Coupling Upscaled Discrete Fracture Matrix and Apparent Permeability Modelling in DFNWORKS for Shale Reservoir Simulation

Abstract

Modelling non-Darcy flow behaviour in shale rocks, composed of nanometer-sized pores and multi-scale fracture networks, is crucial for various subsurface energy applications. However, incorporating multiple physical mechanisms across numerous scales is not trivial. This work proposes an improved and practical upscaling workflow for coupling an Upscaled Discrete Fracture Matrix (UDFM) model and a pressure-dependent apparent permeability (K_app) model to capture the effects of non-Darcy flow in multi-scale fractured shale reservoirs.

First, a 3D DFN is upscaled into octree-refined continuum meshes, where equivalent rock parameters and rock-fluid functions are defined using the UDFM approach. Then, the flow simulation is coupled with a pressure-dependent K_app updating scheme using an existing K_app model and a multiple-restart technique. The effects of non-Darcy flow mechanisms (e.g., slip flow, transitional flow, Knudsen diffusion) are captured. The constructed models are then used to study the impacts of fracture network connectivity and pressure interference on production. The results of this new approach are compared against those obtained from another commercial package while preserving the advantages of DFNWORKS. Neglecting non-Darcy flow behaviours could significantly underestimate gas production and water recovery. It is illustrated that the nanoscale flow mechanisms help to enhance matrix-matrix and matrix-fracture flow. The constructed models are also utilized to study the effects of disconnected or isolated fractures, pressure interference, water retention, and shut-in durations on well performance. The proposed flexible strategies can be adopted in other commercial/open-source fractured-porous-media subsurface-flow simulation frameworks.