See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/266668867

A Method for Pattern Recognition of WOR Plots in Waterflood Management

Article · March 2005

DOI: 10.2118/93870-MS

CITATIONS READS
0 1,096

2 authors:

Zhengming Yang I. Ershaghi

Aera Energy LLC University of Southern California
32 PUBLICATIONS   260 CITATIONS    14 PUBLICATIONS   67 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Waterflood analysis View project

Steamflood View project

All content following this page was uploaded by Zhengming Yang on 05 November 2014.

The user has requested enhancement of the downloaded file.

SPE 93870
A Method for Pattern Recognition of WOR Plots in Waterflood Management
Zhengming Yang, SPE, EXGEO(CGG Venezuela), Iraj Ershaghi, SPE, University of Southern California
Copyright 2005, Society of Petroleum Engineers Inc.
This paper was prepared for presentation at the 2005 SPE Western Regional Meeting held in
Irvine, CA, U.S.A., 30 March – 1 April 2005.
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
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
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
SPE meetings are subject to publication review by Editorial Committees of the Society of
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
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.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.
Abstract
Pressure maintenance by waterflooding is a prominent
recovery process in many of California oilfields. Majority of
these turbidite sandstone reservoirs contain highly
heterogeneous permeability profiles affected by alternation of
sand shale sequences and the clay content. A typical signature
of such heterogeneous profiles is often the existence of a high
permeability streak serving as a ”thief zone”. Mechanical or
chemical blocking of such layers on the injection or
production side has been an objective of water flood
management.
This paper presents a novel method for using a series of
customized type curves to scrutinize the WOR characteristics
plots of producing wells in stratified reservoirs. From a library
of systematically generated performance plots for several
combination of rock and fluid properties, and various
architectural position of the high permeability zone, the actual
data for a given well or a group of wells can be matched
against various patterns to identify the degree of heterogeneity
and the potential existence of a high permeability streak.
With such a pattern recognition tool, opportunities to enhance
the performance of individual wells or group of wells can be
ascertained. Additionally, the impact of permeability
modification by mechanical or deep chemical zonal isolation
has also been studied. In brief, we have generated yardsticks
by which the effectiveness of wellbore isolation or in-depth
plugging of high permeability streak can be examined.
Introduction
Many oilfields in the Los Angeles Basin are currently under
waterflood. Because of stratified nature of these sandstone
reservoirs, inefficient vertical conformances are very common
[1,2]
.
Detection of poor vertical conformance, as caused by presence
of high permeability layers, is critical in the management of
optimized waterfloods. Performance data from producing
fields are often depicted in the form of certain diagnostic plots.
Among them, relating the instantaneous water oil ratio or
some function of this parameter to the cumulative oil
production has been a routine basis for managing waterfloods.
Based on the solution of the one-dimensional Buckley–
Leverett equation and under the assumption of exponential
functional relation of the relative permeability ratio with
saturation, Ershaghi et al [3,4] proposed the so-called X-plot to
interpret and forecast recoveries based on the historical
performance of a well or a group of wells. This derivation
provided a basis for the use of empirical method of Water/oil
ratio (WOR) diagnostic plots [5]. These plots have also been
proposed as a convenient method to identify mechanisms
causing excessive water production [6-7]. These observations
should be in conjunction with other available reservoir and
well completion data. Drawing general conclusions for solving
the water production problems from field observation must be
handled with care. For example, Chan [6] suggested that rapidly
increasing WOR curves with positive derivative slopes are
indicative of water channeling, and gradually increasing WOR
curves with negative derivative slopes are indicative of coning
problems. Based on simulation studies, Seright [8]
demonstrated that in many cases, multi-layer channeling
problems could easily be mistaken as bottom water coning,
and vice versa, concluding that slope of WOR diagnostic plots
should not be used in isolation to identify mechanisms for
excessive water production. A more realistic approach is to
narrow down the problem and make the diagnostic plots more
field-condition specific. Together with other information,
these customized plots then should be able to help diagnose
some of the mechanisms responsible for excessive water
production.
The purpose of this study is to develop diagnostic plots of
water oil ratio vs. oil recovery and/or time with a focus on
stratified systems and in the presence of a thief zone. Further
more, based on reservoir simulation studies, we have
examined the effectiveness of fluid diversion in such systems
by various measures.

