Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
net/publication/254511527
CITATIONS READS
6 58
4 authors, including:
Daniel Mikes
Amsterdam University College
37 PUBLICATIONS 178 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Can one infer sea level curve in time domain from shoreline trajectory in space domain. View project
All content following this page was uploaded by Daniel Mikes on 15 August 2015.
several approaches to obtain two phase upscaling algorithms within the laminae via simple arithmetic averaging. Laminae
suitable for black oil simulation8,24,36,49 . All of them require are assumed homogeneous and isotropic.
dividing the reservoir in appropriate sub-volumes, i.e. the size The main purpose of this work twofold: 1) To improve the
of grid cells and finding averaged phase permeabilities and translation of a geological model to a reservoir model; 2) To
capillary pressures. Problems occur when correlation lengths provide a simple procedure to upscale heterogeneities of a
are comparable to the grid block size. The effects of reservoir. Using a well-defined geometrical facies model,
neighboring grid blocks need then to be taken into account5. where all sedimentological units are represented, will attain
There is a vast literature on upscaling algorithms and we the translation. This is accomplished by using the five-scale
mention those that are frequently employed: hierarchy (see Fig. 1). Next comes the simplification of this
One class of methods seeks to find approximate solutions model to a reservoir model, which can serve as the base of a
of the pressure equations, given more or less appropriate flow simulation.
boundary conditions. Examples are: (1) Renormalization We compare our approach with the SPE ninth comparative
methods 31 (see also refs 14,15,16,17,19,20). (2) The effective study 29. This allows easy duplication of the results. This is
medium approximation10. (3) Homogenisation. A second class why we have modified the original SPE model as little as
of methods uses pseudofunctions6,35,1,18,23,28,37 We use possible. There are three important points we like to
simulation on a smaller scale to obtain the constitutive emphasize. (1) The way of incorporating small-scale
relations 11,12, which constitutes a third class of procedures. Set heterogeneities in a reservoir model is significant for reservoir
up in the x, y, z direction with no flow boundary conditions performance. 2) The upscaling procedure for averaged
along the sides, p=1 at the inlet, p=0 at the outlet. Solve the constitutive relations must deal with capillary trapping in
equations and sum the fluxes. Then the effective directional small-scale lamina. (3) The procedure must be easy to
relative permeabilities can be calculated. Alternatively implement in the current infrastructure of petroleum
periodic boundary conditions 22 lead to the full tensor effective companies and institutes.
permeability30,45 (see also refs48,26 ). This is significantly more
accurate, but it has the disadvantage that it cannot be entered Methods:
in the commercial simulator”. This third class of upscaling Geological model
methods has the additional advantage that it is easy to To demonstrate the procedure we examine a highly
implement. heterogeneous reservoir deposited in a fluvial environment.
Integrated procedures, involving the construction of a Fluvial reservoirs are known to be heterogeneous on a wide
simplified genetically based reservoir model and the range of scales53. For our purposes we generate a model in five
subsequent derivation of averaged constitutive relations have steps using five different scales (Table 1).
been employed before. In many cases, however, the diversity The largest scale comprises the entire reservoir and
of various types of crossbeds is ignored in the constitutive consists of highly permeable sandy sediments deposited in or
relations. A single capillary pressure- and relative permeability near the river channel. These sand bodies are encased in shale,
curve is used for all crossbeds 3,34,41,43,46,55 (see also refs13,33,50 ). which was deposited on the floodplain bordering the river
Ringrose 46,47 however, presents an elegant method to include channels (Fig. 2A).
the effects of small-scale sedimentary structures in the scaling One scale smaller we divide the different elements into
up of heterogeneity. Kjønsvik et al.32 and Panda et al.40 present subfacies: the encasing shale becomes the floodplain subfacies
sophisticated procedures to include all levels of heterogeneity and the sand body is split into a pointbar subfacies and a
in a structure. However, they either do not use the constitutive channel fill subfacies (Fig. 2B). A typical pointbar consists of
relations calculated on the individual laminae, or the method is a series of concentric layers, which dip down toward the outer
too complex to be used routinely. bend of the river. Sometimes these layers are separated by thin
Aigner et al.3 suggest a methodology to systematically shale streaks that extend from the top to halfway the bottom.
divide outcrops into flow units based on the "genetically The channel fill may contain sand in which case there is
driven hierarchy" of heterogeneities, using the fact, that units abundant trough crossbedding, or it may consist entirely of
are repetitive and that heterogeneities occur on several scales. shale.
