Chapter 2

Principles of Transient Test Analysis

2.1 Introduction independent of pressure. As a result of those assumptions, and

This chapter summarizes the basic transient flow theory for the since the common boundary conditions are linear, Eq. 2.1 is linear

well testing and analysis techniques presented in this monograph. and readily solved. Therefore, solutions (dimensionless pressures)

An understanding of the following material should clarify the may be added together to form new solutions, as indicated in

techniques presented later, as well as allow the reader to devise Section 2.9. If d>, c,, or k are

additional testing and analysis techniques. Nevertheless, it is strong functions of pressure, or if varying multiple fluid saturations

possible to use the material in Chapters 3 through 13 without a exist, Eq. 2. t must be replaced by a nonlinear form. That equation

thorough reading and understanding of th is chapter. usually must be solved using computer analysis methods

All basic theory needed in the monograph is summarized here. (numerical reservoir simulation) beyond the scope of this

We neither derive the basic flow equations nor show how to solve monograph.

them. Rather, a general equation is used for transient pressure Boundary conditions are an important factor in solutions to Eq.

behavior with dimensionless pressure solutions for the specific 2.1. Most transient-test analysis techniques assume a single well

conditions desired. Some important dimensionless pressure operating at a constant flow rale in an infinite reservoir. That

functions are presented in this chapter and in Appendix C, and boundary condition is useful because every well transient is like

references to others are provided. The dimensionless pressure that of a single well in an infinite reservoir — at early time. At later

approach provides a way to calculate pressure response and to times the effects of other wells, of reservoir boundaries, and of

devise techniques for analyzing transient tests in a variety of aquifers influence well behavior and cause it to deviate from the

systems. “infiniteacting” behavior. Thus, different solutions to Eq. 2.1 are

Sections covering wellbore storage effects and wellbore required for longer time periods. Superposition or other solutions

damage and improvement are included, since those effects have a are needed to include other factors, such as gradually changing

significant influence on transient well response. The reader is rate at the formation face (wellbore storage), hydraulic fractures,

encouraged to study those sections, even if he only scans the rest layered systems, or the presence of multiple fluids or boundaries.

of the chapter. Chapter 11 provides additional information about Many of those solutions are presented in Appendix C and Chapters

the effects of those two quantities. 10 and 1 i; Matthews and Russell* present others. The solution for
a constant- pressure well is given in Chapter 4.
2.2 Basic Fluid-Flow Equation Although Eq. 2,1 appears to be severely restricted by its basic

The differential equation for fluid flow in a porous medium, the assumptions, under certain circumstances it can be applied to both

diffusivity equation, is a combination of the law of conservation of multiple-phase flow and gas flow, as indicated in Sections 2.10 and

matter, an equation of state, and Darcy's law.1-4 When expressed in 2.11.

radial coordinates, the diffusivity equation is*

2.3 Solutions to the Flow Equation — Dimensionless
d'p j. i dp _ 1 ^i dp n ii Quantities
dr' r dr 0.0002637 k dt.......................... Comprehensive treatments of transient well testing normally
Matthews and Russel]1 present a derivation of Eq. 2.1 and point out use a general approach for providing solutions to the diffusivity
that it assumes horizontal flow, negligible gravity effects, a equation, Eq. 2.1. Such an approach provides a convenient way of
homogeneous and isotropic porous medium, a single fluid of small summarizing the increasing number of solutions being developed
and constant compressibility, and applicability of Darcy’s law, and to more accurately depict well or reservoir pressure behavior over
that /x, ct, k, and 4> are a broad range of time, boundary, and geometry conditions. The
general solutions
♦Symbols and units are defined in the Nomenclature. Normally,
only deviations from that list are discussed in the text.
 PRINCIPLES OF TRANSIENT TEST ANALYSIS 5

rely on the concepts of dimensionless pressure and dimensionless

time, explained later in this section. Some solutions are identical
for the appropriate boundary conditions; and s is the skin
toothers in certain time ranges, but are significantly different in effect, a dimensionless pressure drop assumed to occur at the
others. Thus, throughout the monograph, guidelines indicate where
complex solutions are needed and where the simpler solutions wellbore face as a result of wellbore damage or im-
provement.Skin effect, s, only appears in Eq. 2.2 when r„ = 1. (See
normally give adequate results.
Section 2.5.)
An unfortunate consequence of the generalized dimensionless- I a transient flow ,plt is always a function of dimensionless lime.
solution approach i$ that the dimensionless parameters do not
t) _ 0.0002637 kr ■ ■ (2.3a)
provide the engineer with the physical feel available when normal
<j>pc,rlr2 ’
dimensional parameters are used. For example, a real time of 24
hours may correspond to a dimensionless time range from about when based on wellbore
300 for a light gas reservoir to more than JO 7 for a highly permeable
0.00026 radius,
oil reservoir. The pressure corresponding to a 24-hour time in those
37 kt or
two situations might vary by hundreds of pounds per square inch.
Fortunately, after one works long enough with dimensionless
4ffur,A (2.3b)
when
variables, one does begin to get a feel for them. Nevertheless, it is
always good practice to calculate physical quantities from based on total drainage area. Dimensionless pressure also varies

dimensionless ones; that is easily done because physical quantities with location in the reservoir, as indicated in Eq. 2.2 by the
are directly proportional to dimensionless quantities. dimensionless radial distance from the operating well.
The dimensionless-solution approach can be illustrated by
starting with the familiar steady-state radial flow equation:

□ = o 007OS 2 ~ P,r)
Bp. 1 n (rjr„ )
The point location also may be expressed in Cartesian coordinates.
This equation may be solved for the pressure difference,
Dimensionless pressure is also affected by system geometry, other
system wells, the wellbore storage coefficient of the producing well,
p„ - p„. =
141.2 anisotropic reservoir characteristics, fractures, radial
'±-'‘ 1 n
kft discontinuities, and other physical features.
Dimensionless pressure./?/,, is a solution to Eq. 2.1 for specific
Changing to dimensionless form, the radial flow equation becomes
boundary conditions and reservoir geometry. Practically speaking,

pr - pu, = 141.2 &Lpu, where dimensionless pressure is just a number given by an equation, a
Pi, = ln(rr/rM.). table, ora graph. Some expressions forp„ are given in Sections 2.4,
2.7, and 2.8, and Appendix C. The dimensionless-pressure
Thus, the physical pressure drop in the steady-state radial- flow
approach is used throughout this monograph because of its
situation is equal to a dimensionless pressure drop, which in this
simplicity and general applicability in well-test development and
case is simply ln(rr/r,r), times a scaling factor. The scaling factor
analysis. The approach, which is easy to apply, results in simple,
depends on flow rate and reservoir properties only. The same
general equations that apply to any set of reservoir properties. It is
concept applies to transient flow and to more complex situations —
easily adapted to mathematical manipulation and superposition
only the dimensionless pressure is different. It is this generality that
(Section 2.9), so more complex systems can be considered. For
makes the dimensionless-solution approach useful.
simplicity, the following conventions apply throughout this
The advantages of the dimensionless form occur, as indicated
monograph:
previously, when situations get more complex. In general terms, the
1. Although dimensionless pressure is generally a function of
pressure at any point in a single-well reservoir being produced at
time, location, system geometry, and other variables, we commonly
constant rate, q, is described with the generalized solution of Eq,
2.1: write /?0fo?. ...), PDQD)-, or just p^. Dimensionless pressure, p,it is a
number that may be obtained from an equation, figure or table; it
p, -p(t.r)= 141.2 ^ kh scales linearly to real pressure.
[Pit (Iif 0). geometry, ...) + s], ...............................(2,2) 2. The symbol tf) always refers to dimensionless time
calculated from Eq. 2.3a using the wellbore radius. It is dearly
where />, is the initial, uniform pressure existing in the reservoir
indicated when dimensionless time is based on some other
before production or injection; q is the constant surface flow rate;
dimension. Dimensionless time is just real time multiplied by a
k, h, and p are constant reservoir properties; pi■) is the
scale factor that depends on reservoir properties.
dimensionless-pressure solution to Eq. 2*1
3. Eq. 2.2 includes the van Everdingen-Hurst5,6 skin factor.
That factor appears only when calculating Ap for a producing or
injecting well. In general, s is not shown in equations unless it is
specifically used. The reader should recognize that adding s is
necessary under the appropriate circumstances.
The following example illustrates the use of Eq. 2.2 to estimate
flowing well pressure in a closed system.
Example 2.1

pt ~ 3,265 psi k0 = 90
md /*»= 13.2 cp =
1.02 RB/STB h = 47 ft
 6 ADVANCES IN WELL TEST ANALYSIS