Development of the Diagnostic Plots
Diagnostic plots for this study were designed based
on mechanistic description of the oil-water displacement for
stratified reservoirs. From the point of view of immiscible
displacement in stratified systems, three main parameter
groups are considered; mobility ratio M, modified aspect ratio
RL [12,13] and reservoir heterogeneity defined by heterogeneity
index KHR. As shown in Figure 1, stratified reservoirs
represented by equivalent systems, may be grouped into by
three types:
Homogeneous and anisotropic reservoirs (Type I) are defined
by only two lumped parameters M and RL. Stratified reservoirs
with cross flow (Type II) are characterized by M, RL and
KHR. Layered reservoirs without cross flow (Type III) are
characterized by only two lumped parameters M and a KHR.
These parameter groups are defined as follows:
K rwm K rom
M = (1)
µw µo
L kV
RL = (2)
H kH
∑ ( kh )
(3)
KHR = S
∑ ( kh )
H
where krom and krwm are end point relative permeability values,
µO and µW are fluid viscosities; and kV/kH is the vertical to
horizontal permeability ratio. KHR is defined as the fraction of
flow capacity in thief zone over the total flow capacity, where
∑S(kh) is the summation of flow capacity in the thief zone and
∑H(kh) is the summation of total flow capacity over the whole
producing interval including thief zones. As a minimum, and
from a practical point of view, for a given producing well, the
fluid viscosities µO and µW, the reservoir thickness H and well
spacing L should be known. The relative permeability end
points can be obtained from core flood tests. Vertical-to-
horizontal permeability ratio may be estimated from core
analysis or other geological inferences [9].
Heterogeneity is interpreted in geological terms as changes in
facies types. For mathematical modeling purpose, however,
heterogeneity can be described by variations in petrophysical
properties, such as porosity and permeability in the
longitudinal and transverse directions. In building the
simulation model, formation heterogeneity can be described
by permeability contrast and vertical permeability.
Our study focused on the uniformity of permeability profile.
The parameter KHR is proposed to describe the flow capacity
of the high permeability thief zone as fraction of the total
system flow capacity.
A series of type curves were generated from multitude of
simulation runs with realistic parameter ranges covering
conditions for waterflooding of light-medium oil, with
mobility ratios (M) from 0.1 to 30, modified aspect ratios (RL)
from 0-8 and heterogeneity indices (KHR) from 0 to 0.9.
Simulation Model
Individual runs were conducted on a model consisting of a
layered reservoir for 1/8 section of a symmetric 5-spot well
pattern consisting of one injector and one producer, as shown
in Figure 2.
For our simulation runs, we used the Corey-type relative
permeability equations in exponential forms:
Sw − Swir
Krw = Krwm ( ) nw (4)
1 − Sorw − Swir
1 − Sw − Sorw now
Krow = Krom ( ) (5)
1 − Sorw − Swc
From our sensitivity studies, for the purpose of diagnostic type
curve matching, we noted that the critical parameters are the
fluid viscosities and the end points. The curvature of the
relative permeability plots as affected by the exponential
values nw and now when in the range of 1.5 and 2.0 were
found not to be greatly affecting the type curves.
Type Curves for Specified Lumped Parameters
Results generated for a multitude of reservoir and
flow conditions were plotted in the form of type curves.
These plots were examined on the basis of certain
characteristics that included the extent of recovery before
breakthrough, the shape of the WOR rise after breakthrough,
presence or absence of a stabilized WOR period while
realizing incremental recoveries and the manner for final rise
in WOR, as shown in Figure 4.
Theoretical Expectations
As predicted by the fractional flow curve, for a
homogeneous system, one expects a breakthrough WOR
exceeding 1. Breakthrough WOR’s of less than one signifies
the presence of a stratified system. This is caused by
composite performance including clean oil from tighter layers
and breakthrough water from more permeable ones. As such,
stratified systems that exhibit substantial recoveries before
WOR reaches to 50% are indicative of uniform permeability
profiles.
Application
The type curve library generated in this study and
available as a matching tool allows recognition of the impact
of a thief zone as well as a basis for long term recovery
forecast at a given economic limit WOR.
Base Case
We examined a Type I system (no thief zone) with a
focus on mobility ratio. As shown in Figure 5, for such a
system, one can expect substantial recoveries before
breakthrough for light oil reservoirs and lesser amounts for
medium gravities. For medium gravity oils, recoveries will
continue with an exponential rise in WOR.
As shown in Figure 6, for Type I systems, the influence of
kv/kH as represented by RL seems to be minimal on the type
SPE 93870
curve. Recovery expectation at high WOR can be seen from
the type curves. For example, for light oil with M=1, taking
the economic WOR as 50, indicates an ultimate fractional
recovery of 0.525 which is 1.3 times the recovery observed at
WOR=1. For a medium gravity oil (M=10), the ultimate
