Life Prediction of Electric Connectors for EMUs based on WOA-BILSTM

The aging of electric connectors in EMUs may lead to important power system failures. Online prediction of its life is a convenient strategy for preventing such failures, thus simplifying the application of predictive maintenance plans. In order to construct the trend health index of electric connector degradation process of EMU and improve the life prediction accuracy of electric connector, a BILSTM deep learning model based on whale optimization algorithm (WOA) was proposed for the life prediction of electric connector. WOA is used to optimize some related parameters in BILSTM. BILSTM was used to predict the contact resistance of connectors in operation. Finally, the effectiveness of the proposed method is verified by comparing the predicted value with the measured value. The results show that the fitting degree of this method reaches 92.68%.