Estimating Well Pressure gime; we prefer the term exponential-integral solutionM the Theis51 solution) to the flow
Estimate the pressure at a well '“infinite-acting", since all wells (also called the line- source or equation is
forx <0.0025................................................................................(2.7b)

(-^-J
located in the center of a act as if they were alone in an
infinite system at short flow At the operating welt r„ = l, so
closed-square reservoir after it
times. The infinite-acting period
Pn(lD.rD)= ~ L Ei ’ ................................................
loir,,2 = t„. Since t„ > 100 after
has produced 135 STB/D of dry
is characterized by a straight only a few minutes for most
oil for 15 days. Other data are* systems, there is practi cally
line on the semilog plot, Fig.
(2.5a)
4> = 0.17 2.1b. The portion of the curves no difference between the two
c( = 2.00 X I0~s psi-1 r„. = 0.50fi labeled C in Fig. 2.1 is the - 1 [ ln(f0/rDs) + forms of Eq. 2,5, as illustrated

pseudosteady-state flow' by the following example.

A— 40 acres = 1,742,400 sq ft s 0.80907]...............................................................
— 0. regime that occurs in all closed
systems. During pseudosteady- Eq. 2.5b may be used when
Curve A of Fig, CJ3 is used for
state flow, pressure changes tptrj >100........................................................
pu since it applies to a closed- Example 2.2 Estimating
linearly with time, as shown in (2.6)
square system with a well at Pressure vs Time History
Fig. 2.1a. The B portion of the but the difference between Eq.
the center. of a Well
curves is the transition period 2.5a and Eq. 2.5b is only about
Using Eq. 2.3b,
between infinite-acting and 2 percent when 1,Jrtt- > 5. Use the exponential-
(0,0002637)(90)(15 x 24) pseudosteady-state flow. Thus,1 for practical purposes, integral solution and the dara
(0.17)( 13.2)(2.00 x 10-*)( In Fig. 2.1, flow is transient the log approximation to the of Ex ample 2.1 to estimate the
1,742,400)“ at all times. Some systems exponential integral is pressure vs time relationship
= 0.109. exhibit true steady-state satisfactory when the for a well in an infinite-acting
From Curve A of Fig. C. 13, behavior with p„ constant. exponential integral is system.
0.109) = 6.95. Those systems are most satisfactory. Nevertheless, the Wc calculateplcf at 1 minute
Rearranging Eq. 2.2 and commonly observed in more accurate limit of Eq. 2.6 and at 10 hours to illustrate ihc
substituting values, laboratory core flooding and is used in this monograph. procedure; final results are
(r. r„) = 3,265 - permeability measurement shown in Fig. 2.4. From Eq.
Plrf
The exponential
1141.2^13^1^13^ (6_95) ' experiments; they also may 2.3a,
(47) exist in fluid injection projects integr
with balanced production and al is
= 2,843 psi.
injection and in reservoirs with defin
a strong natural w'ater drive.
ed by
Fig. 2.1 schematically 2.4 Dimensionless x
illustrates three transient flow Pressure During the
regimes fora closed drainage
system. Dimensionless
Infinite-Acting Flow Ei( —x) = — I — dit.............................................................
Period
pressure is shown
function of both t/M and log
as a
Fig. 2.2 is a schematic
J
.V
"

representation of a single well Values may be taken from

(Ji,A)- The portion marked A is
producing at constant rate q in
the early transient or infinite- tables10 or may be
an infinite, horizontal, thin
acting flow re- approximated from
reservoir containing a single-
♦In the examples in this phase, slightly compressible Ei( —jr) ^ ln(jc) -I- 0.5772
monograph, data values are not
always stated to their full fluid. When the assumptions of
number of significant digits. In Eq, 2,1 are satisfied, Eq. 2.2.
such cases, values are
assumed to have three with p,t from Fig. 2.3, describes
Fig, 2,4 Calculated pressure response for a well in an infinite-
significant digits with the pressure behavior at any
significant zeroes omitted. acting system. Example 2.2.
Computations are usually done point in the system. Fig. 2.3
using the intermediate values showsp„ is a function of t„ and
shown. When intermediate
values are not shown, all com- of r,)t the dimensionless radial
puted digits have been used distance from the well, for the
and the final result has been infinite-acting system. {Fig. C. 1
rounded off.
is a full-scale grid- ded version
of Fig. 2.3.) Whenr^ ^ 20 and
rw/rw2 5* 0.5, or whentjr,,2 ≥ 25.
ther„ — 20 and the
"exponential-integral solution"
lines on Figs. 2.3 and C.l are
essentially the same, sopJt
depends only on t,Jrv- under
those conditions.8 The
 PRINCIPLES OF TRANSIENT TEST ANALYSIS 7

Fig. 2.3 Dimensionless pressure function at various dimensionless distances from a well located in an infinite system. After Mueller and
Witherspoon.*1
t _ (0,0002637) = 3.265 -(60,67X2.18)
(90) r = 3.133 pst.
° (0.17X13,2)<2,00 X 10- At 10 hours, I,, — (2,115)( 10) =
21.150 and the log approx-
*)(0.50)2
imation, Eq, 2,5b. can be used:
= 2,115/.
Pi,* ' [ln(2U50/l)
+0.80907]
At I minute, fD = (2,115)(l/60) =
35.25. The exponential- integral
= 5.384. so
solution, Eq. 2.5a, applies
becauset„/rD2 = 35.25/1 > 25. pirAis »0 hours) = 3.265 -
However, since tIflrt/ = 35.25 <
(60.67)(5.384)
100, the log approximation. Eq,
2.5b, should not be used. = 2.938
Using Eq. 2.5a and
psi.
evaluatingplf from Fig. C. 1 ?

we get puib = 35.25) - 2.18.

Then, rearranging Eq. 2.2»

Pn/0 = I minute)
= 3 265 - (i41.2)(i35)
(l.02)(13.2) (2 |g)
' (90X47)
 8 ADVANCES JN WELL TEST ANALYSIS
The response from 0.01 hour (0.6 minute) to 100 hours is shown in the well face. The thin-skin approximation results in a pressure
Fig. 2.4. The dashed line is the log approximation for /j)frj> 4 < 100; gradient reversal for wellbore improvement (s < 0). shown in Fig.
after that time the two lines coincide. Note how well the two 2.5B, Although this situation is physically unrealistic, the skin-
solutions (Eqs. 2.5a and 2 5b) agree, even for tDfrD* = (2,115)(0.0l)/i factor concept is valuable as a measure of wellbore improvement,
= 21. A more physically realistic pressure profile for the negative skin
situation is also shown in Fig. 2.5B.
If the skin is viewed as a zone of finite thickness with
permeability ks, as shown in Fig. 2.6, then"
As mentioned previously, all wells are infinite-acting for some
time after a change in rate. For drawdown, the duration of the
infinite-acting period may be estimated from ..........................................( 2 , 0 )
Either s, k„ or r, may be estimated from Eq. 2.10 if the other two
=
oJmsm .......................................................................... • < 2 ' 8 a l parameters are known.
It is also possible to define an apparent wellbore radius for use
where rM at the end of the infinite-acting period is given in the “Use in Eqs. 2.3 and 2.4 so the correct pressure drop at the well results
Infinite System Solution With Less Than 1 % Error for tDA <“ when s — 0 is used in Eq. 2.2:12
column of Table C. I. For a well in the center of a closed circular
reservoir, ( / / w = 0.1 and r*. ~ .............................................................................(2.11)
For positives, ru.a < rw.; for negatives', rKU >rK..
~ M^t±............................................................(2.8b) Eqs. 2.2 and 2.9 show that the skin factor simply increases or
k
decreases the pressure change at a well proportional to the flow
Equations for buildup are given in Section 5.3. rate of that well. When dimensionless pressure functions include
2.5 Wellbore Damage and Improvement Effects the skin factor (for example,

There are several ways to quantify damage or improvement in

operating (producing or injecting) wells. A favored method
represents the wellbore condition by a steady-state pressure drop
at the weltface in addition to the normal transient pressure drop in
the reservoir. The additional pressure drop, called the “skin
effect,1’ occurs in an infinitesimally thin “skin zone.”3'* In the flow
equation, Eq. 2.2, the

Fig. 2.5A Pressure distribution around a well with a positive skin factor.

