基于改进粒子滤波的锂电池SOH预测  

Lithium-ion battery SOH estimation based on improved particle filter

作　　者：徐超 李立伟 杨玉新 王凯 XU Chao;LI Liwei;YANG Yuxin;WANG Kai(School of Electrical Engineering,Qingdao University;Weihai Innovation Institute,Qingdao Univesity;Library of Qingdao University,Qingdao 266071,Shandong,China)

机构地区：[1]青岛大学电气工程学院 [2]青岛大学威海创新研究院 [3]青岛大学图书馆,山东青岛266071

出　　处：《储能科学与技术》2020年第6期1954-1960,共7页Energy Storage Science and Technology

基　　金：山东省科技发展计划项目(2011GGB01123);山东省重点研发计划项目(2017GGX50114);山东省自然科学基金项目(Y2008F23)。

摘　　要：随着锂离子电池在电动汽车和微电网越来越广泛地使用,人们在保证电池管理系统(BMS)安全可靠运行和降低维护成本方面做了许多研究,电池健康状态(SOH)估计作为锂离子电池管理系统的关键功能之一,准确估计电池当前健康状态对电池管理系统有重要意义。为了提高估计准确性,首先在分析传统布谷鸟搜索优化算法的基础上,提出了一种动态布谷鸟搜索算法,该算法通过改进步长和发现概率,并将函数值变化趋势引入到步长更新方程,平衡了搜索速度和精度之间的关系。为解决传统粒子滤波自身存在的粒子退化问题,通过将粒子用布谷鸟鸟窝表示,对布谷鸟群体搜索的模拟来指导更新粒子的分布,利用改进的动态布谷鸟搜索来优化粒子滤波算法。然后从锂离子电池工作时的可测参数中提取健康指标HI,建立了HI指标与SOH之间的映射模型,并将其应用于状态空间模型的观测,提出了一种基于改进粒子滤波算法的电池SOH估计方法。实验结果表明,该方法优于传统粒子滤波算法(PF),对锂离子电池退化过程预测具有良好的适应性和精确性。With the increasing extensive use of lithium-ion batteries in electric vehicles and microgrid,much research has been performed to ensure a safe and reliable operation and reduce the maintenance costs of the lithium-ion battery management system(BMS).As one of the key functions of the BMS,the state of health(SOH)estimation is very important.A dynamic cuckoo search algorithm is proposed based on the analysis of the traditional cuckoo search optimization algorithm to improve the estimation accuracy.The relationship between search speed and accuracy is balanced by improving the step size and the discovery probability and by introducing the change trend of the function value into the step update equation.A dynamic cuckoo search algorithm is proposed to solve the particle degradation problem existing in the traditional particle filter itself.The particles are represented by cuckoo bird’s nests.Moreover,the cuckoo group search simulation is used to guide the distribution of updated particles,while the improved dynamic cuckoo search is used to optimize the particle filter algorithm.The health index(HI)is extracted from the measurable parameters of the lithium-ion battery,and the mapping model between the HI index and the SOH is established and applied to the state space model observation.A battery SOH estimation method based on the improved particle filter algorithm is proposed.The experimental results show that this method is superior to the traditional particle filter algorithm and has good adaptability and accuracy in predicting the degradation process of lithium-ion batteries.

关 键 词：锂离子电池 粒子滤波算法 粒子退化 改进布谷鸟算法 SOH预测 健康指标 

分 类 号：TM912[电气工程—电力电子与电力传动]