Improved upscaling for flow simulation of tight gas reservoir models
Conference Paper
Overview
Identity
Additional Document Info
View All
Overview
abstract
Tight gas and shale gas reservoirs provide almost half of the current U.S. domestic gas production, with significant projected increases in the next several decades. These reservoirs constitute an important play type, with opportunities for improved reservoir management in the optimization of depleted volumes as functions of well spacing and fractured well design. Reservoir simulation, together with detailed 3D geologic models, may be used for improved management, but only if the flow simulation through these models can capture the essential heterogeneity and 3D continuity of these intermittently connected rock packages. The current work will examine tight gas reservoirs, but we recognize that similar issues will arise in source rock plays. The current approach has three key technical elements: (1) We explicitly preserve the local continuity of the reservoir sands in the geologic model through the design of the simulation grid. Each simulation cell is a 1x1xN amalgamation of the geologic scale corner point cells, where "N" will vary depending upon the local sand thickness. (2) We have developed a "Well Index" based upscaling, to determine the upscaled permeability and which preserves the local reservoir quality. It does not require knowledge of the well locations within the upscaled simulation model, but it will preserve the performance of these wells. (3) The heterogeneity within each sand is preserved through the use of transmissibility upscaling, which we show performs systematically better than the usual permeability upscaling. In combination, these three elements provide simulation results which are almost indistinguishable from the fine scale model. We also examine additional approximations in which we further coarsen the model areally, and reduce the simulation time further. We develop and demonstrate our calculations on a sector model taken from the center of a full field onshore U.S. tight gas reservoir. The results are then validated using the full field model. Our approach differs from earlier studies which have attempted to preserve the global heterogeneity of the reservoir models though layer-based statistical calculations. Although these statistical approaches are superior to uniformly coarsened models, they are not as robust or as accurate as the current work constrained by local continuity. Our approach also differs in the use of accurate property upscaling techniques that simultaneously preserve the internal contrast of permeability within each sand, and the performance of wells within the model. Copyright 2011, Society of Petroleum Engineers.