Fig. 2.6 Skin zone of finite thickness.

Figs. C.6 and C.7), it should normalpD's are used.

not be included explicitly in The flow efficiency (also
Eq. 2.2. Dimensionless called the condition ratio) indi-
pressure functions that cates the approximate fraction
include the skin factor are of a well’s undamaged pro-
particularly useful when the ducing capacity. It is defined

degree of damage (or improvement) is expressed in terms of a skin zone extends for some as the ratio of the well’s actual

“skin factor,” s, which is positive for damage and negative for distance rather than being productivity index to its

improvement. It can vary from about —5 for a hydraulically concentrated within a few feet productivity index if there
of the wellbore. The skin- were no skin.1 For closed
factor concept is used only for systems, the flow efficiency is
Fig. 2.5B Pressure distribution around a well with a negative
skin factor. pressures at the well. When
■factual P ~ Ptt-f ~ Apg ^2 12)
the skin affects some rs Jideal P ~ Ptcf

around the well, the pressure Although the drainage-area

profile in that Tegion will be average pressure, p, should
fractured well to +=c fora well that is too badly damaged to be used in Eq. 2.12, it is
modified (for example, see Fig.
produce. Eq. 2.2 indicates the pressure drop at a damaged (or frequently permissible to use
2.5B). In such situations,
improved) well differs from that at an undamaged well by the the extrapolated buildup
special p0 functions that
additive amount pressure,1 p*. Flow efficiency
includes must be used to
\p. = i .............................................(2,9) determine pressures near the depends on the flowing
kh well. When using Eq. 2.2 to pressure, pir/, and thus it
Fie. 2.5A illustrates the idealized pressure profile for a damaged estimate pressure at distances depends on how long the well
well (s > 0). Since the damage-zone thickness is considered to be greater than ru. or r„.0, the s has been operating unless the
infinitesimal, the entire pressure drop caused by the skin occurs at term is omitted, and well is at pseudo steady-state

(2.13)
| _ factual f ideal
The following
indicators.
(2.14)
 PRINCIPLES OF TRANSIENT TEST ANALYSIS 9

conditions. (At pseudosteady and using transient test would be that Skin factors estimated
state, p — pirf is constant.) For given by Fig. 2.7 if there were from transient tests on
Eq.
wells operating at true steady no true physical damage (or hydrauli-
state, p should be replaced by 2.14. improvement) at the well. If
prt the pressure that the area the there is physical damage, the
will reach after extended shut- calculated skin factor is higher
damag
in. than indicated by Fig. 2.7.
The damage ratio and e cally fractured wells are Section 11.3 provides addi-
damage factor are also relative factor generally negative. Fig. 2.8 tional details.
indicators of wellbore compares the dimensionless
is 1 -
condition. The inverse of the pressure for an ideal, 2.6 Wellbore Storage
flow efficiency is the damage 0.82 = undamaged, unstimulated well Wellbore storage, also
ratio:

f Ideal _ P Pwf factual p

~ p,cf ~ &p 6
(2.16)
By subtracting the flow
efficiency from 1, we obtain
the damage factor:

= ...........................
P ~ Ptcf

example illustrates the

wellbore damage

Example 2.3 Wellbore

Damage Indicators A pressure
buildup-test analysis for a well
with 17 = 83 STB/D, B = 1.12
RB/STB, p = 3.15 cp, h - 12 ft,
rw. = 0.265 ft, and p - pwf = 265
psi gave k = 155 md and s =
2.2. Find the pressure drop
across the skin, the flow
efficiency, the damage ratio,
the damage factor, and the
apparent wellbore radius.
Using Eq. 2.9,

An, =
HiLgX
5.6I46
gMLiat
Fig. 2.7 Pseudoskin factor for partially penetrating wells. After Brons and Marting. ,s
C In
m (2.2) = 0.18. with that for a hydraulically called afterflow, fahr,'
49 psi.
(IJ5)(J2) The apparent wellbore radius
fractured well with
fracture length of 31.63 r(c. At
a half- afterproduction, aftertnjection,
and wellbore unloading or
*
The flow efficiency is is estimated from Eq. 2.11: small tn, the difference loading, has long been
estimated from Eq. 2.12: between the two
rH.„ = 0.265 e-*z = 0.03 ft. recognized as affecting short-
265 — 49_ — o 82 = 82 dimensionless pressure time transient pressure
percent. Damage at this well is
reducing production to about curves, which isj (Eq. 2.2), behavior.2-14 More recently,
265 varies; at larger tt> that several authors15'26 have con-
82 percent of the value that
Using Eq. 2.13, the damage difference is constant. This sidered wellbore storage in
could be expected without
ratio is detail. It is easy to see that
indicates that reasonable skin
damage or stimulation.
values can be estimated from liquid is stored in the wellbore
transient tests for many when the liquid level rises.
hydraulically fractured wells. That situation occurs when a
Wells completed with only However, when large fracture pumping well without a packer
a part of the formation thick- jobs are known to have been is shut in; indeed, bottom-hole
ness open lo the wellbore can performed, the fracture should pressure is often deduced by
appear to be damaged. Partial be accounted for by analyzing measuring liquid level. When
penetration (wells not drilled well tests using the type-curve wellbore storage is significant,
completely through the pro- matching' method (Section 3.3) it must be considered in

ductive interval) and partial transient test design and

with Figs. C.3, C.4, C.5, C. 17,
analysis. If it is not considered,
completion (entire productive C. 18, or C. 19. An important
the result may be an analysis
interval not perforated) are feature of all the log-log plots
of the wrong portion of the
examples.13 (See Section 11.4 of fractured-well p» data is the
transient test data, the
for more information and slope of 'A at small t0, This
deduction of nonexistent
references.) Fig. 2.7 shows slope also will be observed on
reservoir conditions (faults,
theoretical “pseudoskin” a log-log plot of transient
boundaries, etc.), or an
factors for partially pressure difference data from
analysis of meaningless data.
penetrating wells. The skin fractured wells, unless it is
Fortunately, the effectsof
factor estimated from a obscured by wellbore storage.
wellbore storage usually can
 10 ADVANCES IN WELL TEST ANALYSIS
be accounted for in (2.17) where Vw is the total
test
wellbore volume in barrels
analysis — or can be avoided andc is the compressibility of
by careful test design. the fluid in the wellbore at
wellbore conditions.
The wellbore storage Throughout this monograph,
constant (coefficient, factor) is the wellbore storage
coefficient, C, has units of
defined16 by barrels per psi; some authors
prefer cubic feet per psi. The
c = AL.....................................................................
compressibility in Eq. 2.17 is
Ap for the fluid in the wellbore; it
is not c, for the reservoir.
Since the wellbore fluid
compressibility is pressure

Fig. 2.8 Comparison of dimensionless pressures for an ideal well and for a well with a single vertical fracture. Infinite-acting system.
dependent (Appendix D), the
wellbore storage coefficient
may vary with pressure.
Fortunately, such variation in
where C = wellbore storage wellbore storage coefficient is
generally important only in
constant wells containing gas or in
(coefficient, wells that change to a falling or
rising liquid level during the
factor),
test. Those conditions are

