State of health estimation based on PSO-SA-LSTM for fast-charge lithium-ion batteries

Fast-charge technology has been widely used in lithium-ion batteries for electric vehicles. The fast-charge battery has a much different influence on the aging characteristics, compared with the slow-charge battery. Hence, it is essential to further study the state of health (SOH) estimation of fast-charge lithium-ion batteries. However, existing studies on SOH estimation of fast-charge batteries fail to analyze the aging characteristics of different fast-charge batteries, and the effect of health features is poor. In this paper, a novel SOH estimation method of self-attention LSTM based on particle swarm optimization (PSO-SA-LSTM) for fast-charge lithium-ion batteries with novel incremental capacity (IC) features is proposed. Firstly, the IC curves of different fast-charge batteries are analyzed, and the proposed IC features are extracted. Then, Pearson’s correlation coefficient is used to analyze the correlation between the features and SOH. The proposed model with a self-attention mechanism is established and the hyperparameters are optimized with particle swarm optimization (PSO). Finally, various comparison experiments with different features and different models are conducted on two fast-charge battery datasets. The results demonstrate the proposed method with novel IC features has high accuracy and robustness for SOH estimation of fast-charge batteries.