1
pertinent data are
^ SHOUT TIME LAG MAY OCCUR
r BEFORE MEW STABLE RATE
fS REACHED
-=[*-
X 4,
1
1
____1____________________________
FLOW TIME, t, HR
Fig, 4.6
Schem
atic
data
plot for
36 ADVANCES IN WELL TEST ANALYSIS

a two- > After l The p* value may be used with

rate Russell
h material in Chapter 6 to
flow .3
test, q, , estimate average drainage-
B . region pressure.
= 6 =
1.5 c
RB p 3
/S ,
In certain circumstances,
TB < I
the plotting technique
p„. j 6
suggested by Eq. 4.6 may be
XA > 9
simplified.9 Whenfi >>Af, log (/i
r= p
+ At) rr log /1 and log[(ft +
0) = s
Ar)/At] — log ft - log At. Making
= i
these approximations in Eq. 4,6
3,1 0 and rearranging gives
18 . (
psi 0 F pHf = m" log A/ +pim....................................................... (4.13)
h 6 i Thus, for a two-rate test with/,
= g >> A f, a plotofyv/vs log At
59 = , should be a straight line with
ft slope
c 0 4
,= . .
9.3 2 7
2x )
] f
0"s t p
psi i i

-1
n m

Ve ,
= ' =
26, =
40 3
0 — ,
ST 9 5
B 0 1
f, = 0
(2 p
6,4 s p
00 i s
) / i
(2 c
4)/ y (
10 c F
7= i i
5,9 e g
2! .
ho (
urs F 4
(E i .
q, g 7
4.7 . )
) .
/ 4 m" = ^l62-6(^2-^i)fiM kh (4.14)
Permeability is estimated from Eq. 4.10: (I62.6)(107)
i , and intercept
(I-5)(Q.6)
7
= ) (-90X39) ' ’
2
The skin factor is estimated from Eq. 4.11: Pint-Pi + (<7a
0 p “ 4i>
(LRz-im (
L \ 107 — 46/ \ -90 ! - 3.2275 + 0.86859
h
(4.15) J +

---) + _
log
2.9
3.2275]
3
(Q,06)(0.6)(9.32 I0- X0.
X ■Thep2).f2 vs) log
J At data plot may be used to estimate reservoir
tt

permeability from
= -3.6
FLOW TIME, At, HR
p — _ J62.6 (ffs — ») Bp (4.16)
Fig. 4.7 Two-rate flow-test data plot for Example 4.2. After
Russell.3 m*h * ....................
and to estimate skin factor froma

f Plhr ~P«t( A' = 0) _ /

L m*
i]
g
U
+ 3.2275 (4.17)

The analysis suggested by Eq. 4.13 is much faster and

simpler than that suggested by Eq. 4.6. However, the ap -
proximations in Eq. 4.13 cause errors in the results. The error in
the permeability is10

r — actual qx (T*-\)-qtT* ’.................................

while the error in skin factor

<?»
is
<4.18)
■Tactual
$E ~

= —1.1513 Ip,hr ~ —
0)]
MULTIPLE-RATE TESTING 37

should be used. Using Eq. 4.5,

-0.00991 . 0.148

Example 4.3 Two-Rate Flow . j + 3.2275J

Test,
Variable-Rate Case 11
The data in Table 4.2 are
for Well X of Odeh and Jones.
*1 Well X was stabilized for
several days aT<y„ = 1,103
STB/D. Other data are
B = 1. ORB/STB p = 1,0 cp
ru. = 0.26 ft c, = 1.4 x 10"5
psi-1
Parameters to be plotted
are calculated in Table 4,2,
The last value in the right-hand
column is calculated as
follows, (In this case, N = 3
since there are three rates, but
five pressure data points.)
3
Sl—. log(A/ - A/,„i)
<7JV — <?o
= -6.5.
Odeh and Jones11 stale
that pressure drawdown and
buildup tests on the well
indicated kh 2: 1,100 md ft and
*• ^ -7. They also state that
they used the simplified two-
rate flow-test analysis method
(Eq. 4.13) with an average rate
of 832 STB/D. The result was
kh ~ 1,360 md ft and x ~ —
6.3. Thus, the 10-percent rate
variation may have resulted in
a 23-percent error in the value
calculated for kh.
Although the production
time at ihe stabilized rate is not
known for this example, the
range of errors caused by
using the simplified two-rate
analysis may be estimated by
assuming values of the
stabilized production time,
(This requires assuming that
analysis by a two-rate
technique with a constant
second rate is adequate for the
data of this example.) From
Eq. 4.20, for r, = 10 days, and
for Ihe maximum test time of
At = 10 hours,

N the summation on the right

P,rf(At = 0) ~ Pafi&O - mf
XT' j~flj ~ ?J-I x Vs
~ 4o
log (A/ - Arj_,)J + b', ................
-..................................................
(4.21)
where A / is the time from the
start of the varying second flow
side, permeability and skin The error inA7i is estimated
s — 1.1513 from Eq. 4.18 with thicknesses
factor may be estimated from
Eqs. 4.4 and 4.5. The analysis included in the numerator and
- log /----------- denominator;
UO.Ii 61.0
)(1.0)(1.4 x 10 ^(0.26)’
Thus, if r, is indeed 10 days or more, the 23-pcrcent error in£/i when using the simplified two-rate analysis method is largely caused by the
rate variation, not by the inherent errors in the simplified two-rate analysis technique. How-
technique suggested by Eq.
4.21 also applies to variable-

rate. This equation is much
likeEq.4.1 [if<?0 = 0and/>*./(A/
= 0) = Pi, it is identical to Eq.
4.1 ]; the slope, m\ and
intercept, b', are given by Eqs.
4.2 and 4.3, respectively. If the
data are plotted as the quantity
on the left side of Eq. 4.21 vs
-
"
rate test analysis if the well has
been stabilized for a
substantial time at rate q0.
If pseudosteady state is
achieved in the first flow period M ’
of a two-rate flow test, the
analysis techniques of Section
4.5, Eqs. 4.30 through 4.32,
f "

h = 18 ft 61.0 md. LOG

N
A =28.3 x I0*sqfi ~ 650 4N-90 z ‘

acres <£“11 percent. i=i

38 ADVANCES IN WELL TEST ANALYSIS

1 s
?
a
0
5
Fig, 4,8 Two-rate flow test,
variable-rale case, data plot
for
Example 4.3- After Odeh and

♦ Jones.11

M
b

*
<
'
°
=
[{1.3587X
1)] + [<-
0.1794X0.
7782)] +
[(-
0.1794X0.
3010)]
= 1.165.

The data plot (Fig. 4.8)

indicates Ihe slope, m', is
0.14B psi/(STB/D cycle) and
the extrapolated intercept, b',
is -0.00991 psi/(STB/D). Using
Eq, 4.4,
, (162.6X1X1)
(0.148)(18)
and
kh = 1,100 md ft.
 MULTIPLE-RATE TESTING 39

4.2—SUMMARY
TABLE
CALCULATIONS OF
FOR
EXAMPLE 4.3. After
Jones.1 ‘ Odeh and
£
Aqx = Ap
<7v ~ Ap = Pi
~' log (
At Qi. P*iUf) 0} - p„.£At) AQ
■ “ Qi.
N j Qi
(STB/ (STB/D
Qi ~
(STB/D lp (psf)
X
(psi/ST
[hours) D)
1,103 0) )
_
$i)
3.630 0 B) _
*
_
0 0 0 8 -303 -303 3.654 -24 0.0792 0.602
1 1
8 -263 40 3.660 -30 0.114 0.851

8 69
4 8 -263 40 3.665 -35 0.133 0.949
2 2
8 -223 40 3.666 -36 0.161 1.171
i8 -223 40 3.668 -38 0.170 1.165
3 3 10 6

ever, the following table m,' - -162.6 qfiy..........................................................

