Reservoir Modeling Methods and Characterization Parameters for a Shoreface Reservoir: What Is Important for Fluid-Flow Performance?
- F.X. Jian (ChevronTexaco E&P Technology Co.) | D.K. Larue (ChevronTexaco E&P Technology Co.) | A. Castellini (ChevronTexaco E&P Technology Co.) | J. Toldi (ChevronTexaco E&P Technology Co.)
- Document ID
- Society of Petroleum Engineers
- SPE Reservoir Evaluation & Engineering
- Publication Date
- April 2004
- Document Type
- Journal Paper
- 89 - 104
- 2004. Society of Petroleum Engineers
- 5.5.8 History Matching, 2.4.3 Sand/Solids Control, 5.2.1 Phase Behavior and PVT Measurements, 4.3.4 Scale, 5.3.4 Reduction of Residual Oil Saturation, 5.1.1 Exploration, Development, Structural Geology, 5.4.1 Waterflooding, 4.1.2 Separation and Treating, 7.2.2 Risk Management Systems, 1.7.5 Well Control, 1.2.3 Rock properties, 5.5 Reservoir Simulation, 5.1 Reservoir Characterisation, 5.5.3 Scaling Methods, 4.1.5 Processing Equipment, 5.6.1 Open hole/cased hole log analysis, 5.1.5 Geologic Modeling, 5.1.2 Faults and Fracture Characterisation
- 1 in the last 30 days
- 654 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 5.00|
|SPE Non-Member Price:||USD 35.00|
Fifty reservoir models were built by using seven different modeling methods and varying a combination of reservoir characterization parameters using experimental design. These 50 models were all subjected to unconstrained fluid-flow simulation with an active aquifer and limited peripheral waterflooding without history matching. The results show that oil recovery from existing wells at 95% water cut ranges from 29 to 34% in these 50 models despite the large apparent visual differences between them. However, there are large differences (some greater than 200%) in water-breakthrough time, water cut, and cumulative water production. The differences in oil recoveries and water cut among all models were found to be largely related to horizontal permeability heterogeneities perpendicular to the shoreline direction.
Experimental design indicated that there were statistically significant differences in flow-simulation results associated with the definition of reservoir quality trends, 3D facies trends, and variogram range and direction. Do more complex, geologically realistic models provide the best frameworks for predicting flow performance in sandy shoreface reservoirs with good well control? Some of the simplest models we constructed provided flow-simulation results similar to those of the most complex and soft-data-conditioned models. What we do observe, however, is that the more soft conditioning data and the more complex geologic models tend to give a restricted range in full-field flow performance.
Investment decisions for field development are commonly based on future field performance predicted from static and dynamic reservoir models. The predictive capabilities of the reservoir models usually dictate the decision quality for the investment. Challenges to reservoir characterization are (a) identifying what types of reservoir heterogeneities are most relevant to fluid flows so that the right types and amounts of data can be acquired and (b) determining how to build robust reservoir models with limited subsurface information, typically in a short time frame. Overall, in the oil industry and as described later, there is a feeling that more soft conditioning data and more complex modeling procedures make reservoir models have better predictive capabilities.
Previous work has identified parameters influencing waterflood performance for different types of reservoirs. For a low net-to-gross fluvial reservoir, net-to-gross, stacking pattern of channel belts, number of channels within a belt, and channel width and thickness were found to have the most significant effect.1 Small-scale heterogeneities within cross stratification also were found to be significant. 1 In a high net-to-gross fluvial reservoir, permeability contrast was found to be important.2 When permeability contrast is low, detailed reservoir modeling is redundant. However, it becomes necessary if the permeability contrast is large because of waterflood channeling. Larue and Friedmann3 stressed the importance of permeability heterogeneity on waterflood recovery for channelized reservoirs of both high and low net-to-gross. They stressed that because connectivity can be achieved at relatively low net-to-gross values (35%), the channelized geologic models behave like tanks. Still, there was a notable change in recovery (~2 to 5% recovery) between low and high net-to-gross channelized reservoirs. In a shallow marine reservoir, the thickness and stacking patterns of parasequences were found to be the most important.4 Incorporating high-resolution sequence stratigraphic architecture and reservoir quality trends within each parasequence of a reservoir simulation model was found to be significant for improved history match and identifying bypassed oil in a shoreface reservoir.5,6 For submarine channelized reservoirs, net-to-gross, permeability heterogeneity, mean permeability, mobility ratio, and residual oil saturation were found to be significant factors influencing waterflood performance.7 For a tidal reservoir, the sizes of tidal bars and effective permeability of heterolithic facies were found to have the largest impact on single-phase fluid flow.8 Experimental design was commonly used by these studies1,4,7,8 to determine the most efficient number of model runs and to identify the significant modeling parameters. Landa and Strebelle9 used derivative coefficients to identify significant parameters for fluid-flow performance.
Currently, a wide range of reservoir modeling techniques is available. In the studies described previously, Boolean modeling,1,3,7,8 sequential indicator simulation,2 and truncated Gaussian simulation with trend4 were used to model facies distribution. Sequential Gaussian simulation (SGS) with both trend and collocated cokriging were used for petrophysical modelings.5,6,9 Constructing reservoir characterizations also involves choosing modeling parameters for each technique, including variogram type, range, and orientation; vertical and lateral trends; and so on. The objective of this paper is to investigate the impact of some commonly applied modeling methods on fluid-flow performance for a typical shoreface reservoir in west Africa. This study used the same data and stratigraphic framework as those used by Cook et al.5 and Larue and Legarre.6
Parasequences, generally conformable successions of bedsets bounded above and below by flooding surfaces, represent the building blocks of shoreface reservoirs. Groups of parasequences, or parasequence sets, define the larger-scale shoreface reservoir architecture. Two fundamental types of parasequence sets are progradational and retrogradational.10 Because of the characteristic sheet-like geometries, shoreface reservoirs are often described as layer-cake reservoirs.11 Permeability contrast between bedsets and parasequences, vertical and lateral trends in porosity and permeability, and the continuity of mudstones is particularly important for waterdrive and waterflood performance in layer-cake reservoirs. We hypothesize that only those modeling methods and parameters that are able to represent these characteristics can generate robust models of shoreface reservoirs. However, as described later, this is probably not so easily proved.
|File Size||31 MB||Number of Pages||16|