Fault feature extracting for rotating machinery vibration based on blind deconvolution and spectral kurtosis

作者：Huang Weiguo^*; Liu Haiyang; Zhu Zhongkui; He Zhiyong

来源：Journal of Theoretical and Applied Information Technology, 2012, 43(2): 261-266.

• Bearing failures
• Blind deconvolution
• Combination method
• Fault feature
• Frequency domains
• Spectral Kurtosis
• Statistical tools
• Vibration sources

Rotating machinery vibration analysis involves a convolute mixture because of the propagation medium, and the signals recorded by sensors in an industrial application are often disrupted by the environment. Deconvolution is a signal processing method for convolution of vibration sources, spectral kurtosis is a statistical tool which can indicate the presence of series of transients and their locations in the frequency domain in strong noise case. In this paper, we propose an approach for the two characteristics based on blind deconvolution and spectral kurtosis. First the two methods blind deconvolution and spectral kurtosis are reviewed, and then puts forward the combination of the two methods to extract the fault feature from multi-sources convolution and strong noise in rotating machinery vibration, Finally apply the combination method to a bearing failure test, the test results show good performance for extraction of fault features in rotating machinery vibration.

• 出版日期2012