A fault diagnosis method for coal mine machinery bearing

Aiming at the problem that existing adaptive diagnosis methods of coal mine machinery bearing fault were susceptible to the interference of high frequency noise and intermittent noise, which led to the low accuracy of original signal decomposition and feature extraction, a fault diagnosis method for coal mine machinery bearing was proposed which was based on modified local mean decomposition(MLMD). The method adopts adjuvant noise decomposition method in self-adaptive decomposition part of local mean decomposition(LMD) method, namely adding Gaussian white noise to original signal firstly and then carrying out LMD, so as to restrain influence of high-frequency and intermittent noise on signal decomposition. In feature parameter extraction part, MLMD method does Hilbert transformation for product function components, then extracts feature parameters, so as to realize feature parameter extraction in whole value range. The simulation and test results show that MLMD method has good effect on decomposition and feature parameter extraction of bearing fault signal and high diagnosis accuracy of inner and outer ring fault of bearing.