Research On Fault Diagnosis Of Subway Trains Rolling Bearing Based On Acoustic Emission Technique

While the scale of urban rail transit development in China is more and more large, operational security has also become important issues. Train as a carrier of urban rail transport, running speed, large capacity, complex operations environment and the security situation is not optimistic-train failures continue to occur, accidents also often occur, these faults will not only seriously affect the normal order of operations, once the accident will lead to huge loss of life and economic damage to property.Rolling bearings are most widely used in the train, which is one of the most easily damaged parts. Train’s many of the faults are related to rolling bearing, bearing is a direct impact of the State of security of the train. To improve the security of trains, train rolling bearing fault diagnosis research, and find the failure regularity and exploration of effective troubleshooting method, which is a valid safety early-warning security measures.This paper focuses on the safety analysis of brake system and has made the following studies:1. Research put forward rolling bearing fault diagnosis method based on the acoustic emission. The system of rolling bearing summarizes the faults and fault diagnosis method research trend, summarizes the application of the principle of acoustic emission, influence factors, audio emission advantage and audio emission signal processing method。2. In the paper the acoustic emission technology is used, in the rolling bearing on the experiment of inner loop, outer ring, roller bearing fault and normal the acoustic emission signal acquisition, through to the acquisition of the primitive acoustic emission signal data, then decomposition and reconstruction, extract fault feature vector, as the neural network’s input data on Internet training.3. The rolling bearing processing method of acoustic emission signal based on the wavelet packet and a variety of neural network is put forward. In the acoustic emission signal wavelet packet decomposition, and on the basis of use of neural network training samples, the rolling bearing fault type recognition, and find out the rolling bearing fault position. Papers will make wavelet packet theory and neural network of organic combined, drawing on their advantages; improve the train rolling bearing fault diagnosis of efficiency. 4. Offer relatively detailed subway train rolling bearing fault diagnosis of acoustic emission system design thought. From the general introduce the structure of the system framework, then respectively from two aspects of hardware and software introduce the function of the system, provide good design ideas and function design direction for research and development of rolling bearing fault diagnosis acoustic emission system, the research and development will provide good theoretical basis for the next step.