recovery at the same WOR is 0.46 which is 2.6 times the
recovery at WOR=1, indicating a lower recovery at WOR=1
and the longer flooding period required to realize that
recovery. This means that by examining the performance data,
after ascertaining that no thief zone is affecting the
performance, and using realistic estimates of mobility ratio,
one can predict ultimate recoveries given the observed
recovery at WOR=1.
Presence of a Thief Zone (Type II Systems)
For the same range of mobility ratios shown for the
base case, predicted recoveries before breakthrough for both
light and medium gravity oil drop are substantially lower. A
major signature for the light oil is the development of a
stabilized period where increasing amounts of recoveries are
realized before the final rise of WOR.
Effect of the KHR on Type II systems
Figures 7 and 8 show the impact of KHR for
type II reservoirs and for light to medium gravity oils
respectively. For a favorable mobility ratio of 1,
recoveries at WOR=1 are minimally affected by the
presence of a thief zone as long as the flow capacity of
the zone I less than 50 % of the total flow capacity, that
is, KHR<0.5. Thief zones, which by the virtue of their K
and/or H constitute 70% or more of the flow capacity,
can cause a significant drop in the recovery at WOR=1
while the recoveries at WOR=50 are very similar.
For an unfavorable mobility ratio of 10, both the
recoveries at WOR=1 and at WOR=50 are considerably
affected by thief zone with flow capacities of 0.5 or
higher.
Impact of Aspect Ratio RL
Figures 9 and 10 show the effect of KHR and RL
of the reservoirs. Here the influence of cross-flow between the
bulk of the formation and the thief zone is examined. In both
cases, the displacement systems approach the vertical flow
equilibrium (VFE) limit, which enhances the recovery before
breakthrough at RL’s of around 8.
Structural Positioning of the Thief Zone
Our type curve studies show that the recovery before
breakthrough and the overall WOR patterns are independent of
various setting of the thief zone for no or low cross flows, as
shown in Figure 11. For higher cross flows, still with minimal
changes in the early recoveries before the breakthrough, the
role of gravity becomes important depending on the
positioning of the thief zones, as shown in Figure 12. Here,
the gravity negatively impacts the water flood performance for
a thief zone located in the bottom of the stack.
3
Type Curve Matching for Field Diagnostics
Results of the simulation runs were indexed into a
spreadsheet database for generation of customized type curves.
This methodology allows generation of the type curves closest
to the actual field characteristics for given estimation of the
applicable mobility ratio. The matching process can be done
by an iterative process for the estimation of the cross flow as
represented by RL.
The number of simulation runs is limited to discrete values
assumed for M, RL and KHR. In this study, Simulation runs
were made on combination of the following values:
M: 0.1, 1, 2, 3, 4, 6, 8,10,15,20,25, 30.
RL 0, 1, 2, 3, 4, 5, 6, 8.
KHR 0, 0.3, 0.5,0.7, 0.9.
That is a total of 480 simulation runs. For fractional values in
between, an interpolation routine was developed to generate
the type curves from the results of the simulation runs in the
database.
Figure 13 shows a comparison between direct simulation and
interpolation. The error caused by interpolation is insignificant
and will not affect the physical meaning and application of the
diagnostic plots.
Effectiveness of Mechanical Blocking
With the recognition of the presence of a thief zone,
improvements can be obtained in the flood performance with
some method of profile control. There are several techniques
that have been proposed for handling of the fluid diversion
from the thief zone in a layered reservoir [10-11]. These include,
cement squeeze, polymer squeeze, combination squeeze and
liner/casing patches and zonal isolation. With the capabilities
developed in this study to examine the influence of cross flow,
we also made an attempt to scrutinize the required extent of
in-depth plugging before significant improvement is observed
in the water flood.
In case with no cross flow between the bulk formation and the
thief zone, as shown in Figure 14, the earlier the mechanical
blocking and the zonal isolation is exercised, the sooner the
ultimate recovery is realized. In the mean time, the
homogeneous case of KHR=0 can be approximately
considered as the limit of the water flood efficiency for a
heterogeneous case by blocking measure.
For layered formation with large aspect ratio (strong cross
flow) as shown in Figure 15, it is evident that the superficially
blocking high permeability streak at the sandface is not
effective in improving the performance. There is great
potential of improvement bases on comparison with the
homogeneous case. In this case, other in-depth blocking
methods such as polymer gel injection may be necessary to
improve the water flood efficiency.
Effectiveness of In-Depth Polymer Gel Blocking
Figures 16 through 19 show the simulation results of
in-depth plugging with different blocking depth and for
4 SPE 93870
different levels of RL (cross flow). The effect of polymer
