A Control Volume Material Balance Approach and Its Applications to Real-Time Flow Diagnostics

Abstract

Abstract In this paper, we propose a Control Volume Material Balance (CVMB) approach for proxy reservoir simulation and apply it to real-time flow diagnostics. Instead of utilizing a comprehensive reservoir simulator, it estimates the saturations distributions by mapping the mass difference between injected and produced fluids recorded at wells into 3D grid blocks. On this basis, we perform real-time flow diagnostics to evaluate the dynamic heterogeneity of the instantaneous displacement flow field which can be used for making effective and opportune decisions to improve oil recovery. CVMB solves the pressure and flow fields implicitly, and the transport equations explicitly. It incorporates 3D heterogenous rock properties. The fundamental idea of the CVMB method is to divide the 3D flow field into a series of 1D well-pair Control Volumes (CVs). A well-pair Control Volume is composed of grid blocks in the intersection of the sweep and drainage regions of the injector and producer. The fluid flow in and out of the 1D CV can only occur at the wells, and the in-situ fluid volumes are determined by the well flow rates and the well allocation factors. In each CV, we assume the displacement in the grid blocks is piston-like and follows the 1D order of ascending forward time-of-flight. The fluid saturation distributions are determined by defining the cut-off time-of-flight for the displacement front. We show how the CVMB method improves the pattern-based mass balance approaches in the following aspects: 1) enables real-time flow diagnostics in terms of the hydrocarbon dynamic Lorenz coefficient without a comprehensive reservoir simulator; 2) enhances the simplicity and extensibility of the pattern-based mass balance approach without mapping between grid blocks and streamlines; 3) reduces the smearing effects in conventional mass balance approach by defining 1D CVs using forward time-of-flight. The proposed CVMB method utilizes the historical well flow rates as the input to estimate the swept regions and its average saturation with remarkable efficiency and sufficient accuracy for real-time flow diagnostics.