Hybrid Neural Network Method of the Lithium-Ion Battery Remaining Useful Life Prediction

Abstract

In this paper, a hybrid neural network is proposed by combining one dimensional convolutional neural network (1D CNN) and bidirectional long short-term memory (BiLSTM) neural network to improve the accuracy and stability of remaining useful life (RUL) prediction. 1D CNN-BiLSTM hybrid neural network has the ability to extract deep features in data and to save the memory of historical input information. To verify the authority of the proposed method, the lithium-ion battery data of the Center for Advanced Life Cycle Engineering (CALCE) are utilized to make some comparisons among the 1D CNN model, LSTM model and other neural network models. The results show that the hybrid one has higher prediction accuracy than the others.