that computed from the buildup modifications presented in this
indicates that the error could kh test because the correct data section apply when pressure
be entirely caused by the and intercept plot and the correct equation decline at the test well is
analysis technique rather (Eq. 4,25) were not used in the caused by production from
than the variable rate — Error in Two-Rate analysis. That is particularly other wells or from the test well
if production time at the Analysis (percent) true in drillstem-test data itself; the cause of the
first rate is less than 160 50 30 21 16 13 analysis. pressure decline is not
about 4days. 10 8 important.
Fig. 4,10 illustrates
Production Time
bottom-hole pressure from a
at First Rate {days) 4.5 Developed Reservoir
Effects tested well in a developed
1
reservoir. The solid line is the
2 When bottom-hole observed pressure behavior,
3 pressure is declining as a while the dashed line
4 result of withdrawals from the represents the pressure that
5 test well or from other wells in would have been observed
3 the reservoir, the analysis
8 »*-**•*' methods presented earlier in
had there been no flow-rate
change at the well at time The
10 this chapter must be

4.4 Drawdown Testing
After a Short Shut-In
It is common practice to run
a drawdown test after a shut-in
period (pressure buildup test).
If the shut-in is too short for the
well pressure to stabilize at
average reservoir pressure,
the drawdown-test analysis
techniques of Sections 3.2 and
3.4 should not be used;
instead, a multiple-rate-type
analysis is applicable. Fig. 4.9
schematically shows a rate
history For a drawdown test
after a short shut-in. Writing
Eq. 4.1 for such a test and
bfcM
pressure along the dashed line
modified. 12,13 Such modified
for the “no test” case is called
analysis techniques become
pwext. Using the approaches
-3.2275 + 0.86859 increasingly important as the
outlined by Slider12,13 (similar to
J].....................................................................
depiction rate increases and as
those in Section 3.4), Eq. 4.1
test duration increases. The
Reservoir permeability can be may be modified to
estimated by solving Eq. 4.23. PtcextOl d~ At) Pw f( t i ~h Af) reservoir with no other
The equation is the same as < operating wells, Eq. 4.26 is
the normal drawdown-test i identical to Eq. 4.1. A similar
equation (Eq. 3.9), but the data equation can be used in place
plot is different. The skin factor of Eq. 4.21 with p,cext replac-
may be estimated from the N «ng/V/<Ar=0).
data plot with For the two-rate test of
N
s= 1.1513 rPm q
> ~qi~x log (At—Alj-i)l 4* b\
Section 4.3, the applicable
equa* tion is
to J
4 A>.,\ PuaiUt + AO - PwrO, + AO
L '«a' ........................... .............................. ?
\ At*, / = [log(A±A ) +|
MogA,]-APln,.

rearranging gives whercm'andt' are given by

Eq. 4.25 is similar to Eq. 3.10
U for estimating skin factor from
a normal drawdown test, with
................................................................
the exception of the additional
Thus, a plot of pa vs }(9,/^3) logarithmic term. The skin
log[(/, + Atg, + At)/(Atxi + A/)] + factor computed from a
log Ar} should yield a straight drawdown test after a buildup
line with slope test often does not agree with
Eqs. 4.2and4.3, respectively.
The analysis is the same as
outlined in Section 4.2 except
that the pressure quantity used
is puext — pWft the vertical distance
between the solid and dashed
curves in Fig. 4,10. If the test
starts at time zero from Pi in a

Fig. 4.9 Schematic rate history for a drawdown test after a short shut-in.
 40 ADVANCES IN WELL TEST ANALYSIS

ftrextfri + At) ~ PvAh + At) Permeability would be (4.27) where

= Aplhr -m" log Af. (4.34) is estimated from the
estimated from Eq. 4.16 and In this case the data plot has
slope of a Cartesian plot of picf
skin factor would be estimated slope mf, given by Eq. 4,8, and
............................................................... vs / before the second rate
from an intercept starts. By using Eq. 4.33, Eq.
This indicates that a plot of the 4.30 may be rewritten:
extrapolated pressure differ- ..................................................................
Apim = ">i'— flogfv-i—r) - />*./(/, + Ar) - m* Ar = Aplhr +
ence vs log Af should be a 3.2275 m"log Ar........................................................................ (4.35)
This analysis for <?i L
straight line with slope — —
pseudosteady-state flow \«KVV/
wT, where m" i$ given by Hq.
during the first rate can be
4.14. The intercept is
simplified even further, since + 0,86859,] .........................................................

