.

SPE
SPE 22695 Modeling Geological Flow Simulation
q SPE Members
,.. . . . .. 1 ,,.

Heterogeneities

and its Impact on

V. Suro-P&ez, P, Ballin,* K. Aziz, * and A.G. Joyrrwl,* Stanford U,

,,, :..
Copyright 1S91, society of Pdroiium Engineers Inc, :( /

,. ; )

T { /,/, /,) ) l 1 1 / This praperwee prepared for prewntation at the ~th Annual Technical Conference and Exhlbl! n of the Soclety of Pefroleum Englnwr& held Irr Dellea, TX, October S-9, 19S1. P Thla paper wee aalec!ed for preaentatlon by an SPE Program Committee following review of inf rmalion contelned In en abstract submitted by the author(a). Contents of the paper, es preaanfed, have not been reviewed by tha society of Petrobum Englrtaers and awesubject t} correction by the author(s). Tha maferial, as presented, dose not n6cewerlly reflect eny poaitlon of the Soclefy of Petroleum Englnwra, Ila officers, or membere. Papers presented at SPE meetings ere subject to publication review by Editorial Oommltteeaof the Sodety of PetroleumEngineers,Permiaalon10copy Ie reafrfcfadto en abstractof not mcwe than 390words. IIluafratfonsmaynot be copied. The abstractshould contain conspicuousacknowledgment of whera and by whom the papw is prewnted, Write Publlcatlona Manager, SPE, P.O. Sox W3S3S, Richardson, TX 76083-3836 U.S.A. Telex, 730989 SPEDAL

.-:-..;-(,{

Introduction
,,.<., /..,,, /,

Abstract A single algorithm is proposed for the numerical modeling of geological facies and corresponding petrophyeical properties. All reservoir parameters, whether categorical such as Iithofaciea type or continuous such as permeability are coded and processed aa a series of binary indicator variables. Stochastic modeling of the reservoir is baaed on Indicator Principal Component Kriging and the sequential simulation principle. The resulting alternative and equipru~able reservoir models honor the available information at wells and their statistics. These reservoir models are then used to investigate the impact of geological heterogeneities on flow performance prediction in a waterflood scheme. Analysis of various production parameters (cumulative oil production, cumulative water oil ratio and breakthrough times) indicates that they are more sensitive to the geological architecture of the reservoir than to details of the statistical distributions of petrophysical variables. Direct modeling of permeability across Iithofacies, i.e. ignorance of the geological architecture, may lead to severe inaccuracy in reservoir performance prediction, particularly after the first transient years of production.

This paper addresses the stochastic modeling of geological attributes, such as lithology or lithofacies; the generation of reservoir properties, such aa permeability, porosity and saturation; and the prediction of reservoir performance, such as cumulative oil production, cumulative water oil ratio and breakthrough times, The main goal is to evaluate how performance prediction is affected by either explicit modeling of geological heterogeneities or its ignorance, A waterflood scheme is considered with one producer and one injector. Lithofaciea and reservoir petrophysical properties along these two wells are considered aJ+conditioning data. Reservoir description between wells is obtained through stochastic simulation [1, 2], Such stochastic modeling honors the available information at the wells and allows reproduction of patterns of spatial variability. Some of the reservoir models contain explicit information about the lit hofacies distribution. Other models ignore such geological information with petrophysical properties being generated across lithofacies, The waterflood scheme is applied to each reservoiq model and the resulting production parameters are compared to evaluate the importance of accounting for geological information in the modeling process, The stochastic approach used allows quantifying the uncertainty associated to the lack of information between wells. For each production parameter that uncertainty is characterized by the cumulative distribution function (calf) of the various outcomes corresponding to the alternative re,mrvoir models. 399

References and figures at the end of the paper.

,.

