Remaining Useful Life Prediction of Lithium Batteries Based on CNN-GRU Model

Abstract

Aiming at the difficulty in obtaining direct performance parameters such as capacity and internal resistance of lithium batteries, which leads to the low accuracy of predicting the remaining useful life (RUL) of lithium batteries, a prediction model based on Convolutional Neural Network (CNN) and Gated Recurrent Unit (GRU) is proposed. Firstly, four indirect health factors including constant current charging time interval, constant voltage charging time interval, discharge temperature peak time and cycle times are extracted from lithium battery charging and discharging experiments, and the Pearson and Spearman correlation coefficients are established. Secondly, build a lithium battery RUL prediction model based on CNN-GRU combined neural network. Finally, the rationality of extracting health factors is verified by actual data, and the prediction results are compared with SVR model, long short-term memory (LSTM) model, GRU model, and CNN-LSTM model. The mean square of the proposed model is analyzed. Thus, the superiority and effectiveness of the proposed model are verified.