Elman Neural Network Optimized by Swarm Intelligence for SOC Estimation of Lithium-Ion Battery

This paper presents an Elman neural network (ENN) based estimation method for state of charge (SOC) of lithium-ion batteries. Whale optimization algorithm, as a kind of swarm intelligent algorithm, is employed to optimize the ENN to obtain more accurate measurement results. Firstly, good convergence, memory function and better predictive performance are the advantages of the ENN, so it has been fully taken into account in SOC estimation. Secondly, the whale swarm intelligence optimization algorithm (WOA) is applied to optimize the neural network to avoid the over fitting problem. Finally, datasets from multiple standard hybrid driving cycles at different temperatures are applied in experiment to verify the effective estimation theory. The estimation results and errors indicate that the proposed method exhibits better accuracy on the SOC estimation than Back-Propagation neural network (BPNN) and ENN.