Professional Documents
Culture Documents
Practical Well Test Analysis of A Hydraulically Fractured Low Permeability Gas Reservoir: A Case History
Practical Well Test Analysis of A Hydraulically Fractured Low Permeability Gas Reservoir: A Case History
Hazim N. Dmour
Department of Petroleum and Natural Gas Eng ineering, College of Engineering,
King Saud University, P.G. Box 800, Riyadh 11421, Saudi Arabia
Keywords: Hydraulic fracturing, Low permeability gas reservoir, Gas well performance, Non-Darcy effect, Finite
conductivity fracture , Modified isochronal test.
Abstract. The primary objective of hydraulic fracturing is to create a propped fracture with sufficient conductivity
and length to amplifY or at least optimize well performance of low permeability tight gas reservoir. The oil
industry has suggested that hydraulically fractured tight gas wells performance is hindered significantly by non-
Darcy flow effect.
This work will present an investigation of non-Darcy flow effect to hydraulically fractured gas well s
performance and provide the development, validation, and application of actual well test analysis for wells with a
finite conductivity vertical fracture.
Also, this work presents the results obtained in the study of actual post frac modified isochronal test data of
gas wells intersected by a finite conductivity vertical fracture in a tight low permeability gas reservoir. In addition,
the estimation of reservoir properties and fracture properties were carried out to construct a simple analytical
model, which used for rate prediction.
The effect of non-Darcy flow in fractures is clearly seen in the tests data and will lead to limiting production
especially on higher chokes (after one inch). Therefore, non-Darcy effects should be considered in design of
hydraulic fracture treatments, otherwise the design might be far from optimal
47
© 200 (142+). Production and hosting by Elsevier B.V. on behalf of King Saud University. This is an open access article under
the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
48 Hazim N. Dmour: Practical Well Test Analysis of ...
practically apply them to the design and analysis of no-flow outer boundary. At early times, this model
fractured well tests. utilizes the concept of (Lee et aI., 1986) of tri-linear
(Cringarten et aI., 1972) reviewed the theory on flow to represent a finite conductivity fracture (Fig.
transient pressure analysis for both hydraulically and 2). Three linear-flow zones that dominate the pressure
naturally fractured reservoirs. Later, (De Swaan, behavior are:
1976) presented analytical unsteady sate solution for • Fracture flow in the X-direction.
a well producing at constant flow rate in naturally • Formation flow in Y-direction.
fractured reservoirs; he introduced new diffusivity • Formation flow in X-direction.
definitions useful for reservoir characterization. Fracture diffusivity has been assumed constant at
(Crawford et al., 1976) have presented some of 1 E6, as suggested by (Cinco-Ley et al., 1979). The
the best field examples of pressure transient tests on tri-linear fracture flow results merge into the solution
both hydraulically and naturally fractured reservoirs. for infinite-acting radial flow in the middle times.
In fractured wells, the study (Partilmo, 2003) reported Thus, the tri-linear flow solution is truncated as soon
that tests on hydraulic fractured wells often matched as the flow becomes pseudo-radial. Occasionally, the
the early fractured type curves but yield apparent merging of these two solutions is not smooth, and the
fracture length of 10ft when the design lengths where derivative exhibits spikes. These are localized
more than 1000 ft. An early study by (Holditch et al., aberrations and can be ignored as they do not affect
1976) attributed this to high velocity flow in a the rest of the results. Ultimately, at late times, the
fracture. But, the work by (Ramey, 1965) indicated model uses the solution for pseudo-steady state for a
that finite-fracture permeability-width was a more no-flow outer boundary or continues to use the
likely cause of the problem. They used a finite- solution for infinite-acting radial flow. In this model,
element solution. (Prats, 1961) provided the key to the changing wellbore storage or dual porosity effects
this problem in classic study on steady-state flow. are not considered. The flow equations in pressure are
(Cinco-Ley et aI., 1986) presented a truly classic written in the Laplace domain, solved, inverted to real
semi-analytical, real domain solution for a well with a time domain using the Stefhest algorithm (n = 18).
finite conductivity vertical fracture in an infinite
acting reservoir. This semi analytical solution is oundaries
tedious, but accurate. • No Flow (Black
All works presented above address very well the • Infinite (Grey) .
effects of fracture design changes on the well
performance. An analytical model obtained from
different sources has been used in this work without
considering the impact of the equations selection on
the final results. In this paper, a systematic evaluation
of Modified Isochronal Test of hydraulically
fractured two gas wells was performed to examine the
effect and magnitude of non-Darcy flow on pseudo-
Fig. 1. Finite conductivity fracture model.
steady state productivity index (of stimulated gas
well). In addition, this paper shows an application of
this technique to the hydraulically fractured low
permeability tight gas reservoir and demonstrates
how build-up and drawdown data dominated totally
by bilinear flow, which can be rigorously evaluated
with this concept, and by computer model.