Pickup et al.43 present a systematic ‘upscaling box’ in which Based on bedding type we may go to an even smaller
everything from characterization via upscaling to production scale, the bedding scale (Fig. 2C). Here we distinguish three
calculation is incorporated (see also refs44,45 ). For the bed-scale types: horizontal bedding, which is found exclusively in the
they use a number of typical bed-configurations. Compared to floodplain subfacies; longitudinal bedding which is found in
Aigner3 and Pickup al.43 we emphasize more the detail on the the pointbars; and through crossbedding for the (sandy)
crossbed scale regarding both description and measurement. channel fill.
We add to the methodologies described above: (1) Adequate Finally, at the smallest scale the heterogeneity is expressed
description and simplification of the repetitive laminated in the dominant grain size of the laminae that consitute the
facies in terms of flow cells; (2) Measurements of laminae beds: horizontal bottomset laminae are fine or medium
permeability with an optimized laboratory probe-permeameter grained, and the inclined foreset laminae are medium or coarse
set up25,52,54 . Absolute permeabilities of laminae are averaged grained.
SPE 68702 UPSCALING OF FLOW UNITS FOR RESERVOIR FLOW INCORPORATING SMALL-SCALE HETEROGENEITIES 3
Reservoir model Swc=0.0, the sorting factor λ=1.5, the porosity ϕ=0.3 and the
The reservoir model is a direct translation of the geological interfacial tension between oil and water σow =0.03 [N/m]. The
model. Thus, the large-scale model consists of a single layer end point permeabilities for water and oil are taken k’ rw= 1 and
(Fig. 3A). Within this layer the floodplain, pointbar, and k’ro= 1. The permeabilities are found from probe-permeameter
channel fill flow units are placed (Fig. 3B). The flow units are measurements and are summarized in Table (3). For now we
rectangular versions of their subfacies counterparts. Each flow did not use image analysis data to estimate the sorting factor
unit consists of several flow-cells, the equivalent of the and instead used an average value λ=1.5 for both foresets and
bedding type in the geological model (Fig. 3C). We make a bottom sets. In this way we completely assigned the relative
large simplification here: flow cells have the same orientation permeability and capillary pressure fields for each of the three
within each flow unit. Thus, the horizontal bedding in the lamina types we used in our bedding models (Fig. 3). The
floodplain is always horizontal, the longitudinal bedding in the aspect that the trapping is determined by the wetting state is
pointbars and the trough crossbedding in the channel fill outside the scope of this article27,46 and we confine interest to
always dip in the X-direction. As a result each of the flow water wet media39.
units has the same properties throughout the model.
Dimensions for the flow cells were inferred from the The second step consists of performing a simulation (see
geological model and are listed in Table 2. At a smaller scale, Table 4) on a part of the flow unit i.e. the flow cell (Fig. 4),
each flow cell is made up of alternating laminae, to which which is manageable by a simulator. Thus there is one flow
different permeabilities are attributed. Typical lamina cell for every subfacies, i.e. crossbedding, longitudinal
permeability values were compiled from literature and from bedding, and horizontal bedding for the channel, pointbar, and
our own data (Table 3). These values were simplified to the floodplain, respectively (Fig. 2 & 3). For a relatively small
extent that we only have three permeabilities: 10 mD, 50 mD, section we may assume that during reservoir production it will
and 200 mD. The floodplain has horizontal bedding with an go through a slowly varying sequence of quasi steady states.
alternation of 10 mD and 50 mD. The channel fill has This assumption is also implicit in other upscaling methods
crossbedding (alternation of 50 and 200 mD laminae) or it is e.g. homogenisation. Hence, following suggestions made by
shale filled (10 mD). The pointbar has three superimposed many previous authors we will put no flow boundaries at the
regions of equal thickness: the lower part consists of cross top and side of the reservoir and inject oil and water (at
bedding only (alternation of 50 and 200 mD); the middle part various ratios) on one side and produce at the other. We prefer
contains cross bedding with occasional clay drapes (alternation the use of no flow boundaries as opposed to periodic boundary
of 50 and 200 mD and the clay drapes 10 mD). The upper part conditions because the latter give rise to anisotropy that cannot
is horizontal bedding with an alternation of 10 mD and 50 mD. be diagonalized in the (x,y,z) system of coordinates of the
simulation. Most commercially available simulators cannot
Microsimulation handle non-diagonalized anisotropy. The initial oil and water
Here we describe the procedure of obtaining upscaled saturation in the flow cell are determined by assuming a given
phase permeability curves and capillary pressures for the three capillary pressure, Pc =107 kPa. We continue flooding until a
bed types constituting the reservoir. These upscaled steady state is reached.
permeabilities are derived in a three-step process.