blocking measure is modeled by reducing the permeability of
the thief zone value at different penetrating depth from
injection and production wells in the reservoir model. It is
evident that in-depth blocking is much more efficient for
formations with strong cross flow. Also, minimal depth of
plugging will result in similar observations as in the case
observed for superficial isolation. For cases with substantial
cross flow, the in-depth plug may need to be as deep as 1/3 the
spacing between the injector and the producer as shown in
Figure 19.
Field Case Study
The case study below is from one producer under
waterflood. The well produces light oil of viscosity 1.2 cp
from the reservoir. The spacing between injector and producer
is 330 ft and the formation is 200 ft thick. For other
information, water viscosity is estimated at 0.59 cp, end point
of oil relative permeability is estimated at 1.0 and water
relative permeability is estimated at 0.15. Based on other
information, the vertical to horizontal permeability ratio is
high for this formation and is estimated at 0.8. High permeable
thief zone is generally in the middle of the producing interval.
Based on these information, the lumped parameters can be
estimated as M = 0.407 and RL = 1.476. Among the three
main lumped parameters, estimates of M and RL are better
known. One of the purposes of analysis by the diagnostic
plots is to estimate the parameter KHR and also to apply the
diagnostic plots to evaluate the potential of improved
waterflood efficiency from this formation.
Figure 20 shows the results of the diagnostic plot analysis for
the case with the thief zone in the middle of the producing
interval. The parameter KHR is estimated at 0.7. Therefore,
the permeability heterogeneity is severe in this field.
Considering the homogeneous case (KHR=0) as the limit of
the water flood efficiency improvement by any blocking
measure, there would be improvement if blocking measures
were taken in the early time of water breakthrough.
Summary and Conclusions
Based on analytical modeling and reservoir
simulation studies, water flood response in wells producing
from stratified reservoirs have been estimated in terms of
WOR diagnostic plots. In particular the influence of a thief
zone on the oil recovery at the system breakthrough (WOR=1)
as a function of mobility ratio and cross flow has been
characterized for different levels of heterogeneity index. The
type curve method allows a rapid method for history matching
of performance data characterizing the influence of an existing
thief zone and allowing estimates of ultimate recovery based
on recoveries observed at WOR=1.
In this study, the effectiveness of superficial or in-depth
plugging has also been examined. In the presence of cross
flow between the bulk of the formation and the thief zone,
superficial isolation is not effective and a substantial depth of
treatment must be considered before in-depth plugging can
improve the performance.
A spread-sheet Program allowing the generation of the
diagnostic plots and history matching the actual data may be
obtained from http://www.westcoastpttc.org.
Acknowledgements
This study was sponsored by U S DOE PUMP
program and a grant from the California Energy Commission
(CEC). We thank Mr. Tony Wong and Pramod Kulkarni from
CEC for their support during the course of this study.
References
1. J. Roberston, J. Blesssner and S. SooHoo: “Subzone Re-
Development of the Long Beach Unit, Wilmington
Oilfield, A Case Study, JPT, Oct. 1987 (1229-1236).
2. M.K. (Val) Lerma: “Analytical Method to Predict
Waterflood Performance” paper SPE 83511 (May 2003).
3. Ershaghi, I. and Omoregie, O.: ‘‘A Method for
Extrapolation of Cut vs. Recovery Curves,’’ JPT (1978)
203.
4. Ershaghi, I. and Abdassah, D.: ‘‘A Prediction Technique
for Immiscible Processes Using Field Performance Data,’’
JPT (1984) 664.
5. Lo, K.K., Warner, H.R. Jr., and Johnson, J.B.: ‘‘A Study
of the Post- Breakthrough Characteristics of
Waterfloods,’’ paper SPE 20064 presented at the 1990
SPE California Regional Meeting, Ventura, California 4–
6 April.
6. Chan, K.S.: ‘‘Water Control Diagnostic Plots,’’ paper
SPE 30775 presented at the 1995 SPE Annual Technical
Conference and Exhibition, Dallas, 22–25 October.
7. Yortsos, Y., Choi, Y. and Yang Z.M.: “Analysis and
Interpretation of Water/Oil Ratio in Waterfloods”, SPE
Journal, (Dec. 1999) 413-424.
8. Seright, R. S., Lane, R. H. and Sydansk, R. D.: “A
Strategy for Attacking Excess Water Production”, SPE
70067, Midland, Texas, 15–16 May 2001.
9. Weber, K.J: “Influence of Common Sedimentary
Structures on Fluid Flow in Reservoir Models”, Journal of
Petroleum Technology (March 1982) 665-672.
10. R J Fulleylove, J C Morgan, D G Stevens and D.R
Thrasher: “Water Shut-Off in Oil Production Wells -
Lessons from 12 Treatments SPE # 36211, SPE Meeting
Abu Dhabi, UAE, October 1996.
11. M.K. Lerma and M. A.Giuliani: “Cost-Effective Methods
of Profile Control in Mature Waterflood Injectors,” SPE
27849, SPE Western Regional Meeting 1994.
12. Lake, L.W., “Enhanced Oil Recovery”, Prentice Hall,
New York (1989).
13. L.P. Dake, “The Practice of Reservoir Engineering
(Revised Edition)”, Elsevier Science Pub Co (May 2001).
 Figure 1. Equivalent Systems to Stratified Reservoirs.
Figure 4. Basic Features of the Type Curves.