Modeling

Geological Heterogeneities and

and its hnpact

on Flow Simulation

SPE 22695

Modeling Geological Flow Properties

Heterogeneities

Recently, considerable effort haa been put on stochastic modeling of reservoir parameters, suciI as permeability, porosity, lithofscies [2-7]. All algorithms attempt to preserve the spatial variability observed from the well data as reflected by Iiistograms and autocovariance functions. This section presents an integrated approach to the numerical mddeling of oil reservoirs. The algorithm of indicator principal component kriging (IPCK) [8] is used, first to simulate the reservoir geological architecture, then for the simulation of flow properties specific to each of the previously simuhkted lithofacies. The underlying assumption of such a two-steps .approaoh is that flow is largely controlled by the major geological structures. Geological Heterogeneities Consider that K exclusive geological categories are observed at the wells. At any location x, an indicator random variable is defined as: 1 if L(x)= k (1) { O otherwise with k being one of the K geological categories found in the reservoir and L(x) being the actual category observed at x, These K indicator variabhx define the vector,
1(x;

Conditional probability models of type ( 4) provide information about heterogeneities distribution between wells. They can be used to yiehJ images of the reservoir geometric architecture [2]. This is done by drawing. the category m lithof&iea prevailing at any unsampled I* cation from he corresponding probability distribution of type ( 4). F1OW Properties As done above for categorical variables, determination of conditional probability model of type ( 4) provides the likelihood of occurrence of a certain class of permeability or porosity value at any unsampled location. As before, a continuous property, say Z(x), can be coded into a series of indicators:
a

1 if Z(X) ~ .%k (5) { O otherwise with %kbeing anyone of K threshold values discretizing the range of variability of Z(x), The definition ( 5) yields a vector of indicator variables similar to ( 2):
Z(x;

%k)=

k) =

I(x) =

[1(X; ZI) , ssI(X;

ZK)]T

(6)

I(x) = [1(x; 1) . . . I(x; K)]T

(2)

Note that one and only one element of vector ( 2) is equal to 1 since the geological. categories are exclusive. Indicator auto(crces)covariances inferred from the corresponding indicator data characterize the, relative geometry of the K categories: C~(h; k, k) = E[l(x; k)l(x + h; k)] p~pkl (3)

However, as opposed to the indicator vector ( 2), the definition ( 5) entails a vector with a series of Osand 1s. The Sh@.? trLUIShiOn from O tO 1 (~(x; Zk) = (), 1(X; %k+l) = 1) indicates that Z(x) belongs to the interval (zk, %k+l]. A conditional distribution model is agtin provided by a linear combination of indicator data:

Prob*{Z(x)

~ zk!ll~(xa), a = 1,,., ,n}

Pkf + (7),

with pk being the proportion of category k in the reservoir and h, a separation vector. There are K* such indicator Thus, when K is greater than 3 auto(cross)covariances. that inference may become tedious. Indicator auto(croas)covariances can be used to estimate the likelihood of occurrence of a particular category at any unsampled location. Indeed, it can be shown [1] that a model of conditional distribution can be obtained by a weighted linear combination of indicator data. -\ Pro6{L(x) = klL(xa), a = 1,..s, n} = pkt +

with pp being the proportion of values Z(x) below the threshold value %kl. As before, the weights AOare obtained by solving the corresponding system of normal equations. Stochastic Imaging

Expressions ( 4) and ( 7) represent uncertainty models at each unsampled location. They provide quantitative information about how much is ( or is not) known about the attribute Z(x) or category L(x) at location x, The sequential simulation algorithm [1] provides a way to interpolate between sample locations using the uncertainty models of type ( 4) or ( 7). The sequential simulation algorithm is here briefly recalled:
q

The weights & are determined by solving a linear system of normal equations [1].

Define a random path over the domain to be simulated.

.400

At-each location x, determine the conditional probability ( 4) or ( 7) for all K indicator variables. Use aa conditioning data {n} the original sample data as well aa the previously simulated values within a predefine neighborhood. Draw from these conditional distributions model K simulated indicator values i (x; k) or i (x; Zk), k = 1,... , K. The upperscript s refers to a simulated value. Add these simulated values to the set of conditioning data. Loop over all locations of the simulation grid.

