Study And Development On Fault Diagnosis System For Roller Bearing Based On Wavelet Transform

Roller bearing is the common component in rotating machinery. Its running state can influence the performance of the whole machine directly. According to the fault diagnosis of precision engineer bearing, this paper firstly, generalizes the fault mechanism and vibration characteristics of roller bearing. Secondly, this paper introduces several feature extraction methods, such as time-domain analysis, spectrum analysis, envelope technique and wavelet transform and so on. To extract early fatigue feature of engine bearings, the paper presents an average spectrum method with double-wavelet and the diagnosis examples are used to show the validity of the method. Thirdly, expert system application in bearing fault diagnosis is researched. The paper puts forward the storing way of "Diagnosis Tree + Rule" in the knowledge management and the reasoning measure of forward and back in reasoning machine. Finally, combining wavelet and other feature extraction methods with expert system technique, a fault diagnosis system for roller bearing is designed. Then, based on Windows and used object-oriented Delphi6.0, the fault diagnosis system named VGFD1.0 for roller bearing is worked out.The fault diagnosis system is tested by simulations and applications. The results show that the diagnosis system can detect bearing fault accurately, and offer information including fault degree, fault element, fault reason and the corresponding measures to the user.