Method to Improve the Quality of Diagnostic Information
Zhang Haijun,Wen Guangrui,Qu Liangsheng
(School of Mechanical Engineering,Xi'an Jiaotong University,Xi'an 710049,China)
　
Abstract: The essential objective of an engineering diagnostic system is to detect the potential faults existing in a continuously running machine. Complete and high quality diagnostic information is necessary for identifying the faults correctly. However, in practice, because of noise and the mixing of signals due to different components, the signal to noise ratio (SNR) of the signal picked up is usually low. Of course, this deeply affects signal analysis and feature extraction, and adds the difficulty of fault detection. In order to improve the quality of diagnostic signal and identify the faults correctly, the continuous wavelet transform (CWT) and independent component analysis (ICA) are adopted together. Both simulation and example reveals the efficiency of this method.
Keywords: fault diagnosis;wavelet transform;blind source separation;indepentent component analysis