The linear combinations (4) or(7) square regressions of the Yk(x), i.e.,
n
a=l

are replaced by least

followed by the back transfo~m: I*(x) = AY* (X) with Prob* {Z(x) S .Zk I{n]} = Integrating Attributes Flow Properties
Z(X;

(11)

Zk) Geological

It can be shown that this algorithm allows honoring original data at their locations and reproducing global statistics such aa proportions p~ and indicator auto(cross)covariances. Indicator Principal Component Kriging

and

As mentioned before, the derivation of the conditional probabilities ( 4) or ( 7) requires the inference of K2 indicator auto(croas]covariances. IPCK allows shortcutting that inference by working with on a limited number of indicator prinqipal components (ipc) [2, 7, 8]. Consider the indicator covariance matrix for a given separation vector h: ~I(h) = Chv{I(x), I(x + h)} (8)

The integration of geological and petrophysical information is accomplished by making the simulation of flow properties conditional to a previously simulated geological image. The probability distribution for the petrophysical property Z(x) is made specfic to each category k, and would consider aa conditioning data only samples pertaining to that category k, i.e., Prob*{Z(x) S zk I{n) E k]

with I(x) being the indicator vector ( 2) or ( 6). Next consider the unique orthogonal decomposition of that covariance matrix [9]: X1(h) = ADAT (9)

with AT being a K x K orthogonal matrix, defining the vector of ipc: Y(x) = ATI(x) ., with Y(x) =[YI(X) . . . YK(X)]T By definition, the K principal components Yk(x) and Ykl(x + h) are orthogonal to each other, i.e., their covariance matrix at separation h is diagonal. In practice the separation vector h, at which the decomposition ( 9) is done, ia chosen small or zero. This choice ensures in most cases [2, 8, 10], that the principal component covariatice matrix is orthogonal, to an excellent degree of approximation, atall separation vectors h. This property allows reduction of the modeling effort from K2 indicator auto(croas)covariancea for I(x) to only K ipc autocovariances for Y(x). (lo)

Thus, inference of statistical parameters, histograms and autocovariances, is done separately for each geological category. A shortsighted decision would be to avoid such complex inference by pooling together all available data regardless of geological facies. The fact that inference is made easier is no justification for ignoring essential. characteristics in the reservoir, Ignorance of major geological heterogeneities in reservoir modeling may yield inaccurate performance forecast as shown hereafter. A Heterogeneous Oil Reservoir

A number of alternative reservoir models (2D sections) have been simulated using the sequential simulation algorithm and IPCK. The conditioning data is constituted by lithofacies, porosities and permeabilities sampled at two wells. Each reservoir image comprises 2650 cells, 50 in the horizontal and 53 in the vertical direction. Static Properties .. .

Six different Iithofacies are initially considered and later, lumped into four different flow units. Table 1 gives the volume proportion of each Iithofacies and its wisigned flow unit. Figure 1 illustrates one stochastic image of the reservoir. It shows a layer-type reservoir whith each layer corresponding to a different lithofacies, itself constituted by different Iithologies. Also, note that there is no perfect

