3-D Grid Types in Geomodeling and Simulation - How the Choice of the Model Container Determines Modeling Results
Requirements for subsurface modeling have changed substantially over the past years: higher oil prices have made exploration and development of more complex reservoirs feasible; more data has become available for predicting and monitoring performance that now need to be integrated in subsurface models. Targets for (re)development have become more sophisticated and depend more critically on accurate models of geometry and actual properties. This paper attempts to analyze the requirements of ‘static’ modeling at reservoir to basin scales and simulation of dynamic subsurface behavior, covering fluid flow as well as geomechanical response to man-made changes in the subsurface.
Until recently pillar gridding has been seen as the undisputed standard for 3D reservoir modeling. However, when mapping its features onto the evolving requirements it becomes clear that pillar or corner point grids are adequate only for a limited fraction of the total task portfolio within oil and gas producers. Several alternative 3Dgridding methods have emerged lately - among which the faulted s-grid of Jewel Suite - and show significantly better match with critical conditions: For example, accurate full-field modeling is only feasible using grid types that are insensitive to fault directions. Additionally, all gravity-driven processes can only be simulated accurately in grid types supporting vertical cell stacks. Integration of different sources of information typically means dealing with measurements at different scales, which often have a range of 5 orders of magnitude. A consistent property model can, under these conditions, only be built with 3Dgrid types that that support scale-independent geometry handling. Linking and integration of both reservoir simulation and geomechanical simulations is only properly possible if faults are not represented in a voxelized or stair-stepped fashion.
With such variegated and sometimes conflicting requirements it is no surprise that there is no single grid type that optimally addresses all cases. It is hoped that the information provided here makes it more feasible for operators to analyze current and future requirements of their specific asset portfolio and by applying appropriate weights select a 3Dgrid type that optimally supports decisions on subsurface assets for the next 5 - 10 years.
AAPG Search and Discovery Article #90090©2009 AAPG Annual Convention and Exhibition, Denver, Colorado, June 7-10, 2009