Upscaling Technique for High Resolution Coal Bed Methane Reservoir Model

Abstract

Upscaling of a reservoir model is normally conducted when a high resolution model has been generated and it is practically impossible to run the simulation using a high resolution model. Reservoir model upscaling is done in order to have a reasonably coarse model for reservoir history matching and forecasting purposes without losing some of heterogeneity in the reservoir model. A high resolution static coal reservoir model (around 0.1 metre thickness) is generated based on the detailed single coal ply correlation in order to capture complexity and heterogeneity of the coal sediment. A case study is presented in this paper. Several upscaling methods for reservoir properties such as permeability and gas content were tested and analyzed. The upgridding process on the geocellular model mainly focused on vertical direction in order to identify which ply could be merged into the sub layer level. The next step identifies whether the ply can be merged with neighboring ply individually. The results of the upscaling and upgridding are analyzed using the combination of recovery ratio, simulation running time ratio and cell count ratio plot. Histogram and area map analysis is also used to evaluate the result of the upscaling work. The expected results of the upscaled model are lower running time, lower cell count and very close recovery value to the high resolution model, while still capturing the heterogeneity. The results of the upscaling exercise for this case are reduction of running time by 72%, reduction of the cell count by 64% and recovery difference of 3%. Some of the plies in one of the layers cannot be merged into a sub layer level which indicates a high degree of heterogeneity. This upscaling technique for the coal reservoir model provides more information about the reservoir characterization methodology for unconventional reservoirs, and coal bed methane reservoirs specifically.