401
.,

Modeling

Geological Heterogeneities

and its Impact lations define Figure ability

on Flow Simulation

SPE 22695

continuity between the two wells, Table 2 gives the parameters of the first five ipc autocovariances. The sixth ipc is a constant value [10]. Additional information about these ipc autocovariances is given in Appendix. Porosity and permeability have been modeled as lognormal distributions. Table 3 shows the corresponding statistics for each flow unit. Note the heterogeneous behavior between flow units, going from the highly permeable flow unit 1 to an almost impermeable barrier (flow unit 4). Regarding spatial variability, porosity within each flow unit is considered uncorrelated whereas absolute permeability is modeled with spatial correlation. For the simulation of the absolute permeability field, nine threshold values (zk ) corresponding to the deciles of each lognormal distribution were selected. Table 4 shows the parameters of the three ipc autocovariances retained for the evaluation of the conditional distribution of type ( 11), The other ipcs are considered uncorrelated. Also, it was assumed that the permeability spatial variability is the same for all flow units. However, the geometry of each lithofacies is different. Also, note that the correlation ranges of permeability are smaller than the correlation ranges of lithofacies geometry. A small anisotropy ratio of 2:3 is used for vertical to horizontal absolute permeabilities. Dynamic Properties Dynamic fluid properties are assumed constant within each of the four flow units. Fluid saturation values are based on capillary-gravity equilibrium, which is is achieved if the capillary pressure and relative permeability relations. are parametrized by absolute permeability [11]. The dimensionless capillary pressure function group known as Jfunction, is used to parametrize the imbibition capillary presure curve-a as function of the absolute permeability. Further, it is assumed that interracial tension and contact angle changes between flow units are negligible, The connate water saturation (SWi) and residual oil saturation (SOr) are based on the correlations of Kocberber and Collins [11], Table 5 shows the critical saturations used for each flow unit, These values assume a water wet rock system, thus it is expected that the residual oil saturation is not much tiected by the pore geometry as given by the correlation [11]. The capillary pressure curve of the second SPE comparison problem [12] is normalized and transformed to represent an imbibition system with SOr = 0.2. It is used as reference for the parametrization process. The computed capillary presure curves are shown in Figure 2. The relative permeability curves considered in this study depend on the phase end point values and an exponent [13] whose value depends on the rock type. The phase end point relative permeability values are based on the corre-

[14], while the experimental results [15] are used to the curvatures. Table 6 shows the end points and 3 presents the resulting water oil relative permecurves for the four flow units.

Flow Performance A waterflood exercise was designed to measure the impact of geological heterogeneities on flow performance. A twophsse, tw-dimensional black-oil simulator [16] is used to derive cumulative oil production, cumulative water oil ratio and breakthrough times under dfierent scenarios. The waterflood scheme consists of one injector and one producer, with water being injected at constant rate. The injection period was five years. Seven different cases are considered in this study, Some of the cases include an explicit modeling of the geological architecture w obtained from simulation of the lithofacies, Other cases consist of stochastic simulation of the absolute permeability across all lithofacies. For all cases, the same well data are used as conditioning data. The total number of alternative reservoir models considered is 50 for each case. Description of the seven cases follows:
q

Case A: The geological architecture of Figure 1 is considered identically for all 50 realizations, Each flow unit of each realization is informed with an uncorrelated permeability field with the correct histogram. Relative permeabilities are made specific to each flow unit. Case B: This case is in all points identical to case A, except for the permeability fields that now present specific autocorrelation within each flow unit. Case C: No lithofacies differentidion is considered. The spatially correlated permeability field was generated using the average (over all flow units) statistics (histograms and autocovariances) of c,ase B. The relative permeability curves are the same as in cases A and B. Therefore, even though no lithofacies is considered, there is a partial knowledge of flow unit types through using different relative permeabilities. Case D: Same as case C but now a single relative permeability curve is used. The phase end points are computed as a weighted average of the phase end points of the four flow units. The weights are the pr~ portions of the tiow units in the reservoir. The curvature exponent is the same considered in the original relative permeability curves. Figure 4 gives the capillary pr&sure and relative permeability curve used for this case. Case E: The same unique lithofacies geometry used in cases A and B is considered for all 50 realizations of this case, but the permeability field is modified

402

SP.E 22695

V. Suro-P&ez,

P. Ballin, K. Aziz and A. G. Journel

such that the correlation range within each flow unit is increased by about 2570.
q

Case F: There are now 50 different reservoir geometries, i.e. uncertainty about the geological architecture is accounted for. The absolute permeability field is generated using the parrimeters of case E. The relative permeabilities of Figure 3 were assigned to the corresponding flow unit. Case G: The same 50 realizations of Iithofacies geometry generated for case F are used again. The permeability fields are simulated anew using the statistics considered for case B.

