The Research On SVM Fault Diagnosis Of Mine Hoist

The mine hoist equipments are the only transportation media between the mine and theoutside environment. The high frequency of braking during hoisting or descending of hoistingsystem guarantees the mining high speed increase, at the same time brings the commonequipment vibration’s hidden problems. This paper did a further study on SVM, and finally usedLibsvm software to complish the fault diagnosis of “the abstract of dominant frequencycharacteristics+the optimization of classifier kernel function+the optimization of parameters”the mine hoist equipments. First, analyse the fault types, the vibration characteristics and signalprocess, do a site vibration test and diagnosis for the hoist of a mine (the main well)JKM3.25×4(II) based on the analysis of the data spectrum.; second, set up a fault tree of hoistsystem, abstract hoist characteristic information, and then establish a SVM database and do anormalization and PCA to the data; in oder to get the best SVM fault classifier, distribute the testdata again, find a decent optimization of classifier kernel function and parameters, put forward“the abstraction of dominant frequency characteristics+RBF kernel function+the optimizationof Grid model parameters”, a comprehensive fault diagnosis system; finally, with thiscomprehensive SVM fault diagnosis system, establish a SVM fault classifier to diagnose thefault of the mine hoist equipments, contrast and analyse the diagnosis result and the dataspectrum result.The study result showed the SVM technology can apply to the huge mine hoist’s faultdiagnosis, and can guarantee the hoist equipments run reliably and economically, so it has a veryimportant study value and practical meaning.