~Lt~ well+
behavior of hydraulically fractured gas wells. They
f
observed in their study that in a build-up test, non-
k, X Darcy flow can continue for a substantial time after
I Li-l ~ll the well is shut in, and the apparent fracture
conductivity can continue to increase throughout
much of the test. The non-Darcy flow continues as
pressure gradients dissipate during the build-up test
because the flow velocity in the narrow fracture
remains quite high.
Fig. 3. Schematic diagram of a well completed with a vertical It is normally expected to see non-Darcy flow in
fracture.
(Bilinear fracture flow along the fracture and perpendicular to gas well tests. This effect will result in creating what
the fracture). we call rate dependent skin. This rate-dependent skin,
which is caused by turbulent flow, can be added to any
At early time, after the possible effects of model. The skin for a flow period of rate is given by:
wellbore storage have subsided, the response is
bilinear at right angles to the fracture and along the (1)
length of the fracture as shown in Fig. 3. On a log-log
scale, this is characterized by a quarter unit slope on
both the pressure and derivative curves. The quarter where So is the mechanical skin and dS/dQ defmes
unit slope is essentially a very early time feature, and the rate dependence, sometimes called non-Darcy
is very often masked by the effects of wellbore coefficient.
storage. After the bilinear flow, the response will There is also a similar phenomenon, which was
become linear flow in the reservoir, characterized by seen in these tests, which is non-Darcy flow in the
a half-unit slope. When the fracture half-length and fracture itself. The effect of non-Darcy flow in a
formation permeability are known independently, the hydraulic fracture can be significant causing a large
fracture conductivity kJwf can be determined from the pressure drop in the fracture. This will lead to
bilinear flow regime. changing the behavior of the fracture depending on
50 Hazim N. Dmour: Practical Well Test Analysis of ...
(8)
Several relations were proposed for calculating the
apparent conductivity, but the most commonly used
relation is: Therefore, the flow period of drawdown and build-up
is the ~\If / q vs. ~ and \If vs. ~tc + ~ta - ~ . The
(C ) - CID (5) superposition time is also required when multirate
ID APP-l+0.55QD analyses are used to approximate situations in which
the rate is slowly varying. The formulati on of
This relation is only valid for CID < 10, QD < 2 superposition time depends on the flow regime being
(minimal non-Darcy effects, or CID > 10 (for low analyzed. For example, in our case bilinear flow .
permeability formations), QD= 1 to 10. Therefore: Therefore, the following formulae are the generalized
form of superposition time for bilinear flow regime:
J King Saud Univ., Vol. 20, Eng. Sci. (1), Riyadh (2008/ 1429H.) 51
(11) (17)
This data is then plotted on a log-log plot. Taking the For infinite conductivity fractures, when kr is large
logarithm of the left hand side and the right hand side (FcD>20), then the skin factor is:
of the above equations illustrates that a plot of log
Der. vs. log t results in a slope of Y4 (the Y4 fraction in
front of the log(t) term). (18)
The coordinates of any point on the quarter slope
may be used to calculate fracture conductivity.
Pseudosteady state analysis
(12) Pseudosteady state flow is a flow regime that
occurs in bounded (closed) reservoirs, after the
pressure transient has reached all the boundaries of
Therefore, the root fracture conductivity of the reservoir. This includes not only the case of
drawdown (Iog(~\jf 1q) vs.log(t a )) : physically bounded reservoirs, but also the case of a
well surrounded by other producing wells. In these
situations, reservoirs exhibit tank-like behavior. The
(13) purpose of analyzing pseudo steady state flow data is
to determine the reservoir pore volume, V p, and
original hydrocarbons in place, OOlP or OOlP. This
analysis is valid only when the well is flowing. There
And, for build-up (Iog(~\jf) vs.log(~teB)):
is no corresponding pseudo steady state analysis for
build-up or falloff tests. Pseudosteady state flow
cannot be observed in data obtained from build-up or
falloff tests. The constant rate solution for analyzing
(14) pseudosteady state data is:
52 Hazim N. Dmour: Practical Well Test Analysis of ...
-
\jI-\jI wf = 1, 422 qg T Z Ilg [ In - (re J 3
--+S'
] (23)
kh rw 4
(19)
(24)
2348 qT ta
Der = ----'-----"- (26)
A h<p Ilgi cti
where '" a 1S adjusted pressure evaluated at
\jJ (average static drainage area pressure), and If/a,wf where
adjusted pressure evaluated at If/wj. in terms of
pressure itself, an adequate approximation to the (27)
pseudosteady state flow equation is:
The reservoir pore volume is used to calculate the post frac wellhead and sandface pressure
original hydrocarbons in place. From the drawdown measurements (modified isochronal test). The
plot log (fi\lf/q) vs. log(tpssa) using Eq. (25). charactenstIcs of the reservoir as well as the
description of the completion are summarized in
Interpretation methodology of post fracture test Table 2.