All cases consider the same oil in place and total pore volume, This constraint ensures a fair comparison between the production parameters derived from each case. Discussion of Results

Tables 7 and 8 show cumulative oil production for the second and fourth years ; Tables 9 and 10 present cumulative water oil ratios for the same two production years. In all cases, each statistical parameter is derived from the corresponding 50 realizations. Case A, for both production parameters, presents the smaller spread or variance. This is a consequence of the spatially uncorrelated absolute permeability. There are no flow paths or flow barriers since tb.e absolute permeability field lacks any spatial pattern ~f correlation. The flow is entirely controlled by the geological architecture. If the mean (3?)or median (qO.s) are retained as estimators, there is no dramatic difference between this case A and the other cases at least for the first three years. For longer times (> 3 years), the production parameters begin to deviate from the other cases. Caaes B and E share the same geological architecture, differing only by the permeability correlation ranges. For case E the ranges are 25% larger than for B. This causes a reduction of the cumulative oil production and an increase of the cumulative water oil ratio, from case E to B. However, these differences are not significant, and reservoir management decisions based in either case would be equivalent. C&s C and D are very different. Case C presents the minimum cumulative oil production and the maximum cumulative water oil ratio. Its performance appears shifted with respect to the other cases. This is due to using the original relative permeabilities without association to the geometrical architecture. This observation is verified in cake D corresponding to totaI ignorance of geometry. Average values and medians are now closer to the other cases for times lesser than 3 years. Howeverj observe that the variance in case D is small ss compared to the other cases

and that, as time increases (> 3 years), the production parameters begin to deviate from the cases that have access to the geological architecture, This shows that characterization of reservoir geometry is an important factor for accurate forecast of reservoir performance. Finally, cases F and G show the highest variances because the geometry of the reservoir is now changing from one realization to another. There is no practical difference between these two cases, supporting the argument that in heterogeneous reservoirs flow performance is primarily controlled by the geometry of such heterogeneities. Recall that the 50 realizations of lithofacies geometry are the same for both cases F and G. Figures 5a to 5C illustrate this point with water saturation maps at the end of the first year. Each map is one realization of cases B (Figure 5a), C (Figure 5b) and D (Figure 5c). Note that for cases B and C water chain nelling follows the geological structure whereas in case D no such behavior is observed. Cases C and D show a more pronounced gravity effect: the water is coming down as time progresses. Case B shows a less important gravity effect due to the geometrical architecture, Note on Figure 1 below the region of high permeability at the top of the reservoir, there is a layer that acts as a flow barrier. This geometric feature is accounted for in case B and ignored, partly, in case C and totally in case D, Statistics of breakthrough times are illustrated in T& ble 11. The maximum average corresponds to case A which presents also, the minimum variance. Case C departs considerably from the other cases. Its median corresponds approximately to the minimum values of the CMW which use explicit knowledge of reservoir geometry. It is observed that larger variances are associated to cases where the reservoir geometry is accounted for. Uncertainty Models

The very reason for using stochastic reservoir models is to provide an appreciation of the spatial uncertainty left beyond the availabIe information. Uncertainty exists because of incomplete knowledge of geology and flow properties in the reservoir. A good uncertainty model needs to account for all important features of the reservoir. Modeling of geological heterogeneities or equivalently flow units does impact the resulting uncertainty measure, For example, consider the cumulative oil production for the fifth year, Figure 6 shows the cumulative distribution function for cases A, B, D and G. The maximum difference between minimum and mmcimum cumulative oil productions is obtained for the case G, which accounts for the uncertainty about lithofacies. Case A which freezes the geological architecture and considers an uncorrelated permeability field, shows the smaller variation. This figure indicates that excessive simplification of

