Remaining useful life prediction of PEMFC based on CNN-Birnn model
Jiale Luo, Tao Chen, Fei Xiao, Yulin Peng
Abstract
Proton exchange membrane fuel cell (PEMFC) is a new type of clean energy with great development potential. However, as working time increase, the output power of the PEMFC will then be reduced. By predicting the PEMFC degradation trend, faults can be detected in advance to ensure continuous and efficient working of the FC. In this paper, convolutional neural network (CNN) and bidirectional recurrent neural network (BiRNN) are integrated into a new network (CNN-BiRNN) model for voltage degradation and remaining useful life (RUL) prediction of PEMFC. The model is validated by the aging test data of two real FCs. The results indicate that the combination with CNN can significantly enhance the prediction accuracy and calculation speed of the BiRNN model. The CNN-BiRNN model has smaller mean absolute percentage error (MAPE) and mean square root error (RMSE) for voltage degradation prediction than existing models. The mean and standard deviation of the relative error (RE) of the RUL prediction of the FC for five different fault thresholds are smaller. The proposed model is more accurate in predicting the voltage degradation and RUL of PMEFC.