The third step is the computation of the constitutive
First we assign the Brooks-Corey permeabilities for the relations from the simulation results. We use Darcy’s law in
high and low permeable foresets and at the bottom sets which we substitute the flow of oil and water together with the
constituting the facies. average phase pressure difference between the injection and
2 production side to derive with Darcy’s law the phase
+3
S − S wc λ 2 permeabilities. Also the average oil minus the average water
k w = kk 'rw w := kk ' rw S we λ +3 pressure is the capillary pressure. Finally we calculate the
1 − S wc average saturation in the flow cell. In this way we obtain one
point in the phase permeability-saturation curve and capillary
pressure-saturation curve. We continue to use the same
2
+1
k o = kk 'ro (1 − S we ) 1 − S we
2 λ
procedure with water oil injection mixtures ranging from zero
to 100% oil, with steps of 10%. In this way the upscaled phase
permeabilities and capillary pressures for the channel beds, the
1 point bar beds and the flood plain beds are obtained. We
ϕ 2 − S wc λ
1 repeat the same procedure but now we only allow flow in the
Pc = γσ ow vertical direction.
k Sw − S wc In the simulation considered we have three-phase oil-
water-gas flow. Hence three-phase permeabilities are required.
For convenience we use the same values as Van Lingen; Here we use STONE II. Admittedly the use of STONE’s
the empirical constant γ=0.79, the connate water saturation models directly for upscaled relative permeabilities is
4 D. MIKES, O.H.M. BARZANDJI, J. BRUINING, C.R. GEEL SPE 68702
questionable. A much better approach would be to perform saturation; (3) Anisotropy of the relative permeabilities
three phase (steady state) simulations, instead of the two-phase enhancing the effective permeability anisotropy.
simulations described above. The results would, however, be
history dependent and hence three phase upscaling is Macrosimulation
considered outside the scope of this article. The numerical simulation is run on our meandering river
These results are not suitable yet for implementation into a model, using the model of the ninth SPE Comparative
standard simulation run, because the constitutive relations are Solution Project (SPE 29110)29 as basis. This project concerns
anisotropic. Hence we make one further simplification. We a black oil simulation of moderate size (9000 cells) and with a
calculate the average ratio between the relative permeabilities high degree of heterogeneity provided by a geostatistically-
in the vertical and the horizontal direction. We also calculate based permeability field. Here we mention the salient features
the one-phase permeability anisotropy from one-phase of the simulation in order to make clear where we differ from
simulations. The product of the average ratio with the one- this base case. Further details in particular of the fluid
phase anisotropy factor determines the effective permeability properties can be found in SPE 29110.
anisotropy. We use the constitutive relations obtained from the The field contains 25 somewhat randomly spaced
horizontal flow simulation. producers and a single water injector. The dipping reservoir is
divided in 24x25x15 grid blocks in rectangular coordinates
Results for upscaled permeabilities without local refinement. The dimensions of the grid blocks
The results from the microsimulation are summarized in are 300 ft both in the x-direction and in the y-direction. Cell
Fig. 6 for horizontal flow and in Fig. 7 for vertical flow. In the (1,1,1) is at a depth of 9000 ft subsea. The remaining cells dip
left column of Fig. 6 we show the upscaled relative in the x-direction with no dip in the y-direction. In the z-
permeabilities and capillary pressures for oil and water. The direction we have 15 layers, each represented by the height of
relative permeability for oil and gas are the mirror image of the grid block. The height of each of the layers is as given in
the water/oil relative permeabilities and capillary pressures 7. SPE 29110, the average porosity of all the layers is 0.1262.
However, for the runs with the genetically based reservoir
• Channel beds: With respect to the input lamina relative model we use a constant porosity (ϕ =0.3). The added total
permeabilities we observe a substantial increase of the height of the reservoir is 359 feet. All producers were
residual oil saturation both for horizontal and vertical completed over a length of 56 feet in layers 2,3,4 only and the
flow. The higher horizontal relative water permeability at injector was completed over a length of 207 feet in layers
low water saturations must be attributed to the fact that at 11,12,13,14, and 15. The water well is located at the (24,25)-
a given average water saturation of say Sw=0.4, the low corner grid. We use the same constraints on the producer and
permeable lamina are almost completely filled with water injection wells as used in the 9th comparative programme. Also
and hence are able to contribute to the permeability. The the productivity indexes are based on well bore radius of 0.5 ft
vertical relative permeabilities are substantially lower, and a drainage radius of one fifth of the grid block length.
