MPS-Driven Digital Rock Modeling and Upscaling

This paper starts with an overview of the application of multiple-point statistics (MPS) to characterizing and modeling heterogeneous rocks with the aid of digital rock technology. An upscaling workflow from microscopic pore space to inter-well scales is proposed and discussed by illustrative examples. In this MPS-driven upscaling workflow, training images, which play a key role in MPS modeling, are obtained directly from microscopic rock or borehole images or digital outcrops to capture more realistic heterogeneity of reservoir rocks and formations. Unlike the conventional upscaling method that typically applies a certain averaging scheme to coarsening a fine scale model, which tends to lose heterogeneous features that may be critical to fluid flow, the proposed workflow calls for both top-down and bottom-up approaches. The top-down approach aims to divide heterogeneous rock into different representative element volumes (REV) such that each one is more uniform or less heterogeneous; the bottom-up approach attempts to derive effective transport properties by carrying over the computed dynamic properties through running flow simulation from fine to coarser scales with the help of REV identification. This workflow allows us to couple the static with dynamic modeling to derive scale-dependent effective transport properties while capturing more realistic rock heterogeneity using MPS.