The fracture morphology of hydraulic fracturing is a key parameter for evaluating the fracturing effect and predicting the yield. At present, the fracture information of fractured cracks is mainly extracted by microseismic monitoring at home and abroad, and it is difficult to obtain the fracture morphology through the planar identification algorithm by utilizing the microseismic data due to the existence of complex noise. For this reason, this paper proposes a robust planar identification algorithm for hydraulic fracturing cracks, the sampling projection algorithm (RANSAC-MP), which weakens the outlier noise caused by irrelevant rupture events through random sampling, and proposes a maximal projection planar fitting algorithm to minimize the influence of environmental noise, and at the same time, combines the noise resistance of the RANSAC algorithm and the advantages of the projection method with the noise resistance of the RANSAC algorithm. noise immunity and the dimension reduction effect of the projection method. 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 original data when only a single fracture is formed by fracturing.

A machine learning method for simulating compressible multi-medium flows is studied. The regression prediction of multi-medium Riemann solution is realized by using neural network. In order to make the training results more consistent with the physical flow, an additional physical constraint layer is constructed according to the discontinuity relationship of the flow field. A neural network model is established and applied to practical ghost fluid method (PGFM). Through a variety of typical one-dimensional and two-dimensional multi-medium flow problems, the surrogates trained by neural networks of different sizes are verified numerically. It is found that the results of neural network model are more consistent with the real situation after embedding physical constraints. In addition, the relatively simple neural network model can meet the computing requirements. Machine learning method has high computational accuracy and efficiency, and has potential development.