~T) .
Due to non-Darcy flow effects, it was not possible Table 1. MIT wellhead and sandface ~ressure measurements
-...;:io"ds Duration, Wellhead Sandface Flow rate,
in some cases to match the drawdown and the build-up hr I!ressure, ~si p:ressure, I!si MMscf/d
response simultaneously. Therefore, two models were IcitiaiShut in 3366 3366
made; an emphasis was put on matching the fmal build 1" flow 12 1347 1750.21 8.207
and drawdown in each test when two models were 32/64"
1" shut in 12 2223 2654.85
used. The quality of the match depended on: 12 844 1222.41 8.189
2nd flow
- The quality of the rate measurements and, 40/46"
- The span of the fracture conductivity change due 2nd shut in 12 1960 2337.3
to non-Darcy flow" 3'd flow 12 574 969.9 8.146
48/64"
3'd shut in 12 1815 2163.38
Case Studies 4th flow 12 335 626.54 8.661
64/64"
The following two field examples are presented to Extend flow 48 259 580.6 6.696
serve and to illustrate the value of looking at this 64/64"
diagnostic model.
-
Final shut in 110 2564
Field example 1
The analysis of actual data obtained from the gas analyzed (Fig. 5). This test followed a hydraulic
well that stimulated by hydraulic fracturing is fracture of the well. The fracture was clearly
performed using the model of finite conductivity successful, as exhibited by the Y2 slope on the
vertical fracture solution. Table 1 summarized the derivative plot, and the straight line on the fracture
linear flow plot Figs. 6-11 (superposition
Table 2. Reservoir and I!as properties (analvsis results for Field Example 1)
Finite Conductivity Fracture Gas Well Model
(post-frac test)
Model Parameters Fluid Properties "
-
600 3
2 2 XI- 100.00 ft
.00 2
400 ./
600 ~
200 1\
.00
\/
I
,, 11
\.,
I1
If
~
V
/ 1
400 I r"""T
=
o P data
20 ..0 60 w ~ ~
Cum.Time.
~ ~ ~ ~ ~ 260
20 . 100 120 140
Time..hr
160 110 lOO 220
_ Pmodel
240 260
Fig. 5. Post-frac modified isohronal test (bottom hole) - flow Fig. 8. Finite conductivity fracture (total test overview).
rate vs. pressure.
: Kb - lS.300mdft
• Sxt - ·5.116
::~.:soo..!
square root time). A fracture half-length of I Xf - 100.00 ft
600
x,("I"(k»- 30.3n ft.md"
:~o~ft /
~ ."
/ Fig. 9. Fracture model match (finite conductivity fracture type
/. curve).
./ . , ICh - U .lOOrnift
.uo s,r, -...s. lIti
1
./ _
b - 110,00 It.
k-O.O&lmd
Xf - 100.00 It
o " data
~ - slope
12 11 10 a 7 6
Supupositicm Lineal- Pseudo
~-------------4----~
~~~~
~------ o .~
600
PO-3404.79 psi
K- O.1SI md
Kh- 27.136 mdft
/ _ ~modcl ::
S'- -4.463 /
1/ Fig. 10. Finite conductivity fracture model (radial flow
analysis).
h
200
77 \I:SO
11.00
Kh - IS.300 mdft
!~ ~80~~1: X( - 100.00 ft
;:~~~1-'-_0,_,0'_" md
_'-++++-H+__-,----+_++-*"t-I----j--+_-~I-'----i'-
100 ~ l...- - :::::~
.... ---/ ~ ••~~~----+-H-++++--+!-~--~~H-+++-L----i:::::~
~ ::[f\I'III~I-ti-I~I'~~-'-~JEI~
0
10' 10' 10' 10' 10
Superpositioo Radial Pseudo Time
1:~
Fig. 7. Final build-up - radial flow extrapolation. 1
~.~:~n~·06 E
The results of the analyses were used as starting :ooo~- J .. S 6 7 8 9 10 11 12 13
Time. months
14 IS 16 17 18 19 20 21 22 23 24
I
3E8 P match 9.33.10.9{pSi2/Cpy'
accounts for the variation of gas density with pressure kh
.. k
95 .5md.ft
2.91 md
and temperature. The calculated sandface pressures 2E8
skin
fl.p,
-3 .84
·1860
were analyzed by Rate Transient Analysis methods
using FAST program.