403

Modeling

Geological Heterogeneities

and its Impact on Flow Simulation

SPE 22695

the reservoir modeling process yields uncertainty models which may underestimate severely the actual uncertainty, Conclusions This paper has presented the application of stochastic imaging for modeling heterogeneous oil reservoirs. Flow properties, such as absolute permeability and geological attributes such as lithofacies, are the variables considered in that modeling. It has been shown that the geological architecture of the reservoir plays an important role in fluid flow, hence in performance prediction. The modeling of petrophysical properties such as porosity/permeability should be made conditional to previously modeled geological architecture. XIIpretsence of a heterogeneous porous media, flow is mainly controlled by the high permeability contrast between some flow units. It appears that knowledge of sptial variability of flow properties within each flow unit is not critical. Models with different permeability correlation ranges but sharing the same geological architecture, essentially give the sa,me flow performance. Future research should consider systematic variation in the spatial variability of geometric characteristics and measure their impact on flow. This study has also shown that larger uncertainty results from the consideration of geometric architecture. A non-conservative underestimation of actual uncertainty in production forecast may result by ignoring either the influence of geometric architecture or that of the spatial correlation of the petrophysical variables. Acknowledgements This research has been supported by the Stanford Center for k.eservoir Forecasting and SUPRI-B, References [1] Journel, A, G., 1989, Fundamentals of Geostatistics in Five Lessens, American Geophysical Union Press, Washington DC, 40 p. [2] Suro-P&ez, V. and Journel, A. G., 1990, Stochastic simulation of lithofacies: an improved sequential indicator approach, in 2nd. European Conference on the Maihemaiics of Oil Recovery, eds. D, Gu&illot and 0. Guillon, Ed. Technip, p, 3-10, [3] Alabert, F. G. and Massonnant, P., 1990, Heterogeneity in a complex turbidic reservoir: stochastic modeling of facies and petrophysical variability. SPE paper No. 20604. [4] Galii, A., Gu6rillot, D., Ravenne, C. and Heresim Group, 1990, Combining geology; geostatistics and

multiphsse fluid flow for 3D reservoir studies, in 2nd, European Conference on the Mathematics of Oil Recovery, eds, D. Gu6rillot and O, Guillon, Ed. Technip, p. 11-20,
[5] Haldorsen H, H., Brand, P, J. and Macdonald, C. J,,

1988, Review of the stochastic nature of reservoirs, in Mathematics in Oil Production, eds. S. Edwards and P. R. King, Oxford Science Publications.
[6] Hoiberg,

J., Omre, H. and Tjelmeland, H., 1990, Large scale barriers in extensively drilled reservoirs, in 2nd. European Conference on the Mathematics of Oil Recovery, eds, D, Guc%illot and O. Guillon, Ed, Technip, p. 31-41.

[7] Journel,

A. G. and Alabert, F., 1988, Focusing on spatial connectivity of extreme-valued attribut~: stochastic indicator models of reservoir heterogeneities, SPE paper No, 18324. V. and Journel, A. G., 1991, Indicator Principal Component Kriging, Math. GeoL, V. 23, No. 5, in press,

[8] Suro-P&ez,

[9] Golub, G. and Van Loan, Ch., 1983, Matr~z Computations, The John Hopkins University Press, 476 p.

[10] Suro-P&ez, V., 1991, Stochastic simulation of categorical variables for reservoir description, submitted to Math. Geol,. [11] Kocberber, S, and Collins, R, E., 1990, Impact of reservoir heterogeneityy on initial distributions of hydrocarbons, SPE paper No. 20547. [12] Chappelear, J. E. and Nolen, J. S., Second comparative solution project: a three-phase coning study, SPE paper No. 10489. [13] Corey, A. T., Rathjens, C. H., Henderson, J. H. and Wyllie, M. R. J., 1956, Three-phase relative permeability, Trans. AIME, V, 207, p. 349-351. [14] Honarpour, M, M., Koederitz, L. F. and Harvey, A. H., 1982, Empirical equations for estimating twophase relative permeability in consolidated rock , Trans. AIME, V. 273, p. 2905. [15] Morgan, J. T. and Gordon, D. T., 1970, Influence of pore geometry on water oil relative permeability, JPT, p. 1199-1208, [16] Eclipse-100, Exploration Consultants Ltd., Highlands Farm, Oxfordshire (RG9-4PS), England.

