A battery State of Health (SOH) estimation method based on Gated Recurrent Unit (GRU) neural network is proposed to address the issue of low accuracy in estimating the SOH of new energy vehicle power batteries. This method extracts multidimensional input features based on battery charging data, performs data cleaning and normalization on the features, and trains a GRU network to construct a battery SOH estimation model. The results indicate that the proposed method can achieve an average absolute error of 0.26% in estimating battery SOH, which is 1.04% lower than traditional calculation methods. This method can achieve a more accurate estimation of battery SOH and can be used for evaluating the aging status of electric vehicles.

Power battery  /  State of Health (SOH)  /  Adam algorithm  /  Gated Recurrent Unit(GRU)