Extraction and diagnosis of rolling bearing fault signals based on improved wavelet transform

As the continuous growth of the machinery industry, the importance of rolling bearings as key connecting parts in machinery movement is also increasing. However, the extraction and diagnosis of rolling bearing fault signals are difficult, and how to use modern transform analysis methods to raise the extraction efficiency and diagnostic accuracy becomes the focus. For this, a rolling bearing fault signal extraction and diagnosis model is designed based on empirical wavelet transform. The diagnostic model is optimized by using support vector machine and quantum genetic algorithm to design a rolling bearing fault signal extraction and diagnosis model based on improved empirical wavelet transform-support vector machine. The test results show that the research method can obtain four component signals showing different anomalies when generating time domain diagrams. Only five component peaks are generated and one group is extracted as output when generating component peaks. The abnormal amplitude of envelope spectrum basically reaches 0.40x10-6 or above. The judgment accuracy of component diagnosis reaches 98.12%. The above results show that the research method has better fault signal extraction ability and better diagnostic accuracy when performing fault signal diagnosis, which can provide new technical support for rolling bearing fault signal extraction and diagnosis.