科研之友
微信
新浪微博
Facebook
分享链接
摘要
The collected vibration signals from the incipient faulty bearing are generally corrupted by strong background noise. The weak fault feature extraction is crucial to mechanical fault diagnosis and machine condition monitoring. Empirical Mode Decomposition (EMD)-based method is one of the most powerful signal processing techniques and has been extensively studied and widely applied in fault diagnosis of rotating machinery. However, they have also several outstanding problems. Considering these, a new method for bearing weak fault diagnosis is proposed in this paper. It consists of five parts. Firstly, the frequency band ranges of meaningful modes are self-adaptively determined by combining scale-space representation and empirical law. Secondly, the meaningful modes are obtained by using EWT (Empirical Wavelet Transform) to decompose the raw vibration signal, according to the determined boundaries. Thirdly, the Normalized Frequency Energy Ratio (NFER) is presented to define the fault-related components. Fourthly, Teager Energy Operator (TEO) is employed to further enhance the fault-related impulses. Lastly, the fault-related features could be observed from the time waveforms and envelope spectra of the improved modes. Two simulated bearing fault signals and a real bearing fault signal are used to validate the performance of the proposed method. As result, the proposed method delivers a better performance than that of the EMD-based demodulation method.
