Research On Coal Mining Face Safety Warning Based On Data Mining

In recent years,the coal mine safety situation has been gradually improved,but compared with other industries,coal mine safety accidents and casualties are still at a high level,especially major accidents occur from time to time,the safety situation is not optimistic.How to find the hidden danger in advance,eliminate the risk in time and prevent the accident is always the target of coal mine safety early warning research.In the existing researches,there are few researches on the inherent relationship between the hidden dangers and various accidents in mining face.Therefore,starting from the analysis of the causes of accidents in coal mining face,this paper studies the early warning index system of coal mining face,the early warning ontology knowledge base,the data mining of accident hidden danger and the intelligent early warning model.The main research contents and innovations are as follows:(1)On the basis of combing the related literatures,the complexity and risk of production system of coal mining face are analyzed.Based on the statistical data of coal face accidents and typical accident cases,the "2-4" model of coal face accidents is built.Then the frame structure of coal mining face safety early warning system is studied and the hierarchical safety early warning research frame is put forward.The combination and application of different warning methods and models can realize comprehensive warning of coal mining working face,fully reveal the warning information of hidden dangers of coal mining working face and improve the reliability of warning,which can not be achieved by single index or single system.(2)Based on the index system early warning,expert system early warning and intelligent system early warning three different perspectives to carry out security early warning research.The early warning index system covering geological environment,operating environment,equipment,personnel,management and information security is constructed,and the index early warning model integrating dynamic and static indexes is constructed.Based on the knowledge of mining face safety early warning domain,and through the data mining of accident hidden danger,the ontology knowledge base model of mining face safety early warning is constructed.To give full play to the combination advantages of different intelligent algorithms,RS-SVM intelligent early warning model and GA-BP intelligent early warning model are constructed respectively.The multi-angle combination of early warning methods and models overcomes the disadvantages of relying solely on artificial experience,and solves the problem that local early warning is prone to alarm leakage and false alarm.(3)Based on the historical hidden danger data,different data mining methods are used to mine the hidden danger.The text clustering mining method is adopted to find the main types of hidden dangers in the hidden dangers data.The main safety problems existing in coal mining face are found with the theme model mining.The method of association rule mining is adopted to mine the internal association rules between hidden dangers,providing guidance for hidden dangers.The application of data mining from multiple perspectives has changed the situation that most of the hidden danger data is confined to monitoring and alarm,and provided a practical method for the deep application of hidden danger data.