Research On Diagnosis Signal Demodulation Method Of Rolling Bearing Based On Wavelet Packet Entropy

In China’s modern national economy, especially machinery, transport, energy, metallurgy, petrochemical, defense and other industries in the key machinery and equipment,most of its work environment has the characteristics of corrosion, high temperature, high pressure and other complex environment. The key parts and important mechanical structure of the equipment will inevitably have different degrees of fault. Bearing due to the location of the special parts, often prone to imbalance, misalignment, bending, fatigue spalling, pitting and other faults. Once the fault occurs, it will lead to serious economic and property losses, reduce the operating life of the equipment. Therefore, this paper mainly focuses on the fault signal of the rolling bearing as the research object, the research and discussion on the method of the fault signal de-noising and demodulation, the main work of this paper includes:1. Analysis of the commonly used in bearing fault signal processing of several noise reduction methods, comparing the wavelet and wavelet packet de-noising theory. According to the fault signal often contain lots of background noise to remove, is proposed based on wavelet packet entropy and EMD de-noising combining method. In this method, after the wavelet packet entropy effective noise reduction, and then decomposed with EMD, can adaptively from the fault signal extracting weak fault component.2. De-noising method based on wavelet packet entropy and combining autocorrelation analysis is proposed in this paper. Using self-correlation analysis to outstanding periodic fault signal characteristics, combining the wavelet packet entropy noise reduction method. In the wavelet packet entropy de-noise method removal with a lot of noise, at the same time, the self-correlation analysis further restrain noise. The periodicity of the signal is highlighted on the basis of the original fault modulation information has retained.3. The advantages and disadvantages of various signal demodulation methods are compared. The demodulation effect of energy operator demodulation method and Hilbert demodulation method under the same noise reduction method is analyzed. A combination of two kinds of noise signal analysis method. Based on wavelet packet entropy and EMD energy operator demodulation method and based on wavelet packet entropy and correlation analysis of the energy operator demodulation method is proposed in this paper, so as to accurately determine the fault location.4. Introducing of EMD and EEMD multi component analysis method. Combining the wavelet packet entropy reduction method, autocorrelation analysis, energy operator demodulation method and multi component analysis method. A fault signal energy operator demodulation method based on multi component analysis and autocorrelation analysis is proposed. The method on the basis of effective noise reduction of the wavelet packet entropy. The fault signal can be effectively demodulation, and make a judgement of the fault location.In this paper, the research work for analysis and diagnosis of rotating machinery fault signal demodulation method provides a new train of thought, for rolling bearing fault signal demodulation processing has a certain reference.