State-of-Health Estimation of Lithium-ion Batteries Based on WOA-CNN-LSTM-Attention

In this paper, five health characteristics are extracted from battery charging and discharging data, and combined with the whale optimization algorithm, the performance of the two models in battery health state estimation is compared, namely, the long-short term memory neural network and the long-short term memory neural network with attention mechanism. On this basis, the convolution neural network module is added to extract hidden features from the data, so that the model can obtain more information, thus improving the performance of the model. Finally, the relative error of the improved model for battery health state estimation is controlled within 1.6%.