showing the additional anisotropy effect originating from Relative permeabilities and capillary pressures for the runs
the lamina structure. with the genetically based reservoir models are used as derived
• Point bar beds: Here the residual oil saturation appears to above. There is no free gas initially in the reservoir and the oil
be less affected by the cross-bed structure, showing that is at the saturation pressure. PVT properties of the fluids are
we cannot use a single set of relative permeabilities for all taken as in SPE 29110. The oil viscosity is around 1 cP, and
the facies. Again the vertical relative permeabilities are the gas viscosity around 0.015 cP and viscosities are slightly
substantially lower, showing the additional anisotropy pressure dependent. We calculate the initial water saturation
effect originating from the lamina structure. distribution by assuming capillary pressure gravity equilibrium
• Flood plain beds: Here we again observe the typical and given the oil-water contact being at 9950 feet subsea.
horizontal relative permeability enhancement, but this
time for oil. Still the residual oil saturation is higher than Results for the reservoir simulation
for the laminae individually. We note the difference in We make a comparison of runs for six different conditions
horizontal relative permeabilities with respect to the (Table 4 & 5). First runs are made of the SPE 9th comparative
channel and point bar. Again the vertical relative program, using the stochastically generated reservoir model
permeabilities are substantially lower than the horizontal. and all the input data of SPE 29110, whence relative
permeabilities and capillary pressures are shown in Fig. 8.
In summary we observe how much the relative However, we like to compare the situation with two effective
permeability behavior is affected by the way that the laminae permeability anisotropy ratios. First we use a ratio of the
are distributed in the channel, point bar and flood plain. The horizontal and vertical permeability ratio of 100 (indicated by
most important features are: (1) Horizontal relative Hspe). Secondly we use an effective permeability ratio of 5.6
permeability enhancement, due to a more favorable (indicated as Vspe). The other four cases concern the
distribution of the saturations; (2) Increase in residual oil genetically based reservoir model developed in this paper. We
use the channel belt along strike with effective permeability
SPE 68702 UPSCALING OF FLOW UNITS FOR RESERVOIR FLOW INCORPORATING SMALL-SCALE HETEROGENEITIES 5
anisotropy ratios of 100 (indicated by N1), and 5.6 (N1v). The kr = relative permeability
last two runs concern a model with the channel belt along dip, S = saturation, fraction
with again anisotropy ratios of 100 (N3) and 5.6 (N3v). P c = capillary pressure, m/Lt2, Pa
Fig. 9 presents the results of the six runs. Given the s = interfacial tension, m/t2, N/m
preliminary nature of our research we are only able to describe g = empirical constant
the most conspicuous features, without drawing definite l = sorting factor
conclusions. Shown are the cumulative oil, the cumulative gas f = porosity, fraction
and cumulative water productions. Admittedly the results of
Subscripts
our reservoir model (N1, N1v, N3, N3v) with porosities ϕ=0.3
o = oil
are difficult to compare with Vspe and Hspe with an average
w = water
porosity ϕ=0.13 (see Table 5). We note that the different wc = connate water
anisotropy factors have a large effect on the cumulative we = normalised water (saturation)
productions from the stochastically generated reservoirs.
For the genetically based reservoir model, there appears to Acknowledgements
be a slightly more favorable (lower) water production and less
favorable oil recovery (lower) from reservoirs where the We thank B.T. de Haas for reading the manuscript.
anisotropy factor is highest. This is to be expected as the
injection well is injecting water in the bottom part of the
reservoir, and oil is produced near the top. Whether the References
channel belt is along strike or along dip has little effect. This 1. Aasum, Y., E. Kasap, and M. Kelkar, 1995, Analytical
can possibly be attributed to a more or less symmetric Upscaling of Small-Scale Permeability Using a Full Tensor,
situation with respect to the injection well and apparently a Petroleum Geoscience, Vol. 1, pp.365-377.
small effect of dip in this particular situation. 2. Abbaszadeh, M., H. Fuji, F. Fujimoto, 1995, Permeability
For the conditions shown the gas oil ratio for the Prediction by Hydraulic Flow Units - Theory and Applications,
genetically based reservoir model is still approximately equal SPE 30158.
to the solution gas oil ratio indicating that we are still before 3. Aigner, T., J. Heinz, J. Hornung, and U. Asprion, 1999, A
Hierarchical Process Approach to Reservoir Heterogeneity:
free gas breakthrough.
Examples From Outcrop Analogues, Elf exploration production,
Our most important result is, however, that we have shown 11 pp.
that genetically based reservoir models can be simplified in a 4. Alabert, F.G., and V. Modot, 1992, Stochastic Models of
way that makes these models amenable to upscaling and Reservoir Heterogeneity: Impact on Connectivity and Average
reservoir simulation. Our results present the basic concept of a Permeabilities.
procedure, which can be applied using more sophisticated 5. Almeida, J. A., A. Soares, M.J. Periera, and T.S. Deltaban,
aspects in future work. 1996, Upscaling of Permeability : Implementation of a
Conditional Approach to Improve the Performance in Flow
Conclusions Simulation, SPE 35490, European 3-D Reservoir Modeling
Conference (Stavanger Norway, April 16-17 April.
