Rolling Bearing Fault Diagnosis Based on SCSO-LightGBM

Given the frequent occurrence of rolling bearing faults, to effectively enhance the accuracy of rolling bearing fault diagnosis, this paper constructs a bearing fault diagnosis classifier using a Light Gradient Boosting Machine (LightGBM), and applies Sand Cat Swarm Optimization (SCSO) to optimize the optimal hyperparameters, constructing a SCSO- LightGBM model. To avoid the lack of feature extraction at a single scale, the temporal domain and spectral domain features of rolling bearing vibration signals are extracted as fault features. The results show that the SCSO-LightGBM model presented in this paper can diagnose the rolling bearing fault type more accurately, and its classification performance exceeds that of other comparison models.