Fig. 2.9 Effect of wellbore storage on sand-face flow rate, C3 > Cj > C,.
bbl/psi, considered in Section I (.2.
AY * change in Some dimensionless
pressure functions (Appendix
volume of fluid in the
C) for systems with wellbore
wellbore, at wellbore
storage use a dimensionless
conditions, bbl, and
wellbore storage coefficient.
Ap — change in
(2.18)
bottom-hole
pressure, psi. Note that the total
Applying Eq. 2.15 to a wellbore compressibility for the
with a changing liquid level,16 reservoir system* rf, is used in
this definition.
Wellbore storage causes
(**)' the sand-face flow rate
change
to

where F,, is the wellbore

volume per unit length in
barrels per foot. Eq. 2.16 is
valid For both rising and falling
liquid levels. When the
wellbore is completely full of a
singlephase fluid, Eq. 2.15
becomes16
C = Virc.....................................•
 where At and Ap are values read from a point on the log-log unit-slope straight line. C calculated from Eq. 2.20
more slowly than the surface flow rate. Fig. 2.9 schematically shows the ratio of sand-face to surface rate when
the surface rate is changed from 0 to q at time 0. When C — 0, qsfkf = I at all times. ForC > 0, the flow-rate ratio should agree fairly well with C calculated from Eq. 2.16 or Eq. 2.17. If it does not, a reason should be
changes gradually from 0 to 1. The larger C, the longer the transition, as. indicated in Fig. 2.9. The sand-face
flow rate may be calculated from sought. We have observed that
wells producing at high gas-liquid ratios, highly stimulated wells, and
wells used for viscous fluid injection commonly indicate wellbore storage coefficients from the log-log data
24 C dp^ B dl plot that are much higher than those predicted from Eqs. 2.16and 2.17.
The log-log data plot is a valuable aid for recognizing wellbore storage effects in transient tests when early-
— <?[l Cit -j—PidJu-'Cf)»...)]...................................(2.19) time pressure data are available. Thus, throughout this monograph, it is recommended that this plot be made a
dt/j
<? Eq.2.2 relates flowing well pressure to time for a constant flow rate, «r/. Since Eq. 2.19
part of transient test analysis. It often helps the engineer avoid serious analysis mistakes by delineating the
period of wellbore storage dominance as a unit-slope straight line. As wellbore storage effects become less
*/■ = indicates that ^ varies with / and p, it appears that Eq. 2.2 may not be useable. Fortunately, severe, the formation begins to influence the bottom-hole pressure more and more, and the data points on the
</+ the problem is avoided by using a dimensionless pressure that accounts for wellbore log-log plot fall below the unit-slope straight line and finally approach the slowly curving line for zero wellbore
storage and, thus, for the change in flow rate. SuchPt^tp, C/>, ...) are shown in Fig. 2.10, a storage. Such behavior is illustrated by Fig. 2.10 for varying degrees of wellbore storage. Sometimes pressure
simplified version of Figs. C.6 and C.7. The effect of w ellbore storage onpb is clear in those figures. data between the unit-slope line and the zero wellbore-storage line can be analyzed for forma tion properties,
Fig. 2.10 has a characteristic that is diagnostic of wellbore storage effects: the slope of the plt vs tlt graph but the analysis may be tedious. The Gladfelter-Tracy-Wilsey 15 and the type-curve matching techniques 1*24
on log-log paper is 1.0 during wellbore storage domination. Since p„ is proportional to Ap and tt, is proportional (Sections 3.3 and 8.3) apply in this region. Once the final portion of the log-log plot is reached (C f> = 0 line),
to time (see Eqs. 2.2 and 2.3), Fig. 2.10 indicates a way to estimate when wellbore storage is dominant during a wellbore storage is no longer important and standard semilog data-plotting analysis techniques apply. As a
transient test. On log-log paper, plot the pressure change during the test. pir - pa\At = 0), (as a positive number) rule of thumb, that time usually occurs about 1 to 1 Vi cycles in time after the log-log data plot starts deviating
against tesr time. A/, and observe where that plot has a slope of one cycle in pressure change per cycle in time. significantly from the unit slope. The time may be estimated from
(Note that the nomenclature has been generalized here: p^ is the bottom-hole pressure during the test, be it
tp > (60 + 3.5s)C|>, ............................................................(2.21a)
flowing or static; p^iAt = 0) is the pressure at the instant before the start of the test, be it static or flowing; and
A/ is running test time, with the test starting at A/ = 0.) Well test data falling on the unit slope of the log-log plot or approximately,

reveal nothing about formation properties, since essentially all production is from the wellbore during that t> (200,000 4- I2,000*)C .........................(2.21b)
time. The location of the log-tog unit slope can be used to estimate the apparent wellbore storage coefficient (kh/p,) ’•
from1* for drawdown and injection tests. For pressure buildup and fa 11 off tests, Chen and Brigham25 state that a
reasonably accurate analysis is possible when
C- ...................................................................(2.20)
> 50 Cp e014*..................................................................(2.22a)
or approximately when , ^ 170,000 C e0-14*
(khfp.)
Note that skin factor influences pressure buildup (falloff) much more than drawdown (injection).
Fig. C.5 for a horizontally fracLured well without wellbore storage shows unit-slope straight lines for small
h0. Thus, factors other than wellbore storage can cause the unit-slope straight line on the log-log plot.
Fortunately, horizontal fractures are thought to occur rarely.

Example 2.4 Computing Wellbore Storage Coefficients From Well Data Water is injected into a sand at
2,120 ft through 4.75-in.,

(2.22b)
16.00-Ibm casing. Estimate the wellbore storage coefficient for (l) a wellhead injection pressure of 400 psi, and
(2) a wellhead vacuum. Use <fy = 0.15, h = 30 ft* and rw — 3.5 in. to calculate die dimensionless wellbore storage
coefficient.

1. When the wellhead pressure is greater than zero, Eq. 2.17 is used to estimate C. Water compressibility
is estimated to be 3.25 x 10“6 psi-1 from data in Appendix D. For 4.75-in., 16-lb*, casing, Vu = 0.0161 bbl/ft, so Vw
= (0.0161)(2,I20) = 34.1 bbl. Using Eq. 2.17,

C = (34.])(3.25 x 10-fi) = LI 1 X 10“4bbl/psi.

From Eq. 2.18,

c _ (5.6146)(1.11 X 10-«)____________ = g0
D
2ir(0.15)(3.25 x 10-6)(30K3.5/12)' '
2. In this second case there is a changing liquid level in the well, so Eq. 2.16 is used;

In this case, there is a factor of about 340 difference between compressive and changing-liquid-ievel
storage. Note, we assumed ct ~ ck., which is not always a valid assumption.
illustrates dimensionless pressure behavior when the wellbore storage coefficient changes. When the wellbore
storage coefficient increases (fromC, toC 2 in Fig. 2.12), pD (or Ap) flattens, begins to increase again, and finally
Example 2.5 Computing Wellbore S torage Coefficient approaches the response curve for the larger storage coeffi cient. When the wellbore storage coefficient
From Well Test Data decreases,

Use the log-log data plot shown in Fig. 2.11 to estimate a wellbore storage coefficient. That well has 2.5-in.
tubing in 8Vfe-in., 35.5-lbm casing.

LOG <t)

The wellbore storage coefficient can change during transient testing. For example, consider a falloff test in
a water injection well with a high wellhead pressure during injection. When the well is shut in, surface pressure
is high initially but could decrease to atmospheric and go on vacuum if the static formation pressure is below
hydrostatic. The liquid level must start falling as soon as the wellhead pressure drops below atmospheric. As a
result, the wellbore storage coefficient increases from one for fluid compression (Eq. 2.17) to one for a falling
liquid level (Eq. 2.16); the second storage coefficient easily could be a hundred to a thousand times the first.
The reverse situation can occur as well, with a high, rising-liquid-level storage at the beginning of injection
changing to fluid-compression storage as the wellhead pressure begins to increase. Fig. 12.12 schematically

Fig. 2.11 Log-log dataplotter Example 2.5. Pressure buildup lest Fig. 2.12 Theoretical pressure response for both increasing and
in a sandstone reservoir with 6 = L0 RB/STB; q = 66 STB/D; decreasing wellbore storage; C2 > Ct. Adaptation of data
h = 20 ft; and depth = 980 ft. from Eartougher. Kersch, and Ramey.28
there is a rapid pD (or Ap) increase as the pressure response
£ 600 1 i
--- -T " —l
approaches the low wellbore storage curve. More discussion about
I I I —TT 0
this is given in Ref. 26 and in Section 11.2 I I I
c
Stegemeier and Matthews27 showed that gas-liquid (phase) r

redistribution in the wellbore causes anomalous pressure-buildup w 400 o

o

:
curves of the form shown in Fig. 2.13. Fig. 2.14 indicates that phase
3
redistribution is similar to wellbore storage* although it is probably
more complex than anything else presented in this section. It is o
important to understand that the behavior illustrated in Fig. 2.13 is o _

wellbore* not formation* dominated. Fitter, Rice, and Thomas 28 demon- . o .

strated this by testing a well once with surface shut in and a second _ o -

time with bottom-hole shut in. . o -

In summary, wellbore storage effects always should be considered qi a Qi ___1
l 1
0.1 10
___
in transient test design and analysis and in computing the expected " 0.01 I
pressure response of wells. In some cases, transient tests must be 20
designed to minimize or alleviate wellbore storage, or no useful SHUT-IN TIME, At HR. ,
Fig. 2.13 Pressure buildup behavior showing the effect of fluid
information will be obtained. Wellbore storage effects can be
segregation in the wellbore. After Matthews and Russell.1
recognized on log-log data plots if sufficient short-time pressure data
are available. No information about the formation can be determined
from transient test data falling on the unit slope of such a data plot.

