A Statistical Upscaling Workflow for Warm Solvent Injection Processes for Heterogeneous Heavy Oil Reservoirs

Abstract

Hybrid solvent- and thermal-based methods offer an important potential for reducing the environmental impact and enhancing recoveries in heavy oil production processes. Time-consuming compositional simulations of fine-scale models challenge the optimization of operational conditions. This work developed a flow-based upscaling workflow for simulating the warm solvent injection process in heterogeneous reservoirs. This process takes into account the pattern of spatial heterogeneity at different scales. A set of coarse-scale models was considered to be constructed to capture the subscale variability due to coarsening. In this study, scale-up of several static (porosity, permeability) and dynamic or transport parameters (dispersivities) was considered. In order to depict the spatial heterogeneity below the modeling scale, several realizations of sub-scale models of the modeling-scale cell size were created. The difference in the flow simulation responses of a heterogeneous realization and an equivalent average model was minimized to estimate effective longitudinal and transverse dispersivities. Results from all realizations were aggregated to construct the probability distributions of effective dispersivities. Field data of typical Athabasca bitumen reservoirs were used to generate a set of synthetic fine-scale models to test the workflow. The method is flexible, as no explicit assumption regarding the multivariate distribution of the heterogeneity is required. This paper extends our previous work, which focuses on solute transport in single-phase flow, to adopting this statistical scale-up framework for a complex hybrid solvent-thermal process.