• A five scale hierarchy with crossbed and subfacies as key
6. Barker, J.W., and S. Thibeau, 1997, A Critical Review of the
elements i.e. pore scale → lamina scale → flow cell scale Use of Pseudorelative Permeabilities for Upscaling, SPE
→ flow unit scale → reservoir scale can be used to Reservoir Engineering, May, SPE 35491.
construct a reservoir heterogeneity model, directly based 7. Berry, J.F., A.J.H. Little, R.C. Skinner, 1992, Differences in
on geological insight. Gas/Oil and Gas/Water Relative Permeability, SPE/DOE Eighth
• It is possible to simplify the reservoir model such that it Symposium on Enhanced Oil Recovery , April 22-24.
becomes amenable for upscaling and reservoir simulation 8. Botton-Dumay, R., Y.M. Cogrel, Massonat, G.J., N. Eberle,
1997, Realistic Methodology for Permeabilitity Modelling Used
using state of the art technology.
for Conserving Heterogeneity during Assisted History
• The lamina/cross bed structure introduces additional Matching-Applied to Turbiditic reservoir Field Case., SPE
anisotropy via the upscaled constitutive relations. 38677, Annual Technical Conference and Exhibition held in San
• Preliminary simulations show the importance of the total Antonio, Texas, USA, October 5-8.
anisotropy i.e. the product of the one-phase permeability 9. Bruining, J., Batenburg, D.W. van, Lake, L.W. and An Ping
and the additional anisotropy originating from two-phase Yang, “Flexible Spectral Methods for the Generation of Random
flow effects. Fields with Power-Law Semivariograms”, Mathematical
Geology, Vol 29, No 6 (1997).
• Further research is necessary to quantify the effect of the
10. Choy, T.C., 1999, Effective Medium Theory – Principles and
distribution of facies and other parameters used in the Applications, Clarendon Press Oxford, ISBN 0 19 851892 7.
model. 11. Christie, M.A., Upscaling for Reservoir Simulation, J. Pet.
Techn. November 1996, SPE 37324.
12. Christie, M.A., P.J. Clifford, 1997, A Fast Procedure for
Nomenclature Upscaling in Compostional Simulation, SPE 37986, Reservoir
k = permeability, L2, m2 Simulation Symposium, Dallas, Texas, June 8-11.
6 D. MIKES, O.H.M. BARZANDJI, J. BRUINING, C.R. GEEL SPE 68702
13. Ciammetti, G., P. S Ringrose, T.R, Good, J.M.L. Lewis, and K. 30. King, M.J., and M. Mansfield, 1997, Flow Simulation of
S. Sorbie, 1995, Waterflood Recovery and Fluid Flow Upscaling Geologic Models, SPE 38877, Annual Technical Conference
in a Shallow Marine and Fluvial Sandstone Sequence, SPE and Exhibition held in San Antonio, Texas, USA, October 5-8.
30783, Annual Technical Conference and Exhibition held in 31. King, P.R., 1989, The Use of Renormalization for Calculating
Dallas, Texas, USA, October 22-25. Effective Permeability, Transport in Porous Media 4, pp. 37-58.
14. Corbett, P.W.M., and J.L. Jensen, 1991, Variation of Reservoir 32. Kjønsvik, D. J. Doyle, T. Jacobsen, and A. Jones, 1994, The
Statistics According to Sample Spacing and Measurement Type Effects of Sedimentary Heterogeneities on Production from a
for Some Intervals in the Lower Brent Group, The Log Analyst, Shallow Marine Reservoir - What Really Matters?, SPE 28445,
January-February, pp.22-41. pp.27-40.
15. Corbett, P.W.M., and J.L. Jensen, 1991, Flow Performance in 33. Kortekaas, T.F.M., 1985, Water/Oil Displacement
Rannoch Facies, in: Minipermeametry in Reservoir studies, Characteristics in Crossbedded Reservoir Zones, SPE,
1991, Dalmahay Hotel, Kirknewton, Edinburgh, 20 p. December, pp.917-926.
16. Corbett, P.W.M., and J.L. Jensen, 1992, Estimating the Mean 34. Kossack, C.A., S.T. Opdal, 1990, Scaling Up Heterogeneities
Permeability: How Many Measurements Do You Need, First with Pseudofunctions, SPE Formation Evaluation, September,
Break, Vol. 10, No. 3, March, pp.89-94. pp.226-232.
