SPE 90013

Use of Production Data Inversion to Evaluate Performance of Naturally

Fractured Reservoirs
T. Marhaendrajana, Institut Teknologi Bandung; T.A. Blasingame, Texas A&M University; and J.A. Rushing, Anadarko

Copyright 2004, Society of Petroleum Engineers Inc.

fractured reservoirs. Others3,4,5 considered the effect of
This paper was prepared for presentation at the 2004 SPE International Petroleum Conference transient interporosity flow. The transient model is physically
in Mexico held in Puebla, Mexico, 8–9 November 2004.
much more appealing since it allows for all regimes of internal
This paper was selected for presentation by an SPE Program Committee following review of
information contained in a proposal submitted by the author(s). Contents of the paper, as
block flow, i.e. transient, late transient and pseudosteady-state.
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to This was first proposed by Kazemi,3 where the matrix blocks
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at were composed of slabs.
SPE meetings are subject to publication review by Editorial Committees of the Society of The analytical solution in Laplace space for the wellbore
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is pressure response of a dual porosity reservoir has the form:6,7
prohibited. Permission to reproduce in print is restricted to a proposal of not more than 300

( )( ) ( ) ( )
words; illustrations may not be copied. The proposal must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O. 1 K1 reD sf ( s ) I 0 sf ( s ) + I1 reD sf ( s ) K 0 sf ( s )
p fD ( s ) = s → tD
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.
( ) ( ) ( )(
s sf ( s ) I1 reD sf ( s ) K1 sf ( s ) − K1 reD sf ( s ) I1 sf ( s ) )
............................................................................................. (1)
Abstract
Many oil and gas production come from naturally Eq. 1 is the constant well rate solution. The Laplace parameter
fractured reservoirs. The reservoir properties and production function f(s) depends on the type of interporosity model and
performance of this reservoir show a unique behavior, which the matrix block geometries. This paper considers two models:
are different from the homogeneous reservoir. Hence, (i) pseudosteady-state interporosity flow model with matrix
evaluation and forecasting production of these reservoirs blocks as cubes, and (ii) transient interporosity flow model
require special models and approaches. with matrix blocks as slab (strata). The function f(s) for the
This paper attempts to utilize production data inversion two models are:7,8
method to obtain parameters such as permeability, skin factor,
initial fluid-in-place and production potential in a naturally Pseudosteady-state interporosity flow model with matrix
fractured reservoir. The model can then be used to forecast blocks as cubes:
production from this type of reservoir.
The method uses type curve approach that incorporate ω (1 − ω ) s + λ
f (s) = ...................................................... (2)
concepts of both transient and boundary dominated flow (1 − ω ) s + λ
models. Application of this method to oil field data is
presented as well as comparison with results obtained from Transient interporosity flow model with matrix blocks as slab
other methods (such as well test analysis and volumetric (strata):
calculation).
λ 3(1 − ω ) s 3(1 − ω ) s
The results of this research provide engineers a tool to f (s) = ω + tanh ..................... (3)
evaluate and monitor production/reservoir performance of 3s λ λ
naturally fractured reservoirs regularly by analyzing where:
production data (which are always recorded) without
additional testing. k m rw2
λ =α .................................................................... (4)
k f hm2
Introduction
Modeling of single phase flow in fractured reservoir
involve the description of the shapes, orientation and sizes orf (φct ) f
ω= ....................................................... (5)
fractures network, and the mechanism for the transfer of fluid (φct ) f + (φct ) m
from the matrix to the fracture network. The process of
transfer of fluid between matrix blocks and fisssures (or The dimensionless pressure and dimensionless time are
fractures) is the points at which various model differ. Some defined by:
investigators have considered a pseudosteady-state relation for k f h( pi − p wf )
the interporosity flow term. This approach was first suggested p fD = ..................................................... (6)
by Barenblatt et al1 and was taken up by Warren and Root2 in 141.2qBµ
their pioneering study of the pressure response of naturally
2 SPE 90013

