Evaluation on Development Height of Water-Conduted Fractures on Overburden Roof Based on Nonlinear Algorithm

Abstract

Based on the support vector machine theory, the particle swarm algorithm is used to optimize parameters, combined with an analytic network process in order to feature dimensionality reduction of the original data, and a nonlinear algorithm model combining statistical analysis and machine learning analysis is established. Taking the relevant data of overlying rock from two zones in Yingpanhao Coal Mine as an example, the main factors such as rock tectonic development, coal seam dip, mining height, mining operation method and stope width are ranked according to the weights of their contribution to the development height of the two zones in order to determine the main control factors, such as stope width and mining height. Using MATLAB as the experimental platform, 16 sets of two-dimensional mine data similar to the geology of the study area were divided into training and test sets for prediction and simulation, comparing the optimal solutions of various optimization-seeking algorithms to obtain quantitative prediction results based on the nonlinear algorithm model. The kriging interpolation process was carried out by ArcGIS to realize qualitative visualization, and effective classification is carried out according to the natural breakpoint method to obtain six development height divisions of the water-conducting fracture zone in the study area. The results show that the prediction model of the development height of two zones in the coal roof based on the non-linear algorithm has better accuracy and generalization ability. Predicted by the test set, the model’s prediction result MAE is less than 10%, and the accuracy is better than the traditional empirical formula method, which makes up for its lack of rough calculation accuracy and solves the problem of not being able to locate the height of roof crack development at specific borehole locations, which has significance in guiding the prediction and prevention of roof water damage and subsidence disasters.