Abstract
New energy vehicles are becoming a new trend in global transportation development due to the renewable and environmentally friendly nature of the fuel they consume. At the same time, the charging safety of electric vehicle (EV) lithium-ion battery limits the development of the industry. This paper obtains charging data through the EV charging network, takes the lithium-ion battery charging temperature as the observation value, and proposes an early warning method for EV lithium-ion battery based on the charging network according to the nonlinear relationship between the temperature and the charging voltage, current, and battery status. First, we obtain the charging data through the charging network, select the model input parameters, and establish the long- and short-term memory network and temporal convolutional network (LSTM-TCN) model to predict the EV charging temperature. Then, compare the real-time charging data with the predicted data to get the model with the highest accuracy, and analyze the residuals by using the sliding-window method to get the pre-warning thresholds. Finally, by monitoring and calculating the changes in residuals, a thermal runaway warning system is implemented for lithium-ion battery charging to ensure the safety of EV charging. The experimental results show that the LSTM-TCN charging early warning model has higher accuracy compared with other models, which makes the method able to accurately and quickly react to charging accidents and achieve the early warning effect.