57 / 2021-07-11 00:31:58

Rolling Bearing Fault Classification Utilizing Adaptive Density Peaks Search Clustering Based on Wavelet Packet Transform

The development of machinery and equipment is moving in the direction of high speed, complexity and intelligence. Accurate judgment of the health status of precision mechanical equipment is the key to ensure the safety of industrial production. In order to improve the diagnosis level of mechanical equipment faults, this paper proposes a rolling bearing fault classification method utilizing adaptive density peaks search clustering based on wavelet packet transform. Combined with the wavelet packet energy extraction algorithm, a bearing fault classification algorithm is proposed based on the improvement of the Density Peaks Search clustering (DPS) algorithm. The algorithm can adaptively adjust the classification parameters to classify the unmarked fault data, enhance the computational accuracy and reduce the computational effort compared with the DPS algorithm. After bearing test analysis, the performance of the detection method is proved to be suitable for classification and diagnosis of unmarked rotating machinery faults, and the performance is better than other classification algorithms.