Research on gearbox fault diagnosis method based on DTCWT and order spectrum

Abstract

Because the gear boxes of large wind turbine unit operate under complicated working conditions in a long-term, the vibration signals collected from the gearbox are subjected to a large amount of background noise. In order to effectively extract fault features from vibration signals, the heuristic soft threshold and dual tree complex wavelet transform were adopted to denoise the collected signals. Then, according to the speed pulse signal collected synchronously, the rotating shaft frequency and the gear fitting frequency were calculated by time measuring method, and the same frequency sine data was generated, and then the correlation between the sine data and the vibration data was calculated to judge the fault location preliminarily. Lastly, the three order equation fitting method was used to carry out order resampling, and the power spectrum of the order data was calculated to obtain the gear fault feature. The simulation of Matlab and experiment results show that this method is effective in fault diagnosis feature extraction for wind turbine gearbox.