404

.,

SPE 22695 Appendix

V. Suro-P&ez,

P. Ballin, K, Aziz and A. G. Journel

practical range a, defined by:

Five ipc autocovariances were considered for the simulation of the six lithofacies. Two nested exponential structures with a large geometrical anisotropy are used for all five ipc autocovariances. The model is:

Cyk

(h=, h.)

= c~hpal( 7h=

For the absolute permeability field three ipc autocovariances are considered. The general expression of such autocovariancea is:

+r~hz)+

with Cyh (hZ, h~) being the k:h ipc autocovariance, Cl the sill and aI the horizontal cor *elation range for the l:h structure; h= and h. are the coordinates in the horizontal and vertical direction. Exp~ is the exponential covariance with

with C(0) being the variance and COthe nugget effect. The spherical model Spha(.) is: , / i (
\ Spha(h) 3h = 1 Z;+

lh~ ~(z)

Table 1: Equivalence sponds to Jithofacies, in percent.

between Lit hofacies and Flow Units. LF correFU to flow units and VOL to the volume proportion LF 11 24 32 4 52 62 FU VOL 0.46 0.22 0.04 0.09 0.11 0.05

Table 2: Structural Parameters of the IPC Autocovariance Models. Units of C are variance units, a are map units and T units are dimensionless. IPC 1 2 3 4 5

c1
0.1 0.036 0.037 0.025 0,017

C2
0.15 0.05 0.04 0.027 0.02 1::0 11.25 62.5 50.0 25.0 5:0 162.5 12,5 16.25 16.25 4;;8 ;; 2.81 14.1 15.62 4.16 15.62 5.41 7.14 8.12

Table 3: Porosity and Permeability Statistics. ~i and u: are the mean or variance of either porosity (~) or permeability (k) in md. These parameters define a lognormal distribution.
?Tik

u:
360000.0 22500.0 225.0 1.0

m~

u;

1
2

3 4

300.0 50.0 5.0 0.5

0.2 0.15 0.10 0.05

0.01 0,0056 0.0025 0.00062

1
.
z ..-

Table 4: Structural Parameters of the IPC Autocovariance The variances have been standardized to 1. IPC 1 2 3 C(0) 1.0 1.0 1.0 CO 0.10 0.05 0.0 Cl 0.8 0,95 1.0

Models.

3~0 20.0 10.0

JO 4.0 4.0

Table 5: Phase FIJ 1 2 3

4

End Point

Saturations

Table 6: Phase FU 1 2 3 4

End Point 1< .W~a%.(frac.) 0.50 0.35 0.18 0.02

Relative

Permeabilities

SWi(,frac.) 0.20 0.30 0.45 0.60

%(frac.) 0.20 0.225 0.25 0,275

K,O ~o=,(frac,)

1.00, 0.90 0.78 0.65

Table 7: Cumulative Oil Production for the 2nd. Year. Case refers to the different situations studied in the text; z is the arithmetic average; S2 is the variance; qo.25, qo,5 and qo.T5 are the first quartile, median and third quartile respectively; min and max are the minimum and maximum values. The units are STB in all cases, except the variance which is in STB2

c
D E F

7314.26 6824.48 7101.49 7154.93 7261.16 7272.63

53980.39 14925.25 9288.49 96604.41 79047.86 81616,80

7172.93 6742.88 7030.68 6958.35 7087.71 7192.33

7348.41 6812.83 7081.62 7154.60 7323.21 7319.89

7487.10 6892.65 7164.51 7413.05 7468,24 7456.95

6553.86 6593.86 6940.01 6406.39 6266.91 6408.24

7914.66 7396.13 7530.79 7598.67 7599.93 7615.71

Table 8: Cumulative
Case A B s

Oil Production

for the 4nd, Year.

c
D

918t.72 9014.76 8331.09 8730.63 8873.20 8920.46 8919.72

1007.62 38757.57 18691.42 9751.76 65190,74 52771.23 56302.08

9%%0 8922.55 8238.97 8665.96 8653.91 8773.49 8776.09

9;;;:5 9033.88 8319.45 8724.81 8875.84 8940,20 8946.92

9;:t;9 9139.96 8401.62 8782.49 9076.87 9124.57 9102.43

91!;87 8312.82 8082.17 8528.31 8247,06 8281.72 8016.04

92~~4 9299.57 8944.26 9229.44 9243.257 9235.89 9278.93

.,

Table 9: Cumulative dimensionless

Water

Oil Ratio

for the 2nd.