Reservoir permeability is
Puvxi(fi 4 At) = p,r/(Ai = 0) +
m* A/, estimated from Eq. 4.10 and
-3,2275 + (4.33) skin factor from
0.86859jj .....................................................
,»1.1513 It —

log (

L_J)
+3.2275] .
L m. q2
\<t>
p^,r^/ J

....................................................................

If pseudosteady-state
conditionsexist during the
first flow
period in a two-rate flow test,
then Eq. 4.27 becomes

FLOW TIME* HR

Fig, 4.10 Schematic pressure

history for variable-rate testing
in a
developed reservoir.
 MULTIPLE-RATE TESTING 41

wherem* is given by Eq. 4.34 and mN is given by Eq. 4.14. A plot of (/V/ — rn* Ar) vs log Ar would be a straight line with
Fig. slope
4,1 ISm" and representation of rate anti pressure histories
chcmutic
intercept during a constant-pressure test.

The skin factor may be estimated from Eq. 4.17 with Apliir in place of p,hr.
The main difference between equations in this section and in Sections 4.2 and 4.3 is that this section does not assume (4.38)

the system pressure is infinite-

acting at the time the two-rate
or variable-rate test begins. In
particular, if pseudosteady-
state conditions exist at the
start of the test, the analysis
technique can be expected to to wellbore storage, causing constant-pressure tests, so it is
give results different from those desirable to have a method
bottom-hole pressure to vary
of the techniques in Sections during the test. for analyzing such tests.
4.2 and 4.3. Fig 4.11 schematically In a manner similar to that
represents pressure and rate used to express pressure as a
4.6 Constant-Pressure behavior in a constant-pressure function of flow rate and time,
Flow Testing drawdown lest. Such a test is we may express flow rate as a
The transient behavior of a seldom performed since it is function of pressure drop and
well operating at constant much easier to measure pres- time by
sand-face pressure is sure accurately than it is to
analogous to that of a well measure flow rate accurately. (4.37)
operating at a constant flow However, constant-rate tests
rate. In a constant-pressure may inadvertently become Dimensionless time has its
flow test, the well produces at a usual definition:
constant bottom-hole pressure
t„ = 0.0002637 ___
and flow rate is recorded with <b^,r„"
time. Constant bottom-hole
pressure test data are not The principle of superposition
influenced by wellbore storage. may be used with Eq. 4.37 to
However, if the surface compute the flow rate resulting
pressure is maintained from a series of pressure drops
constant, the frictional pressure during pressure-controlled flow.
drop in the flow tubing may act The calculations are analogous
in a manner similar to those given in Section 2,9.
Fig. 4.12,* showing
dimensionless rate as a
function of dimensionless lime
for an infinite-acting system,5 is
useful for test analysis by type-
curve matching. The technique
is similar to that outlined in
Section 3.3. Briefly, test data
are plotted on tracing paper
laid over the grid of Fig. 4.12,
with flow rate on the ordinate
and the time on the abscissa.
The tracing-paper data plot is
moved horizontally and
vertically until the data match
the curve in Fig. 4.12. Then, qM
and t%/ are read from an
arbitrary match point on the
42 ADVANCES IN WELL TEST ANALYSIS

tracing paper and Ufi,)ss and (/ 4.12. Reservoir permeability is

„)„ arc read from the estimated from
corresponding point on Fig. *See footnote on Page 24,
 Using Eq. 4.39,
(141.2)(1.0)(1,720)(L35)
= 6.4 md,
(it000)(0.27)(190) an error of 1.5 percent. Reservoir
porosily*compressibility is estimated from Eq. 4.40:
0.0002637 (6.4)<I90)(1)

(fi
Cf
h (1.35)(1)2(600) =
3.96 x I0-4 ft/psi,

4c = 3-** '°-4= 2.08 x io** psr

= 190
an error of 1.5 percent.
An alternate analysis uses Fig. 4.14, a plot of (t/^r) vs log t.
The slope of the straight line drawn through the data for 0.! <r <
10 hours is
m„ = 1.7 x 10 4 (D/STB)/cycle.
Using Eq. 4.45,

k- (162.6X1-0X1.35) = 6.8 md,

(1,000)( 1.7 X I0-4)(190) an error or 4.6 percent. The skin
factor is estimated from Fig. 4.14 and Eq. 4.46. From Fig. 4.14.
=
0.000578 D/STB.
\M 000578
s- 1.1513
LI x lO*4

-'
-7
6.8
\(1.35)(2.08 x I0*5)(1)!
J+ 3.2275J

H
= 0.28.

