Using Embedded Discrete Fracture Model (EDFM) in numerical simulation of complex hydraulic fracture networks calibrated by microseismic monitoring data

Hydraulic stimulation of unconventional shale and tight reservoirs often creates a complex induced fracture network, which requires a comprehensive characterization for successful exploitation and development. One of the major technologies applied over the past decade to image hydraulic fractures is microseismic monitoring, which analyzes seismic information recorded during hydraulic stimulation to locate the rock deformation. Results of the microseismic data interpretation are then used to generate and calibrate a model of the hydraulic fracture network. However, because of the complexity of the fracture model and the shortcomings of reservoir simulators, direct application of these complex fracture networks has been very limited. Instead, oversimplified models are used to assess the efficiency of the hydraulic fracturing treatment. Such assessment techniques, without further modeling and simulation of hydrocarbon production and pressure drainage, fail to represent an accurate view of the connectivity and complexity of the fracture system. In this paper, we present the application of an Embedded Discrete Fracture Model (EDFM) in numerical simulation of realistic geometry of fractures. With EDFM, each fracture plane is embedded inside the computational matrix grid and is discretized by cell boundaries. We have implemented EDFM in The University of Texas at Austin (UT) in-house reservoir simulator UTCOMP. We discuss the implementation approach using non-neighboring connections. Using the developed simulator, we studied gas production from hydraulic fracture networks calibrated from actual microseismic monitoring data. We investigated the impact of fracture network geometry on the overall performance of these hydraulic stimulations. Simulation results indicate that the efficiency of well treatment is primarily controlled by the interconnectivity of hydraulic fractures and the distribution of conductivity within the fracture network. For a given microseismic cloud, a wide range of production responses was observed by changing the degree of connectivity in the calibrated model. Moreover, the study showed that taking into account the role of aseismic deformations (such as tensile openings) significantly increased cumulative production forecasts. Neglecting the effect of these fractures may lead to underestimation of ultimate recovery.