Research On Fault Diagnosis For Main Circuit Of Mine Hoist

The hoist is a key equipment,which is the throat of mine. So far,the heavy accidents from mining production in our country due to mine hoisting system fault, have caused great economic losses and casualties. The power supply transformer, the inverter in main circuit,hoist and drive roller are included in mine hoist main loop system. Any equipment have fault,which will be not safe for production in coal mine. The fault diagnosis technology and engineering technology are combined, and main circuit fault diagnosis in mine hoist has been studied in order to solve the actual fault diagnosis problem and prevent fault. In the end, the hoist safety and reliability of the system operation are ensured.The constitution of main circuit power supply system in mine hoist is analyzed, for each link of main circuit system, the corresponding fault tree is established based on the fault tree theory. Then the weak links of the system are found out, which will provide guiding ideology for the subsequent fault diagnosis research.On the basis of analyzing the characteristics of the transformer faults, the method of transformer dissolved gas-in-oil analysis is studied. According to the effective information provided by dissolved gas-in-oil, the relationship between the composition content of the fault gas and the transformer fault types has been researched. Because the detection accuracy of traditional transformer fault diagnosis method is not high, the improved binary tree support vector machine(SVM) has been applied, which is a new method and separability measure calculation is combined with. Because its nuclear parameters C andghave great influence on fault classification, the cross validation parameter optimization is used. The better effect of parameter optimization is obtained and the selection problem for the support vector machine(SVM) parameters is solved, the efficiency and accuracy of transformer fault diagnosis are improved.For rectifier part in inverter, according to the characteristics of the power electronic circuit fault, the bilinear grid search algorithm is used to determine the corresponding optimal error penalty parameter and Gaussian kernel parameters. An improved one-to-manyclassification algorithm is proposed and the classifier is built to classify the fault feature.When the value of one classifier is true, it is no longer for the next step of calculation, and the circuit fault type is determined, which effectively reduces the amount of calculation and improves the efficiency of the test. In view of the actual data collected by the frequency converter, the good diagnostic accuracy is achieved by the algorithm simulation of the application.In view of the winding broken bar fault from main drive motor( asynchronous induction motor rotor) in coal mine hoist, fault diagnosis method based on support vector machine(SVM) is proposed, in which multi-layer wavelet packet decomposition is used. The method is based on the fault mechanism analysis of motor rotor winding broken bar, the current signal of motor stator winding is sampled. Then the wavelet packet transform is used, with which the normal signal and fault signal are made for 5 layer wavelet packet transform, the energy characteristic information after reconstruction is extracted, and is made for normalized processing. Then the information works as input vector of the support vector machine(SVM),the diagnosis process of asynchronous motor fault diagnosis system is established, and the actual rotor broken bar fault from mine hoist has been verified and good effect has been achieved.