Lithium-ion Battery State of Charge (SOC) Estimation Based on HHO-LSTM

State of Charge (SOC) estimation is one of the core contents of the Battery Management System (BMS), and it is crucial for evaluating the safety of using lithium-ion batteries as the power source for ships. Based on the existing problems in SOC estimation methods, this paper proposes a method for SOC estimation that uses the Long Short-Term Memory (LSTM) neural network as the core algorithm, optimizes parameters through the Harris Hawk Optimization Algorithm (HHO), and builds a co-simulation model. Firstly, using four batteries, B0005, B0006, B0007, and B0018, from the National Aeronautics and Space Administration (NASA) as the basic data, the random forest algorithm and Pearson correlation coefficient analysis are used to identify the linear and nonlinear correlations with the target variable respectively, eliminate redundant or irrelevant variables, and screen out features with high correlation with SOC to improve estimation accuracy. Then, the obtained features are input into the LSTM neural network model, and the Harris optimization algorithm is used to optimize the model parameter combination. Finally, digital simulation experiments are conducted. The results show that the HHO-LSTM prediction model improves the estimation accuracy of SOC. This method provides a new idea for SOC estimation and has certain effectiveness and feasibility.