State of Health Estimation Method of Lithium-ion Batteries Based on the SA-CDC-GRU-AE Model
HU Yuhang
LIAO Yu
CUI Kun
LI Jingcong
Abstract：In order to address the issue of insufficient prediction accuracy and poor generalization ability of traditional models for the state of health(SOH)prediction of lithium-ion batteries,a SOH estimation method of lithium-ion batteries based on self attention-causal dilated convolution-gated recurrent unit-autoencoder(SA-CDC-GRU-AE)model was proposed.In the convolution module,CDC module was introduced and combined with the SA mechanism to ensure causality in the prediction and suppress the interference of battery capacity regeneration on the prediction results.Additionally,the AE module was incorporated to optimize the GRU model,enabling it to both extract hidden features and capture long-term dependencies.Validation was performed on two public datasets.The results showed that SA-CDC-GRU-AE model achieved the average values of the root mean square error(RMSE)of 1.009%and 0.488%,and the average values of the mean absolute error(MAE)of 0.780%and 0.432%on the two datasets,respectively.SA-CDC-GRU-AE model could accurately estimate the SOH of lithium-ion batteries and had significant engineering application value for battery management systems.
Keywords：lithium-ion batterySOH estimationcapacity regenerationcausal convolutiondilated convolutionautoencoder
Publication Date：2025-06-20
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：6( 266-271 )