17. Corbett, P.W.M., S.G. Stromberg, P.J. Brenchley, and G. 35. Kyte, J.R., and D.W. Berry, 1974, New Pseudofunctions to
Geehan, 1994, Laminaset Geometries in Fine Grained Shallow Control Numerical Dispersion, SPE 5105, SPE - AIME 45th
Marine Sequences: Core Data from the Rannoch Formation Annual Fall Meeting, Houston, Oct 6-9 (1974), (see also SPE-
(North Sea) and Outcrop Data from the Kennilworth Member Reprint Series No. 20, “Numerical Simulation II”, (1986), SPE.
(Utah, USA) and the Bencliff Grit (dorset, UK), Sedimentology, 36. Lemouzy, P.M., R.K. Romeu, anf I.F. Morelon, 1993, A New
pp.729-745. Scaling-Up to Compute Relative Permeability and Capillary
18. Davies, 1987, Pseudofunctions in Formations Containing Pressure for Simulation in Heterogeneous Reservoirs, 68 th
Discontinuous Shales: A Numerical Study, SPE 16012. Annual Technical Conference and Exhibition of the SPE,
19. Deutman, R., 1992, The Impact of Permeability Up-Scaling on Houston, Texas, October 3-6, SPE 26660.
Effective Permeability and Recovery Predictions, Delft 37. Lingen, P.P. van, 1998, Quantification and reduction of capillary
University of Technology, Msc. thesis. entrapment in cross-laminated oil reservoirs, Delft University of
20. Dodge, C.F., D.P. Holler, R.L. Meyer, 1971, Reservoir Technology, PhD thesis, 214 pp.
Heterogeneities of Some Cretaceous Sandtones, The AAPG 38. Martinius, A.W., and R.A. Nieuwenhuijs, 1995, Geological
Bulletin, V. 55, No. 10, October, pp.1815-1828. description of flow units in channel sandstones in a fluvial
21. Dubrule, O., M. Thibaut, P. Lamy, and A. Haas, 1998, reservoir analogue (Loranca Basin, Spain), Petroleum
Geostatistical Reservoir Characterization Constrained by 3D Geoscience, Vol. 1, pp. 237-252.
Seismic Data, Petroleum Geoscience, Vol. 4, pp.121-128. 39. Nouwens, N.M., 1997, Computer simulation of capillary
22. Durlofsky, L.J., 1991, Numerical Calculation of Equivalent Grid entrapment in crossbedded reservoirs, M.Sc. thesis, Delft
Block Permeability Tensors for Heterogeneous Porous Media, University of Technology.
Water Resources Research Vol 27 (5) 699-708. 40. Panda, M.N., C. Mosher, and A.K. Chopra, 1997, Reservoir
23. Fischbusch, D.B., and R.A. Wattenbarger, 1991, Waterflood modeling using scale dependent data, SPE 38745, 18 pp.
Simulation of Stratified Sandstone Reservoirs Using Dynamic 41. Peihua, X., 1986, A point bar facies reservoir model - semi-
Pseudo Curves, SPE, Vol. 5, pp.219-235. communicated sandbody, SPE 14837, pp.103-115.
24. Giliberti, S., M. Erba, M. Rampoldi, and S. Said, 1995, 42. Peters, E. J., The Effect of Instability on Relative Permeability
Optimization of the Numerical Simulation in a Giant Oil Field Curves Obtained by The Dynamic Displacement Method., SPE
by Upscaling of Relative Permeability Curves, International 14713.
Meeting on Petroleum Engineering, Beijing, PR China, 43. Pickup, G.E., P.S. Ringrose, M.M. Forrester, J.L. Jensen, and
November 14-17. K.S. Sorbie, 1994, The geopseudo atlas: geologically based
25. Halvorsen, C., 1991, Probe Permeametry Applied to a Highly upscaling of multiphase flow, SPE 27565, pp.277-289.
Laminated Sanstone Reservoir, in: Minipermeametry in 44. Pickup, G.E., P.S. Ringrose, P.W.M. Corbett, J.L. Jensen, and
Reservoir Studies, Dalmahay Hotel, Kirknewton, Edinburgh, 10 K.S. Sorbie, 1995, Geology, geometry, and effective flow,
p. Petroleum Geoscience, Vol. 1, p. 37-42.
26. Honarpour, M.M., A.S. Cullick, and N. Saad, 1994, Influence of 45. Pickup, G.E., and S. Sorbie, 1996, The Scaleup of Two phase
Small-Scale Rock Laminations on the Core Plug Oil/Water Flow in Porous Media Using Phase Permeability Tensors, SPE
Relative Permeability and Capillary Pressure, University of Journal, December, SPE 28586
Tulsa Centennial Petroleum Engineering Symposium, August 46. Ringrose, P.S., K.S. Sorbie, P.W.M. Corbett, and J.L. Jensen,
29-31, Oklahoma, SPE 27968. 1996, Immiscible flow behaviour in laminated and crossbedded
27. Huang, Y., P.S. Ringrose, K.S. Sorbie, 1996, The Effects of sandstone SPE Formation Evaluation, September 1996, p. 171-
Heterogeneity and Wettability on Oil Recovery from Laminated 176.
