Semi-analytical modeling of shale gas flow through fractal induced fracture networks with microseismic data

Abstract Hydraulic fractures connecting to the adjacent induced fracture network significantly promote the productivity of unconventional gas reservoirs. Precise characterization of the fracture network in stimulated reservoir volume (SRV) is particularly important in modeling shale gas flow mechanisms. The objective of this work, based on available microseismic information of fracture density and field production data, is the integrated modeling of shale gas production with time-dependent pressure, pressure-dependent gas properties, and scale-dependent heterogeneous induced-fracture properties. This paper presents a Fractal Induced Fracture Network Distribution (FIFND) model to characterize SRV heterogeneity. The model consists of a novel fractal induced-fracture density distribution and a fractal permeability/porosity distribution. The FIFND model can accurately estimate the induced fracture permeability and porosity when only the microseismic data of fracture density are available. This is highly useful since microseismic fracture density data are more frequently available than permeability and porosity data. A semi-analytical Fractal Transient Shale Gas Flow model (FTSGF) is then derived for the multi-stage hydraulically-fractured horizontal wells. The FTSGF model is coupled with the FIFND model to better describe the fracture network heterogeneity in the SRV. The transient flow contribution from the matrix is modeled by apparent matrix porosity with the presence of adsorbed gas. The fractal transient flow features are ultimately transformed into a characteristic function of the fractal matrix-fracture flow transfer. The FIFND model is validated though the upscaled microseismic geological data for a Barnett shale well. Fracture density distribution, which is regulated by the Hausdorff dimension, is more significant on well productivity than fracture permeability, which is mainly subject to the fracture tortuosity index. The FTSGF model is verified by the field production data in Barnett shale. The robustness of the FTSGF model is justified by a good fit with the Power Law Exponential Decline model (PLE) and alignment with realistic values of multiple physical parameters, including average induced-fracture apertures. Our model is also validated for predicative robustness using fewer months of production data. Finally, we propose and implement a workflow for the integrated semi-analytical modeling of shale gas production. The workflow provides an effective tool for characterization, history-matching, and forecasting reservoir/well performance of hydraulically-fractured horizontal wells in shale reservoirs. The limitations of the proposed models and potential future expansions are discussed.