RANSAC-MP: A Robust Plane Fitting Algorithm for Noisy Microseismic Data During Hydraulic Fracturing

Abstract Hydraulic fracture morphology is an important parameter for fracturing effect evaluation, reservoir simulation and production prediction. It mainly relies on microseismic interpretation to obtain hydraulic fracture morphology, but the presence of ambient noise and irrelevant rupture events can cause the microseismic monitoring data to be affected by multiple noises, and the accuracy of existing fracture plane fitting algorithms is difficult to be guaranteed under complex noise environments. In order to solve the problem of fracture plane fitting under complex noise, this paper proposes a robust hydraulic fracture plane fitting algorithm - Random Sampling Consensus-Maximum Projection (RANSAC-MP) algorithm, which weakens the outlier noise caused by irrelevant rupture events through the random sampling method and embeds the projection transformation to reduce the environmental noise. The results show that the RANSAC-MP algorithm has stronger robustness and higher accuracy in complex noise environments, and the fitting accuracy is improved by 14% compared with the RANSAC algorithm, 38% compared with the PCA algorithm, and 56% compared with the LS algorithm. The fracture plane fitting is carried out with actual straight well microseismic data as an example, and the results show that the RANSAC-MP algorithm shows stronger robustness and higher computational accuracy under the influence of multiple noises, and the algorithm can directly process the raw data when only a single fracture is formed by fracturing.