0.0002637k f t Cases where well flowing pressure or wellhead pressure

tD = ........................................................ (7) data is available
(φct ) m + f
Next is to use the type curves model to perform production
data inversion to obtain reservoir parameter, such as
Method for Production Data Inversion permeability, skin factor, drainage area and oil-in-place. The
The wellbore rate response (solution) in Laplace space for a production data are matched with the decline type curve. Once
constant well flowing pressure can be obtained from Eq. 1 a good match is obtained, we get match parameters, which are
using the Duhamel principles that is: reD, ω, λ, [q/∆p]MP, [t]MP, [qDd]MP, and [tDd]MP. Hence we
compute permeability and oil-in-place.
1
q D ( s) = 2
.......................................................... (8)
s p fD ( s ) 1. Estimation of Permeability:

The dimensionless rate is defined as Combining Eqs. 9 and 11 we obtain

141.2qBµ 141.2qBµ
qD = ......................................................... (9) q Dd = [ln reD − 0.5] ................................. (12)
kh( pi − p wf ) kh( pi − p wf )
The real domain solution of Eq. 8 is obtained using Rearranging Eq. 9, permeability is computed by
numerical inversion so that for various values of λ, ω, and reD,

[ln reD − 0.5] [q / ∆p]MP ............................ (13)

the dimensionless well rate solution, qD vs tD, can be
141.2 Bµ
generated. Following Fetkovich,9 decline curve dimensionless k=
time, tDd, and decline curve dimensionless rate, qDd, are h [q Dd ]MP
defined by
2. Estimation of Original Oil-in-Place:
tD
t Dd = ....................................... (10) Combining Eqs. 7 and 10
0.5[reD − 1][ln reD − 0.5]
0.0002637k f t 1
q Dd = q D [ln reD − 0.5] ................................................... (11) t Dd = ............... (14)
φct µrw2 0.5[reD − 1][ln reD − 0.5]
The decline curve dimensionless rate for pseudosteady-
state interporosity model is shown in Fig. 1. At early time or
fluid is produced from fracture networks and the flow regime 0.00634k f t 1
t Dd = ................... (15)
0.5[reD − 1][ln reD − 0.5]
is transient. When the entire fracture networks connected to
the wellbore is affected by the production, the well rate φct µrw2
decreases sharply. At this period the flow in the fracture is Rearranging Eq. 15, the oil-in-place can be computed as
under pseudosteady-state flow. During this time the pressure follows
in the fracture reduces substantially and the fluid flow from
matrix into the fracture becomes dominant. This is shown by 1 [t ] MP [q / ∆p ]MP
N= .......................................... (16)
the flattening well rate, as an indication of pseudosteady-state ct [t Dd ] MP [q Dd ]MP
interporosity flow. The well rate, again, decreases sharply as
the entire system (fractures and matrix systems) is affected by 3. Estimation of Reservoir Characteristics:
the production. The relations given below are used to estimate volumetric and
Higher storativity or capacity ratio (ω) means that the flow characteristics of the reservoir based on the results of the
matrix storativity is less. Hence the early response (which is type curve match and the available data.
from fracture networks) is higher (denoted by dash-line). If the Reservoir Drainage Area
storativity ratio, ω, equal to one, fluid production comes from
the fracture networks because the matrix system is not porous. NB
A = 5.615 ................................................. (17)
Therefore, the well performance behaves as the reservoir is φh(1 − S wirr )
homogeneous.
As the interporosity flow parameter, λ, is smaller the fluid Reservoir Drainage Radius
flow from matrix to the fracture system is smaller. When the
A
interporosity flow is zero, there is no flow between the two re = ....................................................................... (18)
systems (matrix and fractures). At this case, the well π
performance also behaves as the reservoir is homogeneous. Effective Wellbore Radius
The decline type curve of naturally fractured reservoir
using transient interporosity model is shown in Fig. 2. The re
rwa = ...................................................................... (19)
obvious difference of the transient model from the reD
pseudosteady-state model is the logarithmic well rate response
decreases linearly (slope is equal to -0.5) with the logarithmic
time. The comparison between both models is shown in Fig. 3.
SPE 90013 3

Skin Factor from the reservoir X. This well has been produced for about
25 years as it is shown in Fig. 4.
⎛r ⎞
skin = − ln⎜⎜ wa ⎟ ........................................................... (20)
⎟
The reservoir consists of reef limestones, and dolomitation
⎝ rw ⎠ process occurred in this reservoir. Core analysis results
showed naturally fractured characteristics. This observation
was supported by well test data taken recently in 2002. The
Cases where well flowing pressure and wellhead pressure Horner and log-log plot analysis of these data are shown in
data are NOT available Figs. 5 and 6.
The Darcy equation for pseudosteady-state flow is At first, the conventional Arps decline curve was used to
kh( pi − p wf ) analyze the production data of Well-1 and to identify reservoir
q= ............................................. (21) character. This is shown in Fig. 7. At beginning, the data
141.2 Bµ[ln reD − 0.5] follows decline curve with b=0 indicating pseudosteady-state
Defining the maximum rate (i.e. the rate of well produced at production under constant well flowing pressure. At about
its maximum capacity) tD=7, the data deviate from decline curve of b=0 and they do
not fit any of b value. Instead, the data follows the half-slope
(qi )max =
khpi
.................................... (22)
decline indicating the transient interporosity flow from matrix
141.2 Bµ[ln reD − 0.5] system to fracture systems. At tD of about 10 the data showing
another sharp decline. An interesting fact is that the
Defining Npi production data do not continue declining, but they show
another level of interporosity flow. This may be an indication
π (re2 − rw2 )φct hpi
N pi = ............................................... (23) of three porosity systems. It also may be due to changing
5.615 B production constraint (e.g., changing choke size, closing and
Writing decline curve dimensionless time (Eq. 10) in term of reopening well, and stimulation). At the end, production
parameters in Eqs. 22 and 23, we have relation as follows declines which indicates the entire system is under
pseudosteady-state flow.
⎡ (q ) ⎤ To avoid ambiguity the effect of production constraint
t Dd = ⎢ i max ⎥t .......................................................... (24) should be eliminated by normalizing production rate with
⎢⎣ N pi ⎥⎦
pressure drop, if flowing well pressure or wellhead pressure
Recalling decline curve dimensionless rate data are available. In our case, those data are not available.
The exercise using the Arps decline type curve showed
141.2qBµ
q Dd = [ln reD − 0.5] ................................ (25a) that the use of Arps decline curve is not appropriate to analyze
kh( pi − p wf ) production data from Well-1, which is producing from a
naturally fractured reservoir. The decline type curve generated
or we can write for this type of reservoir (as mentioned in the previous
q sections) are used to evaluate the production data to obtain
q Dd = ............................................................ (25b) permebility, skin factor, drainage area, effective wellbore
(qi ) max
radius and original oil-in-place. The match results is shown in
1. Estimation of Permeability: Fig.8 with parameters as follow:

Rearranging Eq. 25 we compute permeability as reD = 1000 ω = 0.2 λ = 4 × 10 −8 [t / t Dd ]MP = 400 / 1

141.2 Bµ = 15000 / 1
k= [ln reD − 0.5] [q]MP .............................. (26) [q / q Dd ]MP
hpi [q Dd ]MP
Fluid and reservoir data are given in Table 1 below:
2. Estimation of Original Oil-in-Place:
Oil-in-place can calculated from Eq. 24 and Eq. 25 as follows
Table 1 – Fluid and Resevoir Properties
1 ⎡ q ⎤ ⎡ t ⎤
N= ⎢ ⎥ ⎢ ⎥ ........................................ (27)
ct p i ⎣ q Dd ⎦ MP ⎣ t Dd ⎦ MP Oil viscosity = 1.52 cp
Oil FVF = 1.04 RB/STB
3. Estimation of Reservoir Characteristics: Initial pressure = 3300 psia
Net pay thickness = 100 ft
The procedure and equations for this purpose are the same as
Porosity = 13 %
in the case of pressure history is not available.
Initial water saturation = 26 %
Total compressibility = 4.3421x10-5 1/psi
Wellbore radius = 0.3 ft
Field Application
Production data from well-1 producing from reservoir X were
The analysis results of production data inversion are
used to apply this method. Well-X is the only well producing
summarized in Table 2. The production data inversion yield
very good agreement with well test data analysis for
4 SPE 90013

