An integrated 3D fracture network reconstruction method based on microseismic events

Abstract

3D fracture network modeling remains a challenging work with the wide application of microseismic monitoring in multi-stage hydraulic fracturing. Due to the existence of invalid noise events, conventional methods often have significant deviations in detecting fracture geometry for fracture modeling after multi-stage fracturing.To solve the problem, we present an integrated 3D fracture reconstruction method that incorporated random sampling consensus algorithm (RANSAC), density-based spatial clustering of applications with noise algorithm (DBSCAN), and the alpha-shape method. To evaluate the effectiveness and robustness quantitatively, we also develop a verification workflow based on synthetic events generated by discrete fracture networks (DFN) with different fracture numbers, fracture scales, and dip angles. Through a large number of simulations, it is found that our method has perfect accuracy and good anti-noise ability in handling events with a high proportion of noise. Besides, our method has been successfully applied in a horizontal well stimulated by multi-stage fracturing in Sichuan Basin of China. The reconstructed fracture network is very consistent with the field production after source mechanism analysis and history matching. Our method is helpful to obtain a more accurate 3D fracture network model after stimulation and provides a basis for fracturing modeling in shale gas development.