基于改进红狐算法的粒子滤波锂电池剩余寿命预测  

Residual life prediction of lithium battery with particle filter based on improved Red Fox algorithm

作　　者：孟冠军[1] 李国强 彭裕博 MENG Guanjun;LI Guoqiang;PENG Yubo(School of Mechanical Engineering,Hefei University of Technology,Hefei 230009,China)

机构地区：[1]合肥工业大学机械工程学院,安徽合肥230009

出　　处：《电子元件与材料》2025年第2期184-191,200,共9页Electronic Components And Materials

摘　　要：为了进一步提高锂电池剩余寿命预测精度,提出了一种基于主动探测的红狐粒子滤波算法的在线容量估计方法。在标准红狐优化算法的基础上,引入主动搜索策略:采用混沌遍历技术在搜索空间产生探测参考点,在参考点主动探测可行解,从而改善全局搜索能力;再将采样粒子转移到高似然区域中,降低粒子贫化问题。在建立电池容量退化模型中,从放电曲线中提取固定时间(t=2000 s)的特征电压作为健康特征,建立基于皮尔逊相关分析的容量模型。最后利用NASA数据集设计对比实验进行验证。同时将该方法与标准粒子滤波算法、无迹粒子滤波算法和扩展粒子滤波算法的预测结果进行对比,本文提出的方法平均误差降低35%以上。结果表明,在锂离子电池退化过程的状态估计场景中,该方法具有较优的准确性和鲁棒性。To further improve the prediction accuracy of the remaining life of lithium batteries,an online capacity estimation method was proposed based on an active detection and Red Fox particle filter algorithm.Based on the standard Red Fox optimization algorithm,an active search strategy was introduced.Chaotic traversal technology was used to generate detection reference points in the search space and actively detected feasible solutions at these reference points,which enhanced the global search capability,and then the sampled particles were transferred to the high likelihood region to reduce the particle depletion problem.To establish the battery capacity degradation model,the characteristic voltage at a fixed time(t=2000 s)was extracted from the discharge curve as a health feature,and a capacity model was then established based on Pearson correlation analysis.Finally,comparative experiments were designed using the NASA dataset for verification.Compared with the prediction results among the standard particle filter algorithm,the unscented particle filter algorithm and the extended particle filter algorithm,the average error of the proposed method is reduced by more than 35%.The results show that the proposed method exhibits higher accuracy and robustness in the state estimation for lithium-ion battery degradation process.

关 键 词：锂离子电池 剩余寿命预测 粒子滤波 红狐优化算法 

分 类 号：TM912[电气工程—电力电子与电力传动]