This compares with the actual value of 0.

As opposed to constant-rate testing, the analog to pseudos

ready-state flow does not develop during constant- pressure
testing. When boundary effects influence behavior during the
constant-pressure test, there is a rapid decline of flow rate caused
by declining reservoir pressure. Flow rate goes to zero as
reservoir pressure approaches the wellbore pressure.

Example 4.4 Constant-Pressure Tes ting in an 4,7 Reservoir Limit Testing When Rate Varies
Infinite-Acting Reservoir Section 3,5 discusses reservoir limit testing for constant- rate
The flow-rate data shown in Fig. 4.13 are From a simu lated production; however, it may be difficult to maintain a constant flow
constant-pressure drawdown test. Data used to simulate the test rate during long production periods. If flow rate varies in a cyclic or
are oscillatory manner, reservoir limit testing techniques still can be
4>ct = 2.05 x 10“6 psi-1 Pi used.14,15 The analysis technique is similar to that for constant-rate
k = 6.5 md p =
— p,vf = 1,000 psi rw = 1 ft s = reservoir limit tests, but the results are less accurate.
1,35 cp /1=190 ft
0. To analyze a variable-rate reservoiF limit test, one plots
B = 1.0 RB/STB
The data of Fig. 4.13 were observed flowing pressure vs rime on Cartesian paper. Pressure
type-curve matched to Fig. points must be segregated by the rate occurring when the
4.12, Match-point data are pressure measurement was made. Fig, 4.15 is such a plot for a
qM= 1,720 STB/D (qB)M = 0.27 waste-water injection well; the three sets of pressure points occur
tM = 1 0 hour (tD)M = 600. at three different injection rates. Such a data
 FLOW TIME, t, HR
Fig. 4.14 Plot of \/q vs log / for constant-pressure drawdown test of Example 4.4,
plot should have a straight-line section. Because of flow-rate injection is with one, two, or three pumps, so the rate is -5,140,
variation, there is actually never a definite straight line, but a least- —10,280, or -15,420 STB/D. Pressure and rate data are reported
squares Straight line usually can he fit to the pressure points by month only without indication of how long the rate had applied
observed at one of the rates. The slope, m*, of the straight line when the pressure was measured. Cumulative fluid injected is
may be used to estimate reservoir drainage volume: known accurately. The average injection rate is -9,660 STB/D for
-0-23395 qB ................................................(4.47) the period shown in Fig. 4.15. The formation volume factor, B, is i .
m*ct 0; system total compressibility, ct, is unknown, but is estimated to
be about 5 x 10_6psi“l.

Fig, 4.15 Long-term, variable-rate reservoir limit test of F.xample 4.5. Five years of injection history for a disposal well with three injection
rates; q = —9,660 STB/D during period shown. After Earlougher. 14
Eq. 4.47 is identical to Eq. 3.35 except that the over-all average
flow rate is used. Even though the pressure points are segregated
by the rate in effect at the time the pressure was measured, the
rota/ average flow rate is used to estimate reservoir volume.
Example 4.5 Variable-Ra te Reservoir Limit Test
Fig. 4.15 shows pressure data from Ref. 14 for the last 5 years
of the 11 -year life of an industrial waste-disposal well. Although
the data are for injection, the methods of this section can be
applied by using a negative rate (see Chapter 7),
The well, with casing set at 1,800 ft, is completed open hole to
about 3,000 ft, just into basement. Only a part of the 1,200-ft
section is porous and permeable, but porosity and net interval are
not known. The injection horizon is known to be of large extent.
The lines in Fig, 4,15 are least-squares fits to the pressure since most data points are fori? * —10,280 STB/D.
data for each of the three rates. The slopes, m*, are +2.62 x 10“3 These calculations require that flow be pseudosteady state. A
psi/hr, + 2.15 x 10~3psi/hr, and +3.16 x 10~3 psi/hr for the —5,140, long period of apparently linear pressure increase with lime is a
- 10,280, and -15,420 STB/D rates, respectively. We estimate the good indicator of pseudosteady-state condi tions. If this
pore volume from Eq. 4.47, being certain to use the average assumption is incorrect, estimated pore volume is too small.
injection rate, q — —9,660 STB/D.
-0.23395(-9,660)(1) v (2.62 X 10_3)(5 x 4.8 Deliverability Testing of Oil Wells
I0-6) Deliverability testing has long been used to predict the
capability of a gas well to deliver against a specific flowing bottom-
= 1.73 X 1011 cu ft =
hole pressure.1,16'19 Fetkovich29 demonstrates that such testing can
30,7 x I09STB, be used for oil wells. It is particularly useful for reservoir systems
for the circular data points. Similarly, the estimated pore volume is operating below the bubble point, when fluid properties and
37.4 x IQ9 STB for the square data points and 25.5 X 10 9STB for relative permeabilities vary with
the diamond-shaped data points. These volumes do not contradict
the known geology of the formation. The results for the diamond-
shaped points (q = — 15,420 STB/D) should be considered the
least reliable since there are only five data points, and they are
scattered. The value of 3 7.4 x 10 9 STB maybe the most reliable,