permeability, skin factor, and storativity ratio. The OOIP also recorded during production, in addition to the
obtained from production data inversion 41.9 MMSTB which production rate.
is higher than that obtained volumetrically, which is 33
MMSTB. This may be caused by reservoir drive mechanism 5. Accurate measurement of formation compressibility
being strong water drive (this is supported by the fact that should be guaranteed to obtain accurate original oil-in-
reservoir average pressure just slightly declines for almost 25 place.
years production), which affects the analysis. Nevertheless,
the estimated value of OOIP from production inversion is Nomenclature
reasonable and provides cofidence to the OOIP obtained using
A = drainage area, ft2
volumetric calculation, which has been questioned whether it
B = oil formation volume factor, vol/vol
is too optimistic.
ct = total compressibility, 1/psi
The current recovery factor is about 14.7%, and using the
h = thickness, ft
type curve the production is forecasted based on current
k = permeability, md
performance shown in Fig. 9. The EUR is predicted to be
N = original oil-in-place, STB
15.5%. Low EUR under strong water drive may due to fluid
Pi = initial pressure
flow from matrix to fracture networks is very small because of
Pwf = well flowing pressure, psi
water encroahment into the fracture networks. Small
q = oil rate, STB/D
interporosity flow is indicated by our analysis with small λ of qD = dimensionless rate
4x10-8. As a result, most of the remaining fluid is trapped in
qDd = decline curve dimensionless rate
the matrix (which originally contributes to about 80% of the
re = drainage radius, ft
OOIP as indicated by storativity ratio of 0.2). rw = wellbore radius, ft
sw = water saturation, fraction
Table 2 – Results Summary of Production Data Inversion, t = time, hr or day
Well Test Analysis, Volumetric Calculation tD = dimensionless time
tDd = decline curve dimensionless time
Parameters Prod. Well Test Analysis Volumetric
Data Horner Log-log Calculation
Inversion Plot Plot Symbols
k (mD) 65 62.47 63.71 -
OOIP (MMSTB) 41.9 - - 33 ω = storativity ratio
A (acre) 631 - - -
re (ft) 2958 - - - λ = interporosity coeficient
rwa (ft) 2.96 - - - µ = oil viscosity, cp
S -2.3 -3.81 -3.7 - φ = porosity, fraction
λ 4x10-8 - 2.72x10-5 -
ω 0.2 - 0.21 -
Subscript
Conclusions m = matrix
f = fracture
Our conclusions are as follows: D = dimensionless
1. Conventional Arps decline curve may not be appropriate to
used for analyzing production data from a naturally
fractured reservoir when two ore more porosity systems References
contribute to the fluid flow.
2. We have constructed production data inversion using 1. Barenblatt, G.T, Zheltov, Yu.P. and Kochina, I.N.:”Basic
concepts in the theory of seepage of homogeneous liquids in
decline type curve for naturally fractured reservoirs. For
fissured rocks,” J. Appl. Math. Mech.(June 1960) 1286 – 1303.
this reservoir, the interporosity flow is indicated by 2. Warren, J.E. and Root, P.J.:”The Behaviour of Naturally
flattening production (pseudosteady-state interporosity Fractured Reservoirs,” Soc. Pet. Eng. J. (Sept., 1963) 245 – 255.
flow) or by declining rate with half-slope in log-log plot 3. Kazemi, H.:”Pressure Transient Analysis of Naturally Fractured
(transient interporosity flow). Reservoirs with Uniform Fracture Distribution,” Soc. Pet. Eng.
J. (Dec., 1969) pp. 451 –462.
3. We have demonstrated the application of this method to 4. de Swaan, O.A.:”Analytical Solutions for Determining
Well-1 producing from reservoir X. The results are in Naturally Factured Reservoir Properties by Well Testing,” SPEJ
reasonable agreement with the results of well test analysis (June, 1976).
and volumetric calculation. 5. Cinco-Ley, H. and Samaniego, V.F.:”Pressure Transient
Analysis for Naturally Fractured Reservoirs,” paper SPE 11026
4. The use of this method does not require a specific testing. presented at the 1982 Annual Technical Conference Exhibition
It needs only production rate data. For better accuracy in of SPE, New Orleans, LA., Sept. 26-29.
analyzing well performance and characterizing reservoir, 6. Satman, A.:”Decline Curve Analysis for Naturally Fractured
well flowing pressure or wellhead pressure data should be Reservoirs: A Comparison of Models,” paper SPE 14473.
SPE 90013 5

