Dynamic Embedded Discrete Fracture Multi-Continuum Model for the Simulation of Fractured Shale Reservoirs

Abstract

Abstract The wide implementation of hydraulic fracturing in shale reservoirs and the interactions among hydraulic fractures, natural fractures and the shale matrix have brought great challenge for reservoir studies using conventional finite-difference-based reservoir simulators. This work establishes a dynamic approach for the simulation of fractured shale reservoirs that incorporate the embedded discrete fracture model (EDFM) as well as the multiple porosity model, which is necessary for simulating the complex transportation process in shale reservoirs. In this paper, we extend the EDFM approach for fractured shale reservoirs to a multi-continuum context, in which one fracture segment can have mass transfer with multiple parent grid blocks in favor of the complex porosity types of shale formations. The reservoir model can be updated dynamically during the simulation with our in-house simulator to consider the change of the fracture network due to fracking and refracking at different stages of field development. This is achieved by altering the effective non-neighbor connections with time that are associated with the fracture system. Therefore the reservoir with a changing fracture distribution can be modeled with a single simulation. The formulation is based on finite volume rather than finite difference to facilitate the unstructured nature of the reservoir model. The result is compared with the explicit fracture model with PEBI grid blocks to evaluate its accuracy. For the test case of a horizontal well with multiple hydraulic fractures and large scale natural fractures, the simulation result can be matched with great accuracy. Sensitivity analysis to grid refinement is also conducted, which proves the model only need moderate grid refinement to obtain desirable accuracy. This work established a more flexible and efficient yet more accurate approach for fractured shale reservoir modeling, with the emphasis on improving the ability to model the fluid transportation among different porosity types. The proposed model improves the simulation by reducing the complexity of the gridding process, cutting the total number of grid blocks, and significantly decreasing the CPU time. It provides a coherent method for characterizing the fluid transportation in fractured shale reservoirs, which is usually a difficult task with traditional reservoir models.