TIME, t, HR
distance from the well. Oil flow rate (at surface conditions) has
been empirically related to flowing bottom-hole pres sure and
average reservoir pressure by
h'ip'-p.fr, ..................................................{4.48)
where Jf is a form of productivity index and n is an empirically
determined exponent. Fetkovich39 states that field tests indicate
0.5 ^ n ^ 1.0. Eq. 4.48 is similar to the deliverabil ity equation used
in gas well testing.
Two important deliverability tests are the flow-after-flow test
and the modified isochronal test. Fig. 4,16 schemati cally
demonstrates the rate and pressure behavior of a flow- after-flow
test. The well is produced at rate qi until the pressure stabilizes at
pKfj. Then the rate is changed to q-> until the pressure stabilizes at
pM/2, and so on . Normally, four rates are run but any number
greater than three may be used. Flow rate may be either
increased or decreased. The major disadvantage of the flow-after-
flow test is that each rate must remain constant until pressure
stabilizes. The time required may be estimated from Eq. 2.40,
r, ~ 380 ..................................... .......... (4.49)
For systems that arc large or have low permeability, stabili zation
time can be very long.
To avoid problems with long stabilization times. Cullender 18
proposed the isochronal flow test for gas wells. A shortened
version, the modified isochronal flow test,19 was later suggested
and is generally preferred. Fig. 4.17 schematically illustrates flow-
rate and pressure histories for a modified isochronal flow test. The
well is produced at rate qt for time /, and the final flowing pressure,
plrfly is observed. Then the well is shut in for time /, and the final
shut-in pressure, is observed. The procedure is repeated at rales
q.,, q:l. </4, etc. The well is usually produced to a stabilized
pressure at the final rate, so one stabil ized pressure point,
(/vA1**■ *s available.
Fig. 4.18 illustrates the analysis method for a modified
isochronal deliverability test; Iog(/?z — pirf) is plotted vs log q. The
points usually fall on a straight line with slope l/n. The location of
the line depends on the flow-period duration. Thus, in normal
analysis, the points for the four rates define the straight line and
the single stabilized point defines location of the stabilized
deliverability curve. The stabilized deliverability curve may be
entered at set values of (P~ —purf) to estimate the well's
deliverability (flow rate) at a given drawdown. Alternatively, the
data plot (Fig. 4.18) may be used to estimate J„' and n and the
flow rate may be estimated from Eq. 4.48. Fig. 4.18 and Eq. 4.48
are written with average reservoir pressure, the pressure used for
flow- after-flow and normal isochronal flow tests. The data plot
for a modified isochronal flow test uses the shut-in pressure
occurring immediately before the flow rate instead of the
average reservoir pressure.
Fetkovich20 provides data for many flow-after-flow and
isochronal flow tests in several oil wells. Fig. 4.19 demonstrates
that the isochronal and flow-after-flow tests can give the same
results in oil wells producing from a saturated (both oil and free
gas present) reservoir.

