浙江电力2026，Vol.45Issue(1)：57-65,9.DOI:10.19585/j.zjdl.202601006
基于SAO-BiLSTM-KAN的电池健康状态估计
SOH estimation for lithium-ion batteries based on SAO-BiLSTM-KAN
摘要
Abstract
To improve the estimation accuracy of the state of health(SOH)for lithium-ion batteries,a novel estima-tion method based on snow ablation optimization,bidirectional long short-term memory network and Kolmogorov-Arnold networks(SAO-BiLSTM-KAN)is proposed.Firstly,data is extracted from the battery charging process,and health features are derived through incremental capacity analysis.These features are then fed into a BiLSTM network to capture long-term dependencies within the time series.Subsequently,the outputs from the BiLSTM are passed to the KAN to explore complex nonlinear relationships among the features,thereby enhancing estimation performance.To achieve optimal results,the SAO algorithm is introduced to optimize the model's hyperparameters.Experimental results demonstrate that the proposed model delivers outstanding performance across various comparative tests,with both the root mean square error(RMSE)and mean absolute error(MAE)for SOH estimation remaining below 0.919%,validating the method's superior prediction accuracy and generalization capability.关键词
锂离子电池/健康状态/BiLSTM/KAN/雪消融优化算法Key words
lithium-ion battery/SOH/BiLSTM/KAN/SAO引用本文复制引用
张彬桥,邹霖,万刚..基于SAO-BiLSTM-KAN的电池健康状态估计[J].浙江电力,2026,45(1):57-65,9.基金项目
湖北省自然科学基金(2022CFD167) （2022CFD167）
