Study On Characteristic Analysis And Intelligent Prediction And Early Warning Of Rock Burst In Coal Mine

With the gradual increase of coal mining depth in China,the complexity of high ground stress in deep coal mines has increased significantly,and the problem of rock burst has become one of the main threats to coal mine safety,which is one of the aspects of deep mine safety that need to be solved.At present,the research on rock burst mainly includes monitoring,analysis,early warning and prevention,etc.However,due to the extensive and complicated types of rock burst,and most of them are manually analyzed,which has the limitation of single-parameter analysis,so that many monitoring data have low utilization rate and poor timeliness in joint analysis,and the accuracy of relevant rock burst risk prediction models still cannot meet the needs of the industry,which needs to be further optimized and improved.The accuracy of relevant impact hazard prediction models still cannot meet the needs of the industry and needs to be further optimized and improved.In this paper,theoretical and experimental studies are conducted based on the analysis of the rock burst risk characteristics and the prediction of the rock burst risk level in coal mines,and the main research contents are:Firstly,we proposed to use text mining technology to analyze the rock burst risk features,using 29 rock burst related accident cases and several thesis studies as the sample data base,and carried out in-depth mining of impact ground pressure hazard features through text data pre-processing,linguistic processing,feature representation and extraction,text analysis,features aggregation and other steps,and finally established a set of rock burst risk features including geological aspects,operational aspects and human aspects,including 3 major categories of impact features and a total of 86 sub-categories of specific rock burst risk factors.Secondly,for the study of rock burst risk level prediction,the improved Slime Mould Algorithm(SMA)with good point set and small habitat technology is proposed to optimize the parameters of Support Vector Machine(SVM)to establish the GNSMA-SVM rock burst risk level prediction model,and the simulation study is carried out based on the sample public data of Zhangji Mine of Xuzhou Mining Bureau.The prediction of the rock burst risk level of the mine was carried out,and the final prediction results reached 91.7% accuracy,and the prediction results were compared and analyzed with different models of the same kind,which verified the effectiveness and accuracy of the model.Finally,using the example of 3106 working face of Gucheng coal mine,11 factors with high influence on the risk of rock burst at 3106 working face were selected from the above mentioned rock burst risk factors,and combined with relevant monitoring data,the rock burst risk level prediction model GNSMA-SVM was built to predict the rock burst risk level of 3106 working face.Figure [28] Table [18] Reference [85]...