7. Watson, A.T., Gatens III, J.M., Lee, W.J. and Rahim, Z.:”An
Analytical Model for History Matching Naturally Fractured
Reservoir Production Data,” SPERE (Aug. 1990) 295-306.
8. Stewart, G. and Ascharsobbi, F.:”Well Test Interpretation for
Naturally Fractured Reservoirs,” paper SPE 18173 presented at
the 1988 Annual Technical Conference and Exhibition,
Houston, TX, October 2-5.
9. Fetkovich, M.J.:”Decline Curve Analysis Using Type Curves,”
JPT (June 1980) 1065-1077.
6 SPE 90013

Figure 1 – Decline type curve of naturally fractured reservoir for various ω and λ
at reD=10; pseudosteady-state interporosity model.

Figure 2 – Decline type curve of naturally fractured reservoir for various ω and λ
at reD=10; transient interporosity model.
SPE 90013 7

Figure 3 – Decline type curves comparison; pseudosteady-state versus transient

interporosity models.

Figure 4 – Production History of Well-1.

8 SPE 90013

Horner

28 39 .4 ATCM C alculated Pressure

Pressure
Fracture System
Total System

28 38 .4
Pressure (psi)

28 37 .4

Total System
Slope [psi/log cycle] = -1.579
28 36 .4 Intercept [psi] = 2839
Permeability [md] = 62.47
Skin [0] = -3.81
Delta P Skin [psi] = -5.233
Flow Efficiency [0] = 2.545
P* [psi] = 2839

28 35 .4
0.1 1.0 10 .0 10 0.0 10 00 .0 10 00 0.0

Horner Time Ratio

Figure 5 – Horner plot analysis of well test data; Well-1.

yp
10 .00
Ca lculated Pressure
Ca lcula ted Deriv a tiv e
Measured Pressure
Mea sured D eriv a tiv e

1.0 0
PD and Derivative

0.1 0
Dual P orosity (Pseudosteady)
Permeability [md] = 64. 96 0. 0296
Lambda [0] = 2. 723e-005 (Fixed)
Omega [0] = 0. 2145 0. 0554
Skin [0] = -3. 7 (Fixed)
Wellbore Storage [bbls/psi] = 0. 01 (Fixed)
Delta P Skin [psi] = -4. 878
Flow Efficiency [0] = 2. 458
Radius of Inve stigation [ft] = 406. 1

0.0 1
0.1 1.0 10 .0 10 0.0 10 00 .0 10 00 0.0
tD/CD
Figure 6 – Log-log plot analysis of well test data; Well-1.
SPE 90013 9

Figure 7 – Match of production data on Arps decline type curve.

Figure 8 – Match of production data on decline type curve; naturally fractured reservoir model.
10 SPE 90013

Figure 9 – Production forecast of Well-1.