W </>
o°- f
W
1 cc
2 =3
OWH
W H £
Of
CD ti

TIME, t, NR
er
li Fig. 4. 17 Pressure-rate history for a modified-isochronal-flow test,
lH
d
ttl
5
_i
o <1.

o
u.

If deliverability test data are not available for a solu tion gas below the bubble point have an inflow performance relationship
drive reservoir, Jt stilt may be possible to predict a well’s given by
deliverability by using the “inflow performance relationship” (1PR)
proposed by Vogel*1 and the modification to the 1PR proposed by
Standing.22 Vogel used computer simulation techniques to
demonstrate that many solution gas drive reservoirs operating
 Fig. 4.18 Schematic example of a modified-isochronal- test data plot.
< 4 T O
«■-£?['c#)']-'
where qQ is the oil flow rate (STB/D) occurring at bottomhole
pressure pwf and /* is a productivity index. Given a stabilized q0
and the corresponding^ andp«■/, it is possible to calculate/* from
Eq. 4.50. Then, to estimate</„ at another stabilized pressure, one
uses Eq. 4.50 with the experimen tally determined/*. Standing 22
indicated that, as the reservoir is depleted, it is necessary to
modify Eq. 4.50 because of changes in relative permeability and
fluid properties. He suggested estimating a future value of the
productivity index from the present value by using

Fig. 4.19 Four-hour modified-isochronal and flow-after-flow deliverability curves. Field data from a saturated reservoir. After Fetkovich.20

where the subscript F refers to some time in the future and the
subscript p refers to data at the present time. Values of kro, fj,0, and
B,, in the future are estimated from material balance relationships.
The procedure for developing a future 1PR is (1) estimate/,,* from
current production data and Eq. 4,50; (2) estimate ip* from Eq.
4.51; and {3) estimate the future IPR(<?„)by using ip* inEq. 4.50.
Standing22 gives an example calculation.
It is emphasized that dcliverability tests are not transient well
tests. They do not yield estimates of skin and formation
permeability; rather, they provide empirical relationships between
flow rate and drawdown for stabilized oil wells. They do include
the non ideal conditions existing in the reservoir, particularly
saturation distributions and variation of fluid properties with
pressure. Deliverability testing thus maybe valuable in helping
predict future production rates as a function of available pressure
differential.
10. Earlougher. R. C.. Jr.: “Estimating Errors When Analyzing
Two-Rate Flow Tests." J. Pet. Tech. (May 1973) 545-547.
11. Odeh, A. S. and Jones. L. G.: “Two-Rate Flow Test. Variable-
Rate Case — Application to Gas-Lift and Pumping Weils,”7.
Pet. Tech. (Jan. 1974)93-99;Trans., AIME,257.
12. Slider, H. C.: "Application of Pscudo-Steady-State Flow to
Pressure-Buildup Analysis/* paper SPE 1403 presented at
the SPE-AIME Regional Meeting. Amarillo. Tex., Oct. 27-28,
1966.
13. Slider. H. C.: “A Simplified Method of Pressure Buildup
Analysis for a Stabilized Well.” J. Pel. Tech. (Sept. 1971)
1155-1160,Trans.. AIME. 251.