2*7 Dimens ionless Press ure During the Pseudosteady-State

Flow Period
Fig. 2.1 indicates that in closed systems a transition period follows
the infinite-acting transient response. That is followed by the
pseudosteady-state flow period, a transient flow regime when the
pressure change with time, dp/dt, is constant at all points in the
reservoir. (That is equivalent to the right-hand side of Eq. 2.1 being
constant.) This flow period occasionally has been mistakenly called
steady state, although at true steady state pressure is constant with
time everywhere in the reservoir (Section 2.8).
Fig. 2.15 schematically shows a single producing well in the.center
of a closed-square drainage area. Fig. 2.16, a plot of vs tD at Points A,
B, C, and D of Fig. 2.15, illustrates two properties common to all closed
systems. First, at small dimensionless times, p0 at the well is given by
Eq. 2.5 if vS/r,,. > 50; the well behaves as if it were alone in an infinite
system. Second, at large dimensionless times, p0 at any point in the
system varies linearly with tDA, This is the pseudosteady-state flow
period, which can occur only in bounded systems. During
pseudosteady state, dimensionless pressure is given by29

....................................... (2.23)

CA, the shape factor, is a geometric factor characteristic of the system

shape and the well location, Values are given by Brons and Miller,12
Dietz,30 and others.7,31 BothQ and the final term in Eq. 2.23 are given in
Table C. 1. Eq. 2.23 may be used for any closed system with known
shape factor, if 31.62, the CA value for a well in the center of a circular
system, is used in Eq. 2.23, the last two terms become the familiar
\n(rjru) — 0.75-
 Eq. 2.23 applies any time after pseudosteady-state flow begins; that time may be estimated from From Table C.l.C, = 30.8828. Using Eq. 2.23,

0. 0002637 k
([OA )M (2.24)
where is given in the “Exact for tDA >” column of
Table C. 1. Both CA and (i„A depend on reservoir shape and weli location.
Dimensionless pressure data at the well and at several Then, from Eq. 2.2,

Fig, 2.16 Dimensionless pressure at various points in a closed square caused by a producing well in the center. A, B, C. and D identified in Ftg. 2.15; V/i7#v — 2,000. After Earlougher, Ramey, Miller and Mueller, 31

,VX/.r„) = 3.265 - (l4'-2)(13SX,l.02)(l3.2j (931) (90){47) 1

= 2,700 psi.

2.8 Steady-State Flow

When the pressure at every point-in a system does not vary with time (that is. when the right-hand side of
Eq. 2.1 is zero), flow is said to be steady state. Linear and radial steady-state flow, whose dimensionless
pressure distributions are illustrated in Fig. 2.17, usually occur only in laboratory situations. The
dimensionless-pressure functions arc

steady state, linear

(P»W ^2-rr‘A...............................................................................(2.25)
A
steady state, radial
(/>,,)„,= ..............................................................................................(2-2&0
Fig. 2,17 Dimensionless pressure distribution in linear and radial steady-state flow.
other points in closed rectangular systems are given in Ref. 7. When Eq. 2.26a is used in Eq. 2.2, we obtain after rearrangement

^ _ 0.0Q7Q82 kh (p,. - pu ) ^ ................................(2.26b)

Example (f.(? Estimating Well Pressure During
Bp ln(rjric)
Pseudosteady-State Flow
the familiar radial form of Darcy's law.32
Using ihe data of Example 2.1, estimate flowing well pressure after 60 days. From Eq. 2.3b,
In reservoirs, steady-state flow can occur only when the reservoir is completely recharged by a strong
,_ (0.0002637 )(90)(60 x 24) aquifer or when injection and production are balanced. Muskat33 relates flow rate to tnterwell pressure drop for

(0,17)<I3.2)(2.00 X I0-S)( 1,742,400) several flooding patterns. His equations are easily converted to the dimensionless pressure approach used in
this monograph. Perhaps the most useful is the dimensionless pressure expression for a five-spot flooding
= 0.437.
pattern at steady state with unit mobility ratio and with ra. the same in all wells:
Tabled indicates pseudosteady state exists after i,u =0.1 forawell in the center of a square, so Eq. 2.23 applies.
 >-»11.........................................(2.27) \rtc f
Here, A is the five-spot pattern area, not the area per well.
It is useful to recognize that Higgins-Leighton34 geometric factors are dimensionless pressures for cells As used here, the superposition principle states that adding solutions to a linear differential equation
within streamtubes operating at steady state. By appropriate addition of available Higgins-Leighton geometric results in a new solution to that differential equation, but for different boundary conditions. Eq. 2,2 is a solution
factors, such as those in Ref. 35, one can calculate dimensionless pressures for many irregular steady-state to Eq. 2.1 for a single well producing at constant rate q, Superposition can be applied to include more than one
systems. well, to change rates, and to impose physical boundaries. Superposition is easily applied to infinite systems;
but for bounded systems it must be used with more care — not because the principle is different, but
becausep0 solutions frequently do not give the necessary information for correct superposition.
(2.28) where J is the Higgins-Leighton geometric factor for cell j in streamchannel i of the pattern. The sums To illustrate the principle of superposition in space, consider the three-well infinite system in Fig. 2.18. At t
are taken over all cells and all channels. If the pattern is symmetric, the
(M* right-hand side of Eq. 2.28 must be multiplied by the number of
= 0, Well t starts producing at rate*/,, and Well 2 starts produc ing at rater/;,. We w ish to estimate the pressure

= symmetry units. For the confined five-spot pattern flood of Ref. 35, we at the shut-in observation point. Welt 3. To do this, we add the pressure change at Well 3 caused by Well I to
the pressure change at Well 3 caused by Well 2:
calculate p„ — 10.498; here, VAJr„. = 500. Eq. 2.27 gives the same result
to three decimal places.
Ap, = A+ Ap«|.2.....................................................................(2.29)

To use Eq. 2.29 wc must substitute Eq. 2,2 for Ap. Then, extending to an arbitrary number of wells,/ = 1 , 2 , ....
2.9 The Principle of Superposition n.
So far, only systems with a single well operating at a constant rate from time zero onward have been
considered. Since real reservoir systems usually have several wells operating at varying rates, a more general
Ap(l. r) = >m).................(2.30)
approach is needed to study the problems associated with transient well testing. Fortunately, because Eq. 2.1
is linear, multiple- rate, multiple-well problems can be considered by applying the principle of superposition. j=1
The mathematical basis for this technique is explained by van Everdingen and Hurst,* Collins,3* and others,M*ai
where r,tj is the dimensionless distance from Well /' to the point of interest. Note that Eqs. 2.29 and 2,30 add
pressure changes (or dimensionless pressures), not pressures. If the point of interest is an operating well, the
skin factor must be added to the dimensionless pressure for that well only.
Ftg. 2.19 graphically illustrates the use of Eqs. 2.29 and

I
0 __________I____________________________________
0
tl
TIME—►
Fig. 2*20 Variable rate schedule for superposition explanation.
Fig. 2.19 Calculation of pressure change at observation well of
Fig. 2.18 Multiple-well infinite system for superposition Fig. 2,18. qi = 100STB/D; ri = 100ft;*?z= 150 STB/D; f2 = 316 ft; k
explanation. = 76 md; p = 1.0 cp; = 0.20; cf — 10 x 10"* psi_1,B = 1.08 RB/STB;
h = 20 ft; and 5 = 0.

