State-of-charge Estimation of Lithium-Ion Battery Based on Second Order Resistor-Capacitance Circuit-Pso-tcn Model

Accurate state-of-charge (SOC) estimation of lithium-ion battery is directly related to the reliability, performance, and safety of the battery. In this work, the second order resistor-capacitance (RC) circuit is equivalent to the battery model and the particle swarm optimization (PSO) algorithm is employed for achieving accurate identification of the parameters of circuit model under dynamic stress test (DST) conditions. Furthermore, the values of open circuit voltage (OCV) obtained from the identification results are input into the temporal convolutional network instead of the terminal voltages, and then the SOC of the Li-ion battery is estimated by directly learning the mapping relationship of OCV-SOC curves, which further improves the estimation accuracy and robustness of the proposed method. Finally, the SOC estimation is validated with the public dataset of LiFePO4 batteries under all driving conditions at different temperature and compared with the individual TCN method. Results show that the SOC estimation of the second order resistor-capacitance circuit-PSO-TCN model is optimal with a root mean square error (RMSE) and maximum error (MAXE) less than 1.8% and 7.65%, respectively.