Sedimentary Structures, SPE 30781, SPE Journal, December. 47. Ringrose et al., 1996, Use of geology in the interpretation of
28. Kasap, E., and L.W. Lake, 1990, Dynamic Effective Relative core-scale relative permeability data, SPE Formation Evaluation,
Permeabilities for Crossbedded Flow Units, SPE 20179, pp.71- September, pp.171-176.
82. 48. Saad, N., A.S. Cullick, and M.M. Honarpour, Effective Relative
29. Killough, J.E., Ninth SPE Comparative Solution Project: A Permeability in Scale-up and Simulation, SPE Rocky Mountain
reexamination of Black-Oil Simulation, 13th SPE Symposium Regional/Low Permeability Reservoirs Symposium, Denver,
San Antonio (TX), Feb 12-15 (1995) (SPE 29110) March 20-22 (1995), SPE 29592.
49. Tchelepi, H.A., L.J. Durlofsky, W.H. Chen, A. Bernath and
M.C.H. Chien, 1997, Practical Use of Scale Up and Parallel
SPE 68702 UPSCALING OF FLOW UNITS FOR RESERVOIR FLOW INCORPORATING SMALL-SCALE HETEROGENEITIES 7
Reservoir Simulation Technologies in Field Studies, SPE 38677, 53. Weber, K.J., 1986, How heterogeneity affects oil recovery, in:
Annual Technical Conference and Exhibition held in San Reservoir characterization, L.W. Lake and H.B. Carroll, Jr.
Antonio, Texas, USA, October 5-8. (eds.), Academic Press, Orlando, Florida, pp.487-544.
50. Van Vark, W. van, A.H.M. Paardekam, J.F. Brint, J.B. van 54. Wiefkers, H.L.A., 1995, Minipermeametry; millimetre scale
Lieshout, P.M. George, 1992, The Construction and Validation characterisation of crossbedded sandstone for the modelling of
of a Numerical Model of a Reservoir Consisting of Meandering capillary effects in oil displacement by water, M.Sc. Thesis,
Channels, European Petroleum Conference Cannes November Delft University of Technology, August 1995.
16-18. 55. Yinan, Q., 1984, Depositional model, heterogeneous
51. Volpi, B., G. Tannoia, G.B. Donna, and R. Bouggherara, 1995, characteristics and waterflood performance of sandstone
A complete methodology for reservoir characterization through reservoirs in a lake basin, case study of oilfields, Eastern China,
a 3D high-resolution grid, Petroleum Geoscience, Vol. 1, Reservoir geology and engineering, 13 pp.
pp.319-326. .
52. Waal, W.W. van de, D. Mikeš, J. Bruining, 1999, Inertia factor
measurements from pressure-decay curves obtained with probe-
permeameters, In Situ, volume 22 (4) 339-371.
8 D. MIKES, O.H.M. BARZANDJI, J. BRUINING, C.R. GEEL SPE 68702
Table 1 − Size scales with all respective unit names and examples
used in our procedure. Shaded areas are the key units.
Stratigraphic Units Reservoir Units Geological Units Facies Example Bed levels Scale Range
Formation Reservoir Facies River m-hm
Member Flow Unit Subfacies Pointbar Bedset dm-m
Bed Flow Cell Bed Trough bed Bed Cm-dm
Lamina Lamina Lamina Foreset Lamina Mm-cm
Grain Pore Pore Coarse sand Grain µm-mm
Table 5 - Initial values and production results for the three simulation scenarios.
Hspe9 Vspe9 N1 N1v N3 N3v
6
Total Pore volume 10 STB 453 1324 1324
6
Initial Oil in place 10 STB 186 234 261
6
Mobile Oil volume 10 STB 156 82 92
9
Total solution Gas volume 10 SCF 259 325 327
9
Free Gas volume 10 SCF 0 0 0
6
Water volume 10 STB 246 1061 1063
6
Cumulative Oil produced 10 STB 17.5 25.2 25.9 27.1 25.5 27.2
9
Cumulative Gas produced 10 SCF 71.1 101 29.2 31.2 28.4 30.9
6
Cumulative Water produced 10 STB 2.1 1.6 18.2 23.7 18.9 24.6
6
Cumulative Water Injected 10 STB 1.5 2.5 2.6 3.9 2.7 4.0
SPE 68702 UPSCALING OF FLOW UNITS FOR RESERVOIR FLOW INCORPORATING SMALL-SCALE HETEROGENEITIES 9
RESERVOIR CHARACTERISATION
CONSTRUCTION
MODEL
2
PARAMETER
Pc b
f
c
Kb Kf Kc
3
NUMERICAL SIMULATION
Hor. Vert.