0 20 40 60 80 IOO
t, HOURS
Fig, 2.21 Calculation of pressure change at producing well with the rate history of Fig. 2.20. = IOQ STB/D; 17» = 50 STB/D;
k = 76 md; fj. = 1.0 cp; $ = 0.20; c,= 10 X I0_fi psi_1;£ = 1.08 RB/STB; A = 20 ft; and 5 = 0,
2.30 for the system of Fig. 2.18 and the exponential integral Pn (Eq. 2.5). The lowest curve in Fig. 2.19 is the
pressure change at Well 3 caused by Well 2. The upper dashed curve is the pressure change at Welt 3 caused
by Well 1. Using Eq. 2,29, we add the two pressure changes to get the solid curve, the total pressure change
observed at Well 3.
To illustrate an application of the principle of superposition to varying flow rates, consider a single-well
system with the production-rate schedule shown in Fig. 2,20. The production rate ts^, from / = 0 to / — tt and qt
thereafter. To perform the superposition calculation the single well may be visualized as two wells located at of the summation are needed. Fig. 2.21 illustrates the calculation. The upper dashed curve (including the First
the same point, with one well producing at rate q% from t = 0 to r and the second (imaginary) well producing at portion of the solid curve) is the pressure change caused by r a t e . The lower dashed curve is the pressure
rate (q£ - g,)> starting at and continuing for a time period (r — tThe net rate after time /, would be qx + ( q3 - qt) change caused by rate qt — qt after / = 40 hours; that Ap is negative because (qt ~ 4i) < 0. The sum of the two

= q.t. As in the previous illustration, A/Ts are added for these conditions. The general form of the equation for dashed curves, the solid curve, is the pressure response for the two-rate schedule.
To combine varying rates and multiple wells, First apply Eq. 2.31 for each well in the system to estimate Ap
N rates, with changes at ij.j = 1. 2. ... N. is ”
caused by that well at the point desired. Then add the Ap for all wells, as in Eq. 2.30, to get the total Ap caused
In Eq. 2.31, [r—ri]^ is the dimensionless time calculated at time ( / “/ , ). For Fig. 2.20, N — 2 and only two terms by all wells and all rates. The double summation process is conceptually simple, but can be tedious in
application, as illustrated by the following example.
 Example 2.7 Principle of Superposition 2; that calculation of A p may require use of Eq. 2.31 to account for varying rates.
For the conditions shown in Fig. 2.22, From Eq, 2.3a,
t _ 0,0002637 kt _ (0,0002637)(76)/
estimate the pressure at Well I after 7 hours
and at Well 2 after 11 hours. Assume that ° 4>px:,rw2 (0.2)(1)(10 X 10‘W
the system behaves as an infinite one at = 10,000/.
these short times.
Start by computing the coefFicicnts in
the Ap and tD equations, Eqs. 2.2 and 2.3.
Then, at agiven time, estimate Ap at the
desired well caused by both Well l and Well

WELL t WELL 2
wm----------------IOO*---------------
t, HOURS tt HOURS
Fig. 2.22 Example superposition calculation for two wells, each
produced at two rates. Well l: s = 5, rw — 1 ft. Well 2; s = 1.7, rv - I
ft. fc = 76 md; fi= 1,0 cp; 4> = 0.20; ct = 10 x IQ-6 psi-1;fi = 1.08
RB/STB; h = 20 ft; andpj = 2,200 psi.
PRINCIPLES OF TRANSIENT TEST ANALYSIS 18
For Well 2 , r „ = I:

r„(M hours) = (10.000)01) = 110,000.

Since h, > 100. we use Eq. 2.5b: p„(Wdl 2.

r„ = 110.000. r„ = 1)
= I[ln(r„> + 0.80907]

= 6 .2 1 .
r„(i 1-8 hours) = (10.000)(3) = 30.000.

Recali that ss must be added to p0 to get Ap at Well j (unless po Pi,Ci, = 30.000, r„ = I)

is a function of J, as are some that are given in Appendix C). The
= 1 [ ln(30,000) + 0.80907]
appropriate r also must be used to calculate rD, depending on the pn
function we use.
= 5.56.
Ait = 7 hours, there is a A p contribution at Well 1 from a single
rate at Well I and a single rate at Well 2; so the over-all pressure Estimating Ap at Well 2,
change would be (Eq. 2.30) Ap(Well 2 , 1 1 hours) =
(0.1)(l00)(1.61) + (0,1){50— I00)(0.522)
Ap (7 hours, rlt = 1) + (0.1)(25)(6.2l + 1,7)
+ (0. t)(100—25)(5.56 -f 1.7)
1) + * ] = 87.7 psi.
pw.(Well 2, 11 hours) = 2,200 - 87.7 = 2,112.3 psi.
+ ^-pM„ r„ = 100/1)-
kh

For the contribution of Well !,//> = (10,000)(7) = 70,000. Since tt) >
Additional applications of the principle of superposition and the
100 Eq. 2.5b is used:
Path = 70,000, rt) = 1) method of images are shown in Appendix B.

= i [ln(70.000) + 0.80907} 2.10 Application of Flow Equations to Gas Systems

Although this monograph is concerned only with liquid
— 5.98. systems, much of the material presented can be applied to dry gas
systems if modified slightly.
For the contribution of Well 2, at a distance of 100 ft, Gas viscosity and density vary significantly with pressure, so
the assumptions of Eq. 2.1 are not satisfied for gas systems and
tn g (70,000) = 7 (100/1)* *
the equation does not apply directly to gas flow in porous media.
Sincer„ >20, we can use the line-source solution. Eq. 2.5a, but we
Thai difficulty is avoided by defining a “real gas potential” (also
should not use Eq. 2.5b unless tjr,? > 100. From Fig. C.l or Fig. C.2
commonly referred to as the real gas pseudopressure or just
for t,Jrf,2 = 7 and rIt = 100,
PuCn = 7, r„ = [00) = 1.40. pseudopressure):37,38

p
Calculating at Well I, r
m(p) = 2 P dp...................................................(2.32)
Ap(Weli 1, 7 hours) = (0.1)(I00)(5.98 + 5)
+ (0.1X25)(i:40)
I
P1P)HP)
Sh
= 113.3 psi.
where pt is an arbitrary base pressure. When the real gas potential
is used, Eq. 2.1 essentially retains its form but wilh m(p) replacing p.
The pressure at Well I at 7 hours is fV (7 hours, r„ = 1) =p, - Ap =
That equation can be solved and an analog to Eq. 2.2 can be written
2,200 - 113.3 — 2,086.7 psi.
with ntptto) in place of pD(tD), For radial gas flow it has been
An = 11 hours, we wish to estimatep at Well 2. We must consider
shown37‘39,4° that when
two rates at each well: f
O < fo)p»i
Ap(l 1 hours, r„ = 1)
= (0.0(100)p„(Well 1,1 = II hours, r„ = 100) HOCD) =PDCO)......................................................................<2-33>
where potto) is the liquid flow dimensionless pressure.
+ <0.1)(50—I00)p„(Well I, t = [ i t — 10] hours, rn = 100)
+ (0.l)(25)[p0(Well 2, 1 - II hours, Using Eq. 2.33, the gas analog 2.34, the iermD|qr | accounts
ra= I) + fj of Eq. 2.2, and substituting the for non-Darcy flow around the
+ (0.1)(l00-25){pD(Well 2, r = [l 1 —S] hours, re = 1) + a}. appropriate gas properties, the wellbore. Otherwise, the form
For Well l, use Fig. C.2: flow equation for a real gas is is like the liquid flow equation.
ID(1! hoursw = To use Eq. 2.34 it is necessary
00.000/(10 = n (100)2 - m(p,) - 50,300 £SL. to construct a high-resolution
fl
pw(Welt l, tlt - 11, rw = 100) = 1.61. •sc graph of m(p) vs p from the
kh viscosity and z factor for the
;„( 11 -10 hours)//',/ - (IQ.OOOKJl. = |
gas. If p and z are not known,
[PDM + J + £>|<?|]...............................................
(100)5 (2.34) information presented by Zana

p„(WelI I, f„ = 1, r„ = 100) = 0.522. where q is in Mscf/D. In Eq. and Thomas41 may be used to

and to replace the total system

compressibility by

s
L
kh

(2.39a)
 PRINCIPLES OF TRANSIENT TEST ANALYSIS 19
estimate m(p) vs p. the real gas potential, m(p), uniform saturation distribution. stabilization time and radius of
As a result of the must be used. Eqs. 2.34,2.35, The same modifications are drainage are commonly used
characteristics of the real gas or 2.36 may be used with liquid made when analyzing transient in petroleum engineering and
potential, Eq. 2.34 can be dimensionless pressure for well test data. For example, in transient testing. These
simplified for certain pressure most gas systems. mobility estimating equations quantities are frequently used
ranges. Fig. 2.23 shows pz as a take the form without appropriate
Applica
function of pressure for a understanding of their actual
tion of „ ± 162.6g0B„
typical gas. AT low pressures meaning and limitations. It is
Flow Wo "l/l ................................................
beyond the scope of this
pz is essentially constant,

