Integrated Clustering/Geostatistical/Evolutionary Strategies Approach for 3D Reservoir Characterization and Assisted History-Matching in a Complex Carbonate Reservoir, SACROC Unit, Permian Basin

Abstract

An integrated methodology combining clustering analysis techniques, geostatistical methods and evolutionary strategy technologies was developed and applied to an area in the SACROC Unit (Permian basin). Clustering methods were applied to well logs and core data with high vertical resolution for many wells to predict porosity, permeability and rock type. Geostatistics was applied to extend the characterization into the inter-well area. Evolutionary strategies were used to refine the characterization to match historical production performance. The complete approach was tested on an area within the SACROC Unit, acknowledged as a highly heterogeneous carbonate reservoir with complex production history. Three cored wells provided porosity and permeability measurements on a foot-by-foot basis. These measurements coupled with well logs were used to predict porosity, permeability and flow units. Twenty two wells in the study area having foot-by-foot profiles of porosity and permeability were considered sufficient to characterize porosity and permeability in three dimensions. Geostatistical methods were then used to build porosity and permeability models. As a validation of the characterization procedure, evolutionary strategy jointly coupled with a black oil reservoir model was used to history match production performance of a 0.5 mi2 area. The 65,340 grid-block model had over 50 years of production. Thirteen (13) input parameters were varied during the history match. Among them, a multiplying factor was applied to the permeability realization to account for upscaling effects, varying permeability values without modifying geological heterogeneities identified during the characterization process. No adjustment to porosity characterization was permitted. A very good history match of individual production was achieved for the center wells of the area, and a good match was also obtained for outer wells production and reservoir pressure where boundary effects existed. This validates the new integrated clustering/geostatistical/evolutionary-strategy approach in this highly heterogeneous carbonate reservoir. Introduction An integrated methodology combining clustering analysis techniques, geostatistical methods and evolutionary strategy technologies was developed and applied to a study area in the SACROC Unit (Permian basin). Initially, a two-step "soft-computing" procedure was developed capable of efficiently generating core-scale porosity and permeability profiles at well locations where no core data existed. The approach applies clustering methods based on maximum likelihood principles to well logs and core data for lithology interpretation, reservoir quality characterization, and prediction of "core" parameter profiles, with high vertical resolution for many wells. This procedure permites to populate any well location with core-scale estimates of porosity and permeability (P&P), and rock types facilitating direct application of geostatistical techniques to build 3D reservoir models. Geostatistical methods are then applied to the resulting dataset, and three-dimensional spatial models of variability for clusters, porosity, and permeability are utilized to generate reservoir representations of P&P for flow simulation purposes. Finally, a computer assisted history matching based on application of evolutionary strategy technologies was used to history match the production performance of a selected subregion in the SACROC Unit (Permian basin).

A Method to Integrate Geological Knowledge in Variogram Modeling of Facies: A Case Study of a Fluvial Deltaic Reservoir

“…Geostatistical studies use hybrid simulation approaches that combine two or more conditional simulation techniques [17]. It involves first creating the structural design of the reservoir then the geological properties; beginning with facies then petrophysiscal properties [17]. In the structural design of the reservoir model, faults were first modeled and a grid network to accommodate the data was generated using the fault model.…”

Geostatistics application in spatial analysis of geomechanical properties

“…Geostatistics offers a way of characterizing the spatial continuity of natural phenomena by analyzing them as random variables [1]. -Geostatistics can describe data distribution in various spatial directions.…”

“…Variogram is a mathematical function, basic tool to quantify correlation of spatial variables [1], defined as:…”

“…CoKriging: Cokriging is used to estimate one variable value with co-variable. Two common examples are the estimation of permeability using porosity data and the estimation of porosity data using seismic data [1]. Estimation equation is:…”

Field Testing and Numerical Simulation of Combined CO2 Enhanced Oil Recovery and Storage in the SACROC Unit

“…Cross-section model with homogeneous properties (Brummett et al, 1976,Dicharry et al, 1973,Graue and Blevins, 1978,Langston et al, 1988. Recently, a simulation for history matching has been done with a black-oil model (Gonzalez et al, 2008). Also Weon Shik Han (Han et al, 2010) did simulation studies on the SACROC with the CMG's compositional simulator to evaluate CO 2 sequestration in two artificial models: a reservoir saturated with brine and a reservoir saturated with both brine and oil.…”