Research of Rolling Bearing Fault Diagnosis Method Based on BO-DK-ELM

Aiming at solving the problems that the parameters of the traditional Extreme Learning Machine (ELM) model have significant impact on the fault diagnosis accuracy while it is applied in rolling bearings, a rolling bearing fault diagnosis method based on Bayesian Optimization Deep Kernel Extreme Learning Machine (BO-DK-ELM) was proposed in this paper. Firstly, a model based on Deep Kernel Extreme Learning Machine (DK-ELM) algorithm is constructed by combining the Auto Encoder (AE) with the traditional Kernel Extreme Learning Machine (K-ELM); Secondly, the Bayesian Optimization (BO) algorithm is applied to find the optimal hyper-parameters in the above proposed DK-ELM model, in order to minimize the sum of classification error rates of the training and validation sets in the DK-ELM model. And then the testing set would be fed into the trained BO-DK-ELM model for fault diagnosis. Finally, our proposed method was validated using the Case Western Reserve University bearing fault dataset, and the final results show that its fault diagnosis accuracy can reach 99.0%. Meanwhile, compared with other traditional intelligent fault diagnosis algorithms, the proposed method has higher fault diagnosis accuracy.