-±f
Equatio
while at high pressures it is monograph to investigate the
ns to
essentially directly 162.6 (1,000) problems associated with
Multiple radius of drainage and
proportional to pressure. When -Phase stabilization time. However,
this behavior is used in Eq. Flow [q, - 0,001 because of their wide use, the
2,32, Eq. 2.34 can be simplified
Eqs. 2.1 and 2.2 and the (4»S» + quantities are defined and
lo
dimensionless pressure (k\ _ + 162.6 qtrBv equations are given for them in
information in this monograph this section.
“ Mh .........................
are derived for single-phase Stabilization time has been
(2.35) flow. However, they may be The ± sign in Eq. 2.39 indicates defined in many ways by
used for certain multiple-phase application to any of the various authors.46,47 Most
at high pressures, while at low various test analysis definitions correspond to the
flow situations with some
techniques. beginning of the
pressures it modifications.42'45 The basic
approach is to replace the pseudosteady-state flow

becomes mobility terms in Eqs, 2.1

2,12 Radius of Drainage period. Using that as the
and Stabilization Time definition of stabilization time,
through 2.3 by the total flowing
-50,3000^,) The concepts of we can estimate the
mobility,
stabilization time for any shape definitions are presented by
\ — K, +■ Xg + AMM ............. given in Table C. 1 from Eq.
[pi>(lv) + S+D\q\]
van Poollen,46 Kazemi*60 and
(2.37 a) or . .. (2.37b)
..................................................................... 2*24, For a well in the center of Gibson and Campbell/’1 In most
Cl +. <■»+«/■
most symmetrical shapes, definitions, the radius of
40 ~$ u t
As a rule of thumb, Eq. 2.36
(t)
i,m\
drawdown stabilization time is
yi 11 n

drainage defines a circular

generally applies when p < estimated from
te system with a pseudosteady-
2,000 psi, while Eq. 2.35

r -l i /aMl
- state pressure distribution from
generally applies forp > 3,000 . I, -380
(2.40a) the well to the * ‘drainage
psi; for 2,000 < p < 3,000 use
+s
"1 B
[
Op ) B„ k radius/* As time increases,
Eq. 2.34. We suggest that the -
more of the reservoir is
»
p z vs p plot be made for the
| 1
L
B
where t, is in hours. If we
particular gas flowing before - (2.38a)
choosing between the With these modifications,
single-phase liquid influenced by the ■ well and the
equations. If neither situation assume the system is radial*
dimensionless pressures may radius of drainage increases, as
prevails at the pressure level
given by
observed or expected, then be used to describe multiple- i, - 1.200
phase systems containing
r„ = 0.029 J — , ..................................................................(2,41)
immiscible fluids with fairly (2.40b)
(2,38b)
k

For stabilization time in days where rrt is the radius of

the coefficient in Eq. 2.40b is 50 drainage in feet and / is in
and the equation takes the form hours. If ( h expressed in days,
given by van Poollen.46 It is the constant 0.029 rounds to
important 10 recognize that 0,14, and Eq;2.41 corresponds
stabilization time can be to equations given in Refs.
considerably longer than 46,50, and 51, Eventually, must
indicated by Eq. 2,40 when the Stop increasing —- either when
shape is not a symmetrical one reservoir boundaries or
with the well in the center* or if drainage regions of adjacent
two or more noncoinmunicating wells are encountered, SO Eq.
layers intersect the wellbore,42- 2.41 can only apply until .
4
"
Radius of drainage is also
defined in several ways. Many

