Study Of Fault Diagnosis Of Crusher Based On Neural Network

Crusher plays an important role in mine production, which not only has a direct inference on the entire production line but also causes significant economic loss and even crash accident. In order to guarantee safe operation of equipment, reduce equipment repairing costs and improve the utilization rate of equipment, designing a monitoring fault diagnosis method which could acquire knowledge automatically and inference in high speed has became one of the main researching directions of the monitoring fault diagnosis of equipment.In this paper, fault diagnosis and monitoring the state of mining crusher were studied.Using data acquisition analyzer PDM2000 to collect crusher vibration signal, obtain eigenvalues. According to the theory of equipment vibration diagnosis technology,Analysis and discuss the fault symptoms of crusher failure, study vibration mechanism of each failure. Selecting the method of artificial neural networks, fault diagnosis for the crusher was studied.This paper elaborates the structure and algorithm of BP neural network, three method s were used to improve neural network algorithm; The three improved algorithms were co mpared and a more suitable improved algorithm was chosen; Matlab software was used to select feature data; Finally, different fault types of data were measured, and it was subject ed to tested by selected neural network, the fault type is determined to obtain a more accur ate diagnosis.Verification of the measured vibration data of crusher was conducted in this paper, th e results show that fault diagnosis method of crusher which was proposed based on BP ne ural network in the paper has the characteristics of practicality, effectiveness etc.