Automatic history matching of multistage fractured shale gas reservoir constrained by microseismic data

Implementing an effective automatic history matching (AHM) is essential for the shale gas reservoir characterization and production optimization. Embedded discrete fracture model (EDFM) has been recognized as an efficient forward model used in AHM, in which the properties of multiscale fracture planes are required to be properly assigned. However, the multiscale fracture characterization based on the induced microseismic events (MEs) is still unclear, resulting in the over-simplified fracture planes used in the EDFM simulator. To estimate the properties of multistage fractured reservoir more accurately with both production and microseismic data, a new two-step history matching workflow is proposed. The energy-based K-means clustering (EBKC) is applied to characterize the distribution and geometry of multiscale fracture planes based on mutistage microseismic events (MEs). The obtained fracture planes are then used as prior model parameters in the Ensemble Kalman Filter (EnKF) based history matching method to assimilate the production data. The performance of the proposed workflow is verified by a synthetic case study, and the updated ensemble of production forecast successfully matches the observation data. The uncertainties associated with the updated ensemble of fracture property and permeability of primary fractures are significantly reduced, but substantial uncertainties are still remain for the secondary and tertiary fracture permeability. It indicates that the production data is more sensitive to the permeability of primary fracture. Notably, the highest-frequency component of the updated ensembles matches the observations with less than 20% discrepancy. With the case study, it elucidates that our EBKC approach can successfully characterize the multistrand fractures implemented into a computationally efficient forward model, i.e., EDFM for multistage fractured shale gas reservoir. Therefore, the proposed AHM workflow is recommended for field application.