The 10th International Conference on Energy, Resources, Environment and Sustainable Development (ERESD 2024)

383 / 2024-03-27 20:23:29

Motor Bearing Fault Recognition Based on Wavelet Packet Transform and S-K-means Unsupervised Learning

fault identification; wavelet packet entropy; S-K-means unsupervised learning; MATLAB

Intelligent Equipment Technology > 7．Equipped with Safety Monitoring and Control Technology

To achieve fault identification of motor rolling bearings, a fault identification method based on wavelet packet transform and S-K-means unsupervised learning is proposed. Firstly, perform time-domain and frequency-domain analysis on the vibration signals of rolling bearings under different conditions to preliminarily obtain the working conditions of the rolling bearings; Then, combining wavelet packet transform and Shannon entropy, a wavelet packet entropy that can be used for fault feature analysis is proposed. By performing S-K-means unsupervised learning on the entropy value, the range of wavelet packet entropy values for different faults is obtained, thereby achieving the goal of identifying unknown signal faults. Finally, four recognition tasks were designed, and the experimental results showed that the proposed method had a high fault recognition rate of 95.5%.