Research On The Key Techology And Application Of Mine Hoist Braking System Fault Diagnosis

Mine hoisting equipment is the key equipment linking uphole and downhole in coal mine energy exploitation.The safe and reliable operation of the mine directly relates to the economic benefits of the enterprise and the life safety of the workers.The braking system is also one of the most important part of the lifting equipment.Therefore,the fault diagnosis of the braking system of mine hoisting equipment is of great significance for safe and efficient production.This paper takes the mine hoist as the engineering research background,and focuses on the key technologies such as knowledge acquisition and knowledge reasoning in fault diagnosis of the hoist braking system.A fault diagnosis method that integrates Rough Set,MATLAB,Bayesian Net,and heuristic search algorithms.This method realizes the fault diagnosis of hoist braking system by taking advantage of their respective advantages.The main research content is as follows:First,Kingview technology is used to achieve real-time data monitoringand storage of the hoist braking system.Through ODBC data source and code in MATLAB,connecting MATLAB and database,you can directly call data from the database and perform data processing in MATLAB.Secondly,for the problems of difficult diagnosis and large amount of calculation,the improved rough set and MATLAB technology are used to establish the knowledge acquisition model of fault diagnosis rules.The improved rough set contains the improved discernibility matrix and the importance of the improved attribute importance.This method is applicable to the reduction of any decision table;finally the improved rough set conversion implemented in MATLAB and implemented Fault diagnosis knowledge is automatically obtained.Thirdly,aiming at the more uncertain problems in hoist fault diagnosis,the Bayesian network structure for fault diagnosis of hoist brake system is established;the structure of the network structure is studied,and the desired maximum optimization algorithm(EM))is adopted to reduce the external strong interference factors.The Bayesian network parameters learning,as well as the knowledge of failure Bayesian network structure such as the lack of braking torque of the braking system,establishes the braking system's uncertain reasoning model.Fourthly,according to the uncertain reasoning model of the shortage of braking torque,a search tree for the failure of braking torque fault is established;then,an improved A-star algorithm is written in Java language based onMyeclipse software.The fault search tree is searched to obtain the best path for the fault tree search.Finally,diagnostic faults and maintenance suggestions are issued to facilitate user fault analysis and maintenance.Fifth,set up test rigs and formulate practical test plans.We conducted experimental research on data acquisition,knowledge acquisition of fault diagnosis rules,and uncertain reasoning.We obtained expected results and verified the correctness and effectiveness of fault diagnosis results.And the above results have been successfully applied to the fault diagnosis of a mine hoist.