Year.

Units are

c
D E

1
G

0.041 0.113 0.070 0.065 ().049 0.047

0.0011 0.0003 0.0002 0.0021 0,0017 0.0018

0.013 0.099 0.060 0.024 0.016 0.016

0.032 0.114 0.073 0.060 0.038 0.039

0.054 0.125 0.080 0.088 0.060 0.054

0.0003 0.029 0.0111 0.002 0.!)02 0,0001

0.157 0.150 0.094 0,172 0.209 0,182

Table 10: Cumulative

Water

Oil Ratio

for the 4th. Year.

c D
E F
G

0.670 0.804 0.722 0.696 0.687 0.687

0.0013 0.0008 0.0003 0.0023 0.0019 0.0021

0.643 0,785 0,711 0,655 0.647 0.651

0.663 0.805 0.723 0,690 0.681 0.681

0.679 0.820 0.731 0,733 0.710 0.708

0,619 0.682 0.631 0.628 0,630 0.622

0.807 0.858 0.762 0.822 0.814 0.873

Table 11: Breakthrough are days2

Case A B a

Times.

Units are days, except for S* for which

C
E F

64;70 607.87 460i35 577.24 590.17 583.69

26;.84 2515.83 1258.21 4643.61 4162.69 5301.83

6&5 565.75 428.87 520.12 547.50 548.50

6?+;;7 611.37 456.25 574.88 602.25 602.25

6?~;i0 642.75 474.5 629.63 642.75 642.75

6;2~5 492.75 401,50 428.88 419.75 410.62

6;3:;0 693.50 %02.25 702.63 693.50 711.75

vroduux

inlectw

1
q

50

Figure 1: one stochastic lithofacies simulation. There are six lihofaci~ that later are summarized into four flow units. 70
n

is 60 :
50 :

~ : ---------............... ; ii ~ ;

0.0

0.2

0.4

0.6

0.8

1.0

WATER SATURATION

(frac.)

e Figure2: Capillarypressure curvescorresponding to the fourflowunits

co&idered in this study.

1,0 L
0.8 ~

0.6 f 0.4

0.2

0.0 l

0.0

0.2 0.4 0.6 WATER SATURATION

0.8

(frac.)

1 ,, i

.1

Figure 3: Relative permeability curves for the four flow units.

.. . . .

408

,.

1.0
0.8

Case

- D

0.6

GrJ 0.4
al
4

0.2 0.0

i,nfi.,,t,

O.O ;ATER i&uRA#oN

(a)

1.0 (fiit~j

18 16 14 12 10 8 6 4 2 0

Case - D

O.O

0.2
WATER

0.4
(b)

0.6

0.8
(frac.) ----

l.O

SATURATION

. Figure 4: Dynamic properties used in case D. (a) Relative permeability. (b) Capillary pressure.

409

Sw Distribution
50

Sw Distribution

40 30
N

20
10

1 0.2

10

20

30

40

50

10

20 (U)

30

40

50

x
(a)

*
o

Sw Distribution
50

0.70 0.65 0.60

0.9 : 0.8 : o.? 0.6 R K 0.5 0.4

Cumulative 011 Production

, * # * 1 * 1 1 D # * 9

(5th.

Year)

40 30
N

0.55 0.50

20 10 0

0.45 0.40 0.35

0.3 0.2 0.1

10

20

30

40

50
z [STBI

x
(c)
q

. Figure 5: Sw distribution at the end of the lst. year. (a) case B. (b) case C. (c) caae D.

Figure 6: Uncertainty models for the cumulative oil production at t~e&end of 5th. year. () case A; (*) case E ( ) case D and (+) we ~ Ml G. . .1