基于KF-BPNN融合算法的电池循环寿命预测方法  

Battery cycle life prediction method based on KF-BPNN fusion algorithm

作　　者：张宁[1] 刘一飞 汤建林 李佳宽 ZHANG Ning;LIU Yi-fei;TANG Jian-lin;LI Jia-kuan(College of Weaponry Engineering,Naval Univ.of Engineering,Wuhan 430033,China)

机构地区：[1]海军工程大学兵器工程学院,武汉430033

出　　处：《海军工程大学学报》2022年第5期39-44,共6页Journal of Naval University of Engineering

基　　金：国家部委基金资助项目(20170588)。

摘　　要：为了解决实际应用过程中电池循环寿命预测精度较低的问题,提出卡尔曼滤波-BP神经网络(KF-BPNN)融合算法对电池的循环寿命进行预测。该方法选用电池内阻作为循环寿命的评估参数,利用BPNN预测电池的内阻值,并将预测内阻值作为KF算法的观测值来修正卡尔曼滤波观测方程系数,从而提高循环寿命预测精度。实验结果表明:融合算法的预测精度有了明显提高。In order to solve the problem of low accuracy for battery cycle life prediction during practical applications, a fusion algorithm employing a Kalman filter(KF) and a back-propagation neural network(BPNN)(KF-BPNN fusion algorithm) was proposed to accurately predict the battery cycle life. The battery life cycle was evaluated using the battery internal resistance whose value was predicted using a BPNN. The predicted resistance values were used as observations in the KF algorithm to update the KF observation coefficients, and thereby improve the accuracy of cycle life prediction. The experimental results show that the prediction accuracy of the KF-BPNN fusion algorithm has been significantly improved.

关 键 词：电池循环寿命 电池寿命预测 内阻 KF-BPNN融合算法 

分 类 号：TM912[电气工程—电力电子与电力传动]