基于RBF-BSA的锂离子电池SOC混合估算算法

State of charge estimation of lithium-ion battery based on hybrid algorithms of RBF-BSA

为提高锂离子电池荷电状态(SOC)预测精度,提出利用回溯搜索算法(BSA)优化径向基函数(RBF)神经网络的输出权值与阈值的混合算法.通过对锂电池模型中的目标函数进行优化求解,并寻找最佳的目标权值和阈值降低预测误差,提高了RBF网络模型的预测精度.为验证算法的有效性,搭建锂离子电池的充放电实验平台获取数据并对网络进行验证,实验结果表明:混合算法相比标准RBF网络算法具有更好的SOC预测精度,并将网络输出预测误差降低到2%以内,符合锂电池荷电状态估算要求.

A hybrid algorithm was proposed to improve the prediction accuracy of state of charge(SOC)estimation of lithium battery by optimizing the output weight and threshold of the radial basis function(RBF)neural network using backtracking search algorithm(BSA).The prediction accuracy of the RBF network model was improved by optimizing the target function in the lithium-ion battery model and finding the optimal value of the target weight and threshold to reduce the prediction error.To assess the effectiveness of the algorithm,the experimental platform for charging and discharging lithium ion batteries was built to obtain data and verify the network.The experiment results show that the hybrid algorithm has better SOC prediction accuracy than the standard RBF network algorithm,and the prediction error of the network requirements of lithium-ion battery.