MICRO
K rw K rw
K ro K ro
Flow-Cell model
4
SIMULATION
Production
MACRO
o
w
g
Reservoir model
Fig. 1 − The 4-step procedure. Step 1: Geological description; Step 2: Parameter assignment; Step 3: Microsimulation; Step 4:
Macrosimulation.
10 D. MIKES, O.H.M. BARZANDJI, J. BRUINING, C.R. GEEL SPE 68702
Floodplain
C) Bedding
B) Flow Units
Floodplain
C) Flow Cells
Channel
Pointbar
Floodplain
Fig. 4 − The three flow-cell models for Channel, Point bar, and Floodplain, resp.
SPE 68702 UPSCALING OF FLOW UNITS FOR RESERVOIR FLOW INCORPORATING SMALL-SCALE HETEROGENEITIES 13
1.0 120
0.8 100
0.6 80
60
0.4
40
0.2 20
0.0 0
0 0.2 0.4 0.6 0.8 1
Sw
1.0 120
0.8 100
0.6 80
60
0.4 40
0.2 20
0.0 0
0 0.2 0.4 0.6 0.8 1
Sw
Bottom Set
1.0 120
0.9
0.8 100
0.7 80
0.6
0.5 60
0.4
0.3 40
0.2 20
0.1
0.0 0
0 0.2 0.4 0.6 0.8 1
Sw
Fig. 5 − The capillary-pressure- and relative-permeability curves for the laminae of the three flow-cells.
14 D. MIKES, O.H.M. BARZANDJI, J. BRUINING, C.R. GEEL SPE 68702
Channel Channel
1 120 1 120
0.2 0.2 20
20
0 0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.5 1
Sw Sgt
100 100
0.8 0.8
80 80
0.6 0.6
60
60
0.4
40 0.4
40
0.2 20 0.2
20
0 0
0 0
0 0.2 0.4 0.6 0.8 1
0 0.2 0.4 0.6 0.8 1
Sw
Sgt
1 120 1 120
80 80
0.6 0.6
60 60
0.4 0.4
40 40
0.2 0.2 20
20
0 0 0 0
Sw Sgt
Fig. 6 − Capillary-pressure- and relative-permeability curves for the entire flow-cells for horizontal flow. The oil-gas rel-perms and capillary
pressures are the mirror image of the water-oil rel perms.
SPE 68702 UPSCALING OF FLOW UNITS FOR RESERVOIR FLOW INCORPORATING SMALL-SCALE HETEROGENEITIES 15
0.8 100
80
0.6
60
0.4
40
0.2 20
0 0
0 0.2 0.4 0.6 0.8 1
Sw
0.8 100
80
0.6
60
0.4
40
0.2 20
0 0
0 0.2 0.4 0.6 0.8 1
Sw
100
0.8
80
0.6
60
0.4
40
0.2 20
0 0
0 0.2 0.4 0.6 0.8 1
Sw
Fig. 7 − The capillary-pressure- and relative-permeability curves for the entire flow-cells for vertical flow.
16 D. MIKES, O.H.M. BARZANDJI, J. BRUINING, C.R. GEEL SPE 68702
1.2 450
400
1
350
0.8 300
250
0.6
200
0.4
150
0.2 100
50
0
0
0 0.2 0.4 0.6 0.8 1
-0.2 -50
S w
1.2 4.5
4
1
3.5
0.8
3
0.6 2.5
0.4 2
1.5
0.2
1
0
0.5
0 0.2 0.4 0.6 0.8 1
-0.2 0
S g
Fig. 8 − The capillary-pressure- and relative-permeability curves for the SPE model.
SPE 68702 UPSCALING OF FLOW UNITS FOR RESERVOIR FLOW INCORPORATING SMALL-SCALE HETEROGENEITIES 17
Channel
Pointbar
Floodplain
Fig. 9 − Along strike and along dip flow simulation models for IMEX.
18 D. MIKES, O.H.M. BARZANDJI, J. BRUINING, C.R. GEEL SPE 68702
30
Cumulative Oil
25 Hspe
20
MMBbls
Vspe
15
N1
10
N1v
5
N3
0
N3v
0 300 600 900
Time (days)
25
Cumulative Water
20 Hspe
MMBls
15 Vspe
10 N1
N1v
5
N3
0
N3v
0 300 600 900
Time (days)
Cumulative Gas 10^9
120
90 Hspe
Vspe
SFC
60
N1
30
N1v
0
N3
0 300 600 900
N3v
Time (days)