Figure 2. Well Pattern Used in the Simulation Study.

Figure 5. Effect of Mobility Ratio for

Homogeneous Formation.

Figure 3. Effect of the Curvature of Relative

Permeability on Diagnostic Plot.

Figure 6. Effect of RL on WOR plots for

Homogeneous Formation.
6 SPE 93870

Figure 7. Effect of Mobility Ratio for Heterogeneous Figure 10. Effect of RL on WOR Plots for
Formation. Heterogeneous Formation.

Figure 8. Effect of KHR for Favorable Mobility Ratio.

Figure 11. Effect of the modified aspect Ratio RL on

WOR plots (RL =0.5).

Figure 9. Effect of KHR for Unfavorable Mobility for

Unfavorable Mobility Ratio.
SPE 93870 7

Figure 12. Effect of the Modified Aspect Ratio RL on

WOR Plots (RL =2.0). Figure 14. Effect of Mechanical Blocking on WOR
Plots for Heterogeneous Formation With RL =0.

Figure 13. Comparison Between Direct Simulation Figure. 15. Ineffectiveness of mechanical isolation timing
and Interpolation. on WOR behavior when high cross flows exist in the
formation.
8 SPE 93870

Figure 16. Effect of Blocking Depth, M=4, RL=1, KHR=0.7). Figure 19. Effect of Blocking Depth (M=4, RL=4, KHR=0.7).

Figure 17. Effect of Blocking Depth (M=4, RL=2 , KHR=0.7). Figure 20. Diagnostic Plots Results for the Field Case Study.

Figure 18. Effect of Blocking Depth (M=4, RL=3, KHR=0.7).

View publication stats