AN INTEGRATED RESERVOIR CHARACTERIZATION, GEOSTATISTICAL ANALYSIS, OPTIMIZED HISTORY-MATCHING AND PERFORMANCE FORECASTING STUDY OF THE 9-SECTION, 30-WELL PUMP CANYON CO2-ECBM/SEQUESTRATION DEMONSTRATION SITE, SAN JUAN BASIN, NEW MEXICO

摘要

The Pump Canyon CO2-ECBM sequestration demonstration site, located in the northern New Mexico portion of the San Juan basin, is part of the Southwestern Regional Partnership for Carbon Sequestration sponsored by the United States Department of Energy. The objective of the project is to field-test the efficacy of CO2 sequestration in deep, unmineable coalseams. In support of the demonstration, an integrated reservoir characterization, geostatistical analysis, optimized history-matching and performance forecasting study was performed. Based on available reservoir data such as coal depth and thickness, isotherms, initial pressure and cleat orientation, a three-layer, nine-section, 30-well reservoir simulation model was built. Production data revealed a high degree of reservoir heterogeneity. In order to account for this heterogeneity a geostatiscal approach based on a sequential Gaussian simulation algorithm was used to generate a permeability distribution. A loose correlation between permeability and porosity was applied. An automated history-matching procedure, an iterative technique utilizing a global optimizer and a fractured reservoir simulator, was utilized to match the production data in an unbiased, time-efficient manner. A total of thirteen input parameters, a combination of formation, relative permeability and well properties, were varied during the optimization process. A total of 125 simulator iterations requiring 8-days of unsupervised computing time was required to achieve the match. A performance forecast under CO2-injection was then performed.