Field example 2
'Ea .
-<r-r..,..,rtT¥F'fT(;-' ,
l ~
I I i I I I I I I I
4
I i i · ......' i I I
~
I I I , i i i rr
h~~~~~.,...,-~r~l--r- ~~r=;:::';:::::;:=i=r=r-.--.---1
Q (IMSCF/d a y) v. ra ua T (Y-~J
'000 .
100000 -; I
. · .lJI~r·"·· ..· . . . . 1
90000
aoooo
P p g ) & Q [MSCFI"'-Y) ".rau. T .....,
70000 -:::
Fig. 12. History match for the whole test sequence. LIT (m(p): l.P.R.
Test type: MIT
Reservopir pressure, 2701. 187
AOFP 4562.2 Mscfld
A history match for the whole test sequence is A 151246
shown in Fig. 18 and in Fig. 19 for the semi log plot B 12.0242
10
7
r------ Conclusions
1 Smoothing 0. 1
Radius of investing. 707 ft
~p, (skin) 499.9676 psi.
1 From the results of this work, the following
6
-'~~I '"
10
'0 ' , conclusions are warranted:
d on(P) ( p ~l21cp] ve r sus d ~ (hrJ
Shales." SPE, 9397, (Sept. 1980). Prats, M. "Effect of Vertical Fractures on Reservoir Behavior-
Larry, K.; Jack, R.J. and Harmon, H.J. "Application of After- incompressible Fluid Case." SPEJ, 1575-G, (June 1961),
closure Analysis Techniques to Detennine Permeability in 105-118.
Tight Fonnation Gas Reservoir." SPE, 90865, (Sept. 2004). Raghavan, R. and Ohaeir, C.V. "Unsteady Flow to a Well
Lee, J. and Brockenbrough, R. "A New Approximate Produced at Constant Pressure in a Fractured Reservoir."
Analytical Solution for Finite Conductivity Vertical SPE, 9902, (March 1981).
Fractures." SPE, 12013-PA, (February 1986),75-88. Rahman, A.; Poolad, M. and Mattar, L. "Perforation Inflow
Lee, W.J. and Holditch, S.A. "Fracture Evaluation with Test Analysis (PIT A)." Petroleum Society-Canadian
Pressure Transient Testing in Low-permeability Gas International, 95510-MS, (June 2005).
Reservoir." JPT, 9975-PA, (Sept. 1981),1776-1792. Ramey, J. "Non-Darcy Flow and Wellbore Storage Effects in
Mattar, L. and Santo, M. "Well Testing of Tight Gas Pressure Buildup and Drawdown of Gas Wells." JPT, Vo!.
Reservoirs." SPE, 100576, (May 2006). 17, (February 1965),223-223.
Najurieta, H.L. "A Theory for Pressure Transient Analysis in Russell, D.G. and Truitt, N.E. 'Transient Pressure Behavior in
Naturally Fractured Reservoirs." JPT, 6017-PA, (July Vertically Fractured Reservoir." JPT, 702-PA, (Oct. 1964),
1980), 1241-1250 1159-1170.
Odeh, A.S. "Unsteady-state Behavior of Naturally Fractured Streltsova, T.D. "Well Pressure Behavior of a Naturally
Reservoirs." SPEJ, 966-PA, (March 1965),60-66. Fractured Reservoir." SPEJ, 10782-PA, (Oct. 1983), 769-
Ohaeri, C.U. "Pressure Buildup Analysis for a Well Produced at 780.
a Constant Pressure in a Naturally Fractured Reservoir." Van Everdingen, A.F.V. and Burst, W. "The Application of
SPE, 12009, (Oct. 1983). the Laplace Transfonnation to Flow Problems in
Partikno, H. "Decline Curve Analysis Using Type Curves- Reservoirs." Translated, AIME, 186, (1949), 305-324B.
fractured Wells." SPE, 84287, (Oct. 2003). Warren, J.E. and Root, P.J. "The Behavior of Naturally
Pollard, P. "Evaluation of Acid Treatment from Pressure Fractured Reservoirs." SPPJ, 426-PA, (1963), 245-255.
Buildup Analysis." Translated, AIME, 981-G, (1959), 216.