PRESSURE, p, PSI A
Fig. 2.23 Isothermal variation of pz with pressure.
 20 ADVANCES IN WELL TEST ANALYSIS
Example 2.8 Radius of discuss various kinds of to decline, but at a more rapid that affect both pressure
Drainage decline and pressure
Estimate the radius of drainage created during a 72-hour• test on a well in a reservoir with kip = 172 md/cp and 4K-{ = 0.232 x 10' psr * Using s l
distribution. Each well will
Eq. 2.41, establish a drainage area that
supplies all fluid removed from
that well — if there is no fluid
reservoir simulators have rate than if the boundary had injection into the system.
= 2,100 ft. not been encountered. If, on the When boundaries are
appeared in the petroleum
This estimate is valid only if no literature. Three of the classics other hand, the transient encountered (either no-flow or
are by Aronofsky and Jenkins, 53 pressure response reaches a constant-pressure), the
boundaries are within about
replenishable outcrop that pressure gradient—not the
2,100 ft of the test well, and if Bruce, Peaceman, Rachford,
maintains constant pressure at pressure levels tends to
no other operating wells are and Rice54, and West, Garvin,
55 some point, pressures nearer stabilize after a sufficiently long
within about 4,200 ft. and Sheldon. Many facets of
the withdrawal well will decline production time, the
reservoir simulation were
more slowly than if a no-flow stabilization time. For the
summarized by van Poollen,
boundary had been closed-boundary case, the
Bixel, and Jargon56 in a series
encountered. Rate changes or pressure behavior reaches
In systems completely of articles appearing in the Oil
additional wells will cause pseudosteady state with a
recharged by an aquifer or and Gas Journal. The SPE-
additional pressure transients constant gradient and an over-
when production and injection AIME Reprint Series booklet57
all pressure decline every*
are balanced, the concepts of on numerical simulation
where that is linear with time. AIME. Dallas (1967) 25-
stabilization time and radius of contains many useful papers. 43.
For reservoirs with constant-
drainage are meaningless. Chapter 12 of this 5. van Everdingen, A. F.:
pressure boundaries, a steady
However, Ramey, Kumar, and monograph presents some “The Skin Effect and Its
Gulati52 define a readjustment state may be approached. In Influence on
information about the
time, the time required for a that case, both pressure the Productive Capacity
application of computers to of a Well,” Trans.. AIME
short-lived transient to die out, gradient and absolute pressure
transient well testing. (1953) 198, 171-176.
for such systems. For a single values become constant with
Also Reprint Series, No. 9 —
well in the center of a constant- time. The pseudosteady-state
Pressure Analysis Methods,
2.14 Summary — A and steady-state solutions to
pressure square, which is Society of Petroleum
Physical Viewpoint
equivalent to a balanced five- Eq. 2.1 have a simple form and Engineers of AIME,
After presenting the Dallas (1967)45-50. ’
spot water injection pattern represent the simplest
dimensionless-parameter 6. Hurst. William;
with unit mobility ratio, the approach to future
approach to solution of "Establishment of the
readjustment time is52 performance predictions, when
transient flow problems and Skin Effect and Its
they are applicable. Impediment to Fluid
= 9« explaining some of the factors
....................................................... Flow lntoa Well Bore,"
k that influence those solutions,
Pet- Eng. (Oct.
In this equation, A would be it seems worthwhile to References J 953) B-6 through B-16.
approximately one-half the five- summarize the situation from a
1. Matthews. C. S. and 7. Earlougher. R. C., Jr.,
spot pattern area. physical viewpoint.
Russell, D. G.: Pressure and Ramey, H. J., Jr.:
Fluid withdrawal from a well Buildup and Flow Tests in "Interference Analysis in
2.13 Numerical Solution of penetrating a pressurized Wells, Monograph Series, Bounded Systems,” J.
petroleum reservoir containing Society of Petroleum Cdn. Pet. Tech. (Oct.-Dec.
the Diffusivity Equation
a compressible fluid results in a Engineers of A1ME, 1973)33-45.
It is possible to obtain
pressure disturbance. Although
Dallas (1967) I, Chap. 2. 8. Mueller, Thomas D. and
analytical solutions to Eq. 2.1 2. vanEverdingen, A. F,
we might expect that Witherspoon, Paul A.:
only for the simplest systems. and Hurst, W.; “The “Pressure Interference
disturbance to move with the Application of the
Most dimensionless-pressure Effects Within
functions are from numerical
speed of sound, it is quickly Laplace Transformation Reservoirs and
attenuated, so for any given to Flow Problems in Aquifers," J. Pet. Tech.
solutions of Eq. 2.1 or its
length of production time there Reservoirs," Trans., (April 1965)471-
analogs for gas and multiple- A1ME (1949) 186, 305-
is some distance, the radius of 474;Trans.. AIME,234.
phase flow'. Computer solution 324.
drainage, beyond, which no
3. Hubbert. M. King: “The 9. Theis, Charles V.: “The
is the only practical method for
appreciable pressure change Theory of Ground-Water Relation Between the
obtaining dimensionless Lowering of the
can be observed. As fluid with- Motion J. ofGeol. (Nov.-
pressures for extremely
Dee. 1940) XLVIII, 785- Piezometric Surface and
drawal continues, the
heterogeneous systems, the Rate and Duration of
disturbance moves farther into 944.
Discharge of a Well
layered systems, systems with 4. Homer, D. R.: “Pressure
the reservoir, with pressure Using Ground-Water
two or three phases flowing, Build-Up in Wells," Proc., Storage,” Trans., AGU
continuing to decline at all
systems with water or gas Third World Pet, Cong., (1935) 519-524.
points that have started to The Hague (1951) Sec.
coning, or systems with 10. Abramowitz, Milton and
experience pressure decline. II, 503-523. Also Reprint
significant gravity effects, for When a closed boundary is Series, No, 9 —Pressure Stegun, Irene A.
example. During the past Analysis Methods, Society of Handbook of Mathematical
encountered, the pressure
Petroleum Engineers of Functions With Formulas,
several years, many papers that within the boundary continues Graphs and Mathematical
 PRINCIPLES OF TRANSIENT TEST ANALYSIS 21
Tables,National Bureau Instantaneous Charge K. M.,and Ramey, H.
of Standards Applied of Water." Water J., Jr.: “Wellbore
Mathematics Series-55 Resources Res. (1967)3, Effects tn Injection Well
(June 1964) 227-253. No. 1,263-269. Testing,” J. Pet, Tech.
M. Hawkins. Murray F., Jr.: 19. Ramey, H. J., (Nov. 1973) 1244-1250.
“A Note on the Skin Jr.: “Short-Time Well 27. Stegemeier, G.
Effect,” Trans., AIME Test Data Interpretation L. and Matthews, C. S.:
(1956) 207,356-357. tn the Presence of Skin “A Study of Anomalous
12. Brans, F. and Miller, W. Effect and Wellbore Pressure Build-Up
C.: "A Simple Method for Storage,”/. Pet. Tech. Behavior,” Trans., AIME
Correcting Spot (Jan. 1970) 97-104; (1958) 213, 44-50. Also
Pressure Readings,” J. Trans., AIME, 249. ” Reprint Series, No. 9 —
Pet. Tech. (Aug. 1961) 20. Agarwal, Ram Pressure Analysis Methods,
803-805;Trans., AIME, G., Al-Hussainy, Rafi, Society of Petroleum
222. and Ramey, H. J., Jr.: Engineers of AIME,
13. Brons, F. and Marring, "An Investigation of Dallas (1967) 75-81.
V. E.: “The Effect of Wellbore Storage and 28. Pitzer, Sidney
Restricted Fluid Entry on Skin Effect in Unsteady C.t Rice, John D., and
Well Productivity,” J. Pet. Liquid Flow; L Thomas, Clifford E,: “A
Tech. (Feb. 1961) 172- Analytical Treatment,” Comparison of
174; Trans., AIME, 222. Soc. Per. Eng.J. (Sept. Theoretical Pressure
Also Reprint Series, No. 9 — 1970)279-290; Trans.. Build-Up Curves With
Pressure Analysis Methods, AIME. 249. Field Curves Obtained
Society of Petroleum 21. Wattenbarger, From Bottom-Hole
Engineers ofAIME, Robert A. and Ramey, Shut-In Tests,” Trans.,
Dallas (1967) 101-103. H. J., Jr.: “An Inves- AIME (1959) 216, 416-
14. Chatas, Angelos T.: “A tigation of Wellbore 419. Also Reprint Series.
No. 9—Pressure Analysis
Practical Treatment of Storage and Skin Effect
Non-Steady State Flow in Unsteady Liquid Methods, Society of
Problems in Reservoir Flaw: II. Finite Petroleum Engineers of
Difference Treatment, AIME, Dallas (1967)
Systems,” Pet. Eng., Part
"Soc. Pet. Eng. J, (Sept.
83-86.
1 (May 1953) B-42
through B-50; Part 2 1970) 291-297; Trans., 29. Ramey, H. J.,
(June 1953) B-38 AIME, 249. Jr.t and Cobb, William
through B-50; Part 3 22. McKinley, R. M.: “A General Buildup
(Aug: 1953) B-44 M.r “Wellbore Theory for a Well in a
through B-56. Transmissibility From Closed Drainage
15. Gladfelter, R. E,, Tracy, Afterflow-Dominated Area,”/. Pet. Tech. (Dec.,
G. W., and Wilsey, L. E.: Pressure Buildup 1971) 1493-1505.
“Selecting Wells Which Data,” J. Pet. Tech. (July 30. Dietz, D. N.:
Will Respond to 1971) 863-872;Trans., "Determination of
Production-Stimulation AIME, 251. Average Reservoir
Treatment," Drill, and Prod. Pressure From Build-
Prac., API (1955) 117-
23. Barbe, J. A. and Up Surveys,”/. Pet. Tech.
129. Boyd, B, L.: “Short- (Aug. 1965)955-959;
16. Ramey, H. J., Term Buildup Test- Trans., AIME,234.
Jr.: “Non-Darcy Flow ing,” /, Pet, Tech. (July 31. Earlougher,
and Wellbore Storage 1971) 800-804. Robert C., Jr., Ramey,
Effects in Pressure 24. Earlougher, H. J., Jr., Miller, F. G.,
Build-Up and Robert C., Jr., and and Mueller, T. D.:
Drawdown of Gas Kersch, Keith M.: “Pressure Distributions
Wells,” J. Pet. Tech. (Feb. “Analysis of Short-Time in Rectangular
1965) 223-233; Trans., Transient Test Data by Reservoirs," /. Pet. Tech.
AIME, 234. Also Reprint Type-Curve Match- (Feb. 1968) 199-208;
Series, No. 9 — Pressure ing,”/. Per. Tech. (July Trans., AIME, 243.
Analysis Methods, Society 32. Amyx, James
1974)793-800;Trans..
of Petroleum Engineers AIME, 257. W., Bass, Daniel M.,
of ATME, Dallas (1967) Jr,, and Whiting, Robert
233-243. 25. Chen, Hsiu-Kuo
L.: Petroleum Reservoir
and Brigham, W. E.:
17. Papadopulos, “Pressure Buildup fora
Engineering: Physical
lstavros S. and Cooper, Properties, McGraw-Hill
Well With Storage and
Hilton H., Jr.: “Draw- Book Co., Inc. New
Skin in a Closed
down in a Well of Large York (I960) 78-79.
Diameter,” Water
Square,” paper SPE
Resources Res. (1967) 3,
4890 presented at the
No. 1,241-244. SPE-AIME 44th Annual
California Regional
18. Cooper, Hilton Meeting, San
H,, Jr., Bredehoeft, Francisco, April 4-5,
John D., and Papa- 1974.
dopulos, lstavros S.:
“Response of a Finite- 26. Earlougher.
Diameter Well to an RobertC., Jr., Kersch,