基于支持向量回归的锂电池健康状态估计  

Estimation of Lithium Battery State of Health Based on Support Vector Regression

作　　者：张新锋 饶勇翔[1,2] 姚蒙蒙 ZHANG Xin-feng;RAO Yong-xiang;YAO Meng-meng(Key Laboratory of Automotive Transportation Safety Security Technology of the Ministry of Communication,Chang’an University,Xi’an 710064,China;School of Automobile,Chang’an University,Xi’an 710064,Chiha)

机构地区：[1]长安大学汽车运输安全保障技术交通行业重点实验室,陕西西安710064 [2]长安大学汽车学院,陕西西安710064

出　　处：《中北大学学报（自然科学版）》2019年第6期511-516,536,共7页Journal of North University of China(Natural Science Edition)

基　　金：中央高校基本科研业务费专项资金(CHD2012JC048,72105473);长安大学基础研究支持计划专向基金;汽车运输安全保障技术交通行业重点实验室开放基金资助

摘　　要：针对锂电池动态工况下健康状态估计困难的问题,设计了一种基于支持向量回归机的健康状态估计方法.提取电池运行时可监测的电压、电流、温度、荷电状态融合成一种新的健康因子,采用支持向量回归机的方法训练得到健康状态估计模型,并选用网格寻优算法对模型的参数进行优化,实现基于可监测参数的动态工况下的锂电池健康状态估计.仿真结果表明,本文选取的健康因子能准确地反映电池的健康状态,健康状态的平均估计精度在1%以内.In order to solve the problem of state-of-health estimation of lithium battery under dynamic operating