4.9 Factors Complicating Multiple-Rate Testing 14. Earlougher, Robeit C., Jr.; “Variable Flow Rate Reservoir
Limit Testing,” J. Pet. Tech. (Dec. 1972) 1423-1429.
Multiple-rate tests exhibit their greatest advantage when
15. Kazemi, Hossein: "Discussion of Variable Flow Rate Reser -
changing wellbore storage makes normal transient test analysis
voir Limit Testing,’’/. Pet. Tech. (Dec. 1972) 1429-1430.
difficult or impossible, That is because such tests can eliminate
16. Rawlins, E. L. and Schellhardt. M. A.: “Back-Pressure Data
changes in the wellbore storage coefficient — even though the
on Natural-Gas Wells and Their Application to Production
effects of wellbore storage still exist. Multiple-rate tests also Practices," Monograph 7. USBM (1936).
reduce the loss of current production. However, such tests arc
17. Theory and Practice of the Testing of Gas Wells, 3rd ed.,
difficult to control since they are flowing tests. Rate fluctuations
Pub. ERCB-75-34, Energy Resources and Conservation
are difficult to measure, especially on a continuous basis. The Board, Calgary, Alta., Canada (1975).
analysis techniques are much more bothersome and difficult than
18. Cullender, M. H.: “The Isochronal Performance Method oF
those for constant-rate tests; they frequently require the use of a Determining the Flow Characteristics of Gas Wells," Trans.,
computer. AIME (1955) 204, 137-142. Also Reprint Series. No. 9 —
To assure the best possible multiple-rate test, the engineer Pressure Analysis Methods. Society of Petroleum
must have an idea of a well’s flow characteristics. The rate change Engineers of AIME, Dallas (1967)203-208.
imposed must be large enough to give significant change in a 19. Katz, Donald L., Cornell. David, Kobayashi, Riki, Poett-
pressure transient behavior of the well. The effect of rate change mann, Fred H.. Vary. John A., Elenbaas. John R,, and
Weinaug, Charles F .'.Handbook of Natural Gas
on pressure response must be estimated from Eq. 4.1, Eq. 4.6, or
Engineering, McGraw-Hill Book Co.. Inc., New York (1959)
Eq. 4.13. Normally, rate is changed by a factor of two or more. Chap. 11.
References 20. Fetkovich, M. J.: “The Isochronal Testing of Oil Wells," paper
1. Matthews. C. S. and Russell, D. G,; Pressure Buildup and SPE4529 presented at the SPE-AIME 48th Annual Fall
Flow Tests in Wells. Monograph Series, Society of Petroleum Meeting, Las Vegas, Nev., Sept. 30-Oct. 3, 1973.
Engineers of AIME, Dallas (1967) 1, Chap. 6. 21. Vogel, J, V.: “Inflow Performance Relationships for Solution-
Gas Drive Wells/’ J. Pet. Tech. (Jan. 1968) 83-92; Trans.,
2. Odeh, A. S. andJones, L.G.: “Pressure Drawdown Analysis.
AIME, 243.
Variable-Rate Case/’ J. Pet. Tech. (Aug, 1965) 960-964;
Trans., AIME, 234, Also Reprint Series. No. 9—Pressure 22. Standing, M. B .: “Concerning the Calculation of Inflow Per-
Analysis Methods, Society of Petroleum Engineers of AIME, formance of Wells Producing Solution Gas Drive Reser-
Dallas (1967) 161-165. voirs,”/. Pet. Tech. (Sept. 1971) 1141-1142.
3. Russell, D. G.: “Determination of Formation Characteristics From
Two-Rate Flow Tests/' J. Pet, Tech. (Dec. 1963) 1347-
l355;7>a/u., AIME, 228. Also Reprint Series, No. 9 —
Pressure Analysis Methods. Society of Petroleum Engineers
of AIME, Dallas (1967} 136-144.
4. Doyle, R. E. andSayegh, E. F.: “Real Gas Transient Analysis
of Three-Rate How Tests," J. Pet. Tech. (Nov. 1970) 1347-
1356.
5. Jacob, C. E. and Lohman, S. W.: “Nonsteady Flow to a Well
of Constant Drawdown in an Extensive Aquifer/’ Trans., AGU
(Aug. 1952) 559-569.
6. Gladfelter, R, E.. Tracy, G. W..and Wilsey, L. W.: “Selecting
Wells Which Will Respond to Production-Stimulation
Treatment," and Prod. Prac.. AP1(1955) I 17-129.
7. Winestock, A. G. and Colpitts, G. P.: “Advances in Estimat ing
Gas Well Deliverability,”/. CcJn. Pet. Tech. (July-Sept.
1965)111-119.
8. Ramey, H. J., Jr.: “Verification of the Gladfelter-Tracy- Wilsey
Concept for Wellbore Storage Dominated Transient
Pressures During Production/’ J. Cdn. Pet. Tech., (April-Junc
1976) 84-85,
9. Pinson, A. E., Jr.: “Conveniences in Analyzing Two-Rate Flow
Tests/’ J. Pet. Tech. (Sept. 1972) 1139-1141,