It is difficult to identify the characteristic frequency for the fault diagnosis of rolling bearing by the traditional time-frequency analysis method. In this paper, a fault diagnosis method for rolling bearings was proposed based on EEMD auto-correlation denosing and envelope spectrum. Frist of all, signals were decomposed into several intrinsic mode functions(IMFs), then calculated the concentrated energy ratio of IMFs autocorrelation function and found the demarc K by the coefficient P. Some IMFs containing the main fault information were selected and reconstructed signals by soft threshold filtering for those chosen and other IMFs. At last, the denosing signal was taken as envelope spectrum analysis to identify the bearing failure type. The effectiveness of the method was evidenced by using simulated signals and experimental signals. The experimental results showed that this method has a higher denosing effect than that based on the wavelet packet or EMD method, and can be used to classify and identify bearings fault patterns accurately and effectively..
展开▼
