A new computational model for flow in karst-carbonates containing solution-collapse breccias

Abstract

We develop a new three-scale (micro/meso/macro) computational model based on a reiterated homogenization procedure to describe flow in carbonate rocks containing complex geological structures, such as fractures and solution-collapse breccias in the sense of Loucks (AAPG Bull. 83(11), 1795–1834, 1999). In this setting, we construct a hierarchical karst-fracture model wherein the larger geological objects are incorporated explicitly whereas the higher density microscopic structures are homogenized and replaced by equivalent continua with properties computed from self-consistent homogenization schemes. In the upscaling method, we subdivide the different clastic arrangements in the breccia into crackle, mosaic, and chaotic substructures where equivalent permeability and elastic constants are assigned to each layer within the breccia. After reconstructing the mesoscopic coefficients, we adopt a flow-based upscaling to the macroscale, where the characteristic length is associated with a typical coarse grid cell of a reservoir simulator. The mesoscopic flow equations are constructed based on the discrete fracture model (DFM) and discretized by a robust computationally scheme with the ability to handle strong heterogeneity induced by the collapse breccia and pressure jumps across flow barriers. In addition to the scenario wherein the collapse breccia network is composed of disconnected objects (isolated chambers), we also develop a reduced model for the case of connected karst facies playing the role of a network of enlarged fractures. Numerical results, with input data extracted from outcrops drone images, are presented illustrating the influence of different settings on flow patterns and their effect upon the magnitude of macroscopic properties.