Efficient Multiscale Numerical Framework of the Subsurface Reservoir Simulation in Heterogeneous Porous Media

Faculty Development Competitive Research Grant Program 2019-2021

In reservoir simulation of subsurface flow and transport, numerical simulations of entire reservoir in fine scale are computationally excessive. Accurate solution of subsurface model involves a vast amount of computer memory and running time. In order to reproduce fine scale features of subsurface model in heterogeneous porous media multiscale approaches have been established. We are using recent development for the flow and transport problem, an adaptive numerical homogenization method, which has been developed in [1,2]. The numerical results for the SPE 10 problems were showed a good agreement with fine scale simulation at production well. The adaptivity criteria for flow should be selected in an intelligent way, that is why we will conduct a posterior error analysis that provides good indicators of elements in mesh refinement strategy. We plan to investigate the error indicators for mesh refinement in the flow problems and then apply for numerical simulation with realistic applications.