Research On Bearing Fault Diagnosis Based On Wavelet Packet And Support Vector Machines

Rolling is the most commonly used in rotating equipment and key components, and the frequency of fault is high.Its working condition has influence on performance of the whole machine directly, even the whole production line. So it is very significant to research on bearing fault diagnosis techniques.For the samples of fault diagnosis of bearing with limited data, fault vibration signal has the characteristics of nonstationary and strong noise background, this paper adopts the method of wavelet packet transform of fault vibration signal denoising, reconstruction and feature extraction, and combining Support Vector Machine(SVM) to fault type identification. In order to improve the recognition rate, the fault type of wavelet packet to extract the fault feature vector normalization processing; secondly, using the genetic algorithm optimization for the parameters of SVM; Finally, with test samples respectively for parameter optimization. parameter optimization of SVM is used to identify the fault type.Under the advantage of MATLAB platform.by comparing the experimental results,show that the proposed method has higher classification recognition rate,and within the scope permitted misjudgment,it can satisfy the need of industrial field.