Research On Fractal Analysis For Bearing Fault Status Classification

Fault feature extraction and representation is the most crucial problem for the reliability and accuracy in the bearing condition monitoring and fault diagnosis. In this paper we select the fractal dimension as the characteristic parameter, because it can reflect signal complexity. We calculate the fractal signal separately from the signal of time domain and frequency domain.In chapter 1, at first, we expounded technical means for the condition monitoring of bearing, then briefly introduced and compared the existing test plan and adopted technology. Then the applications of time-domain statistics analysis, Fourier analysis, pour spectral analysis, envelope analysis, short-time Fourier transform, wavelet transform, bilinear time-frequency analysis, cyclostationarity analysis, fractal analysis. Lastly, the contents, the focuses and the innovations of this dissertation are pointed out.In chapter 2, at first, introduces the main failure types of bearings, including wear failure, fatigue and fracture failure, plastic deformation failure, etc. Then calculated the characterisitic frequency of outer circle and inner circle fault and the characterisitic frequency of roller fault, finally research the temporal characteristics of bearing vibration signal, when the bearing is at fault.In chapter 3, choose grid dimension for extracting the feature of bearing vibration signal. By changing the sampling points, calculating different grid dimension of vibration signal. Use different grid dimension to emerge the feature vector of bearing state. Then establish model space about bearing state. The distance of different feature vector can show the different of bearing state.In chapter 4, the applications of the fractal analysis based on wavelet transform to the classification of bearing's vibration signals are studied. Firstly the acceleration signals are decomposed to detailed signals at different wavelet scales by using the discrete wavelet transform. According to the relationship of the original signal and 1/ f process.The variances of detailed signals are calculated and then fractal dimensions of the acceleration signals are estimated from the slop of the variance progression. The fractal dimensions are significantly different among the different working conditions of the bearings and showes a high reproducibility. The results suggest that the wavelet-based fractal analysis is effective for classifying the working conditions of bearing. Chapter 6 gives the conclusions and the prospect about this study.