Microseismic monitoring has been widely used to image the growth of fracture network in unconventional reservoir development. It is the key for stimulation optimization to understand clearly geometric features of discrete fracture network (DFN) created by hydraulic fracturing. However, due to the complex interaction between natural fractures and hydraulic fractures, it meets great difficulties in building a realistic 3D DFN. In this study, we go beyond the existing methods that develop analytical model to simulate propagation of created-fracture in formation with pre-existing natural fractures. We present an enhanced fracture regression method based Random Sample Consensus (RANSAC) to provide a rough 3D DFN model from microseismic events directly. In this approach, the laboratory fracturing experiments are conducted and convex polygon model based on alpha-shape algorithm is developed to characterize complex fracture geometries. RANSAC combined with a novel occurrence calculation method is used to detect the fracture planes in microseismic events. Simulations by Monte Carlo and case study show that the integrated RANSAC method is able to achieve fast fractures regression and is also robust and reliable enough for a rough 3D complex DFN modeling.
An Enhanced RANSAC Method for Complex Hydraulic Network Characterization Based on Microseismic Data
Liu, Xing, Jin, Yan, Lin, Botao, Xiang, Jianhua, and Hua Zhong. "An Enhanced RANSAC Method for Complex Hydraulic Network Characterization Based on Microseismic Data." Paper presented at the 2nd International Discrete Fracture Network Engineering Conference, Seattle, Washington, USA, June 2018.
Download citation file: