A NUMERICAL SIMULATION APPROACH FOR EMBEDDED DISCRETE FRACTURE MODEL COUPLED GREEN ELEMENT METHOD BASED ON TWO SETS OF NODES

For the original embedded discrete fracture model (EDFM), the linear-distribution assumption is adopted in calculating the pressure distribution in matrix grids containing fracture elements, which leads to the lack of accuracy in solving unsteady interflux in the early stage of oil reservoir development. Therefore, this paper proposes a numerical simulation approach for EDFM coupled Green element method based on two sets of nodes. The main idea of the Green element method with two sets of nodes is to distinguish pressure nodes from flux nodes, in which one set of pressure nodes is set at the vertex of grids and another set of flux nodes is set at the edge-midpoint of grids. It not only meets the local material conservation and has second-order accuracy, but also can be applied to any grid type. In this paper, the Green element method based on two sets of nodes is coupled with EDFM, and a new scheme of mass transfer between matrix cell and fracture elements is derived by adopting the boundary integral form of the unsteady flow control equation, which replaces the linear distribution assumption to improve the simulation accuracy. In addition, the modified EDFM adapts to any form of matrix mesh generation, which extends the limitations of the original EDFM which is only suitable for rectangular matrix mesh and difficult to consider complex reservoir boundaries. The research shows that the proposed model has high accuracy in the early stage and it is verified by the LGR module of commercial software tNavigator^® and the original EDFM. Taking the SRV-zoning model considering fracture networks and complex reservoir boundaries as an example, the flexibility of the proposed model for solving complicated problems is demonstrated by comparing the business simulation software named SFEM-COMSOL. This study can be used for the accurate simulation of dynamic production performance in fractured oil reservoirs.