Lithium-ion Battery SOC Estimation by Integrating Electrochemical Model with Deep Learning

The development of new energy vehicles has led to a surge in interest in lithium-ion batteries. In order to ensure the safe and efficient use of these batteries, an accurate assessment of the State of Charge (SOC) is essential for battery management systems. Data-driven estimation methods provide a new approach for SOC estimation, but they always neglect the internal state information of the battery. To overcome this limitation, the paper proposes a method to estimate the SOC of lithium-ion batteries. This method integrates an electrochemical model with deep learning. Firstly, an electrochemical simplified model, the Single Particle Model (SPM), is introduced to obtain internal battery information. Secondly, we propose a Bidirectional Long Short-Term Memory (BiLSTM) network as a data-driven model. The input current, temperature, terminal voltage of the SPM and the deviation between the terminal voltage and the experimental data are included in the input vector of the neural network. The output vector represents the state of charge determined by the ampere-hour integral method. Finally, the neural network was trained using data collected from the DST test at 25°C, and data collected from the US06 and FUDS tests were selected to evaluate the predictive performance of the neural network. Simulation results demonstrate that the SOC estimation method, which integrates the electrochemical model with deep learning, achieves high prediction accuracy with an error of less than 2% under different driving cycles. This paper serves as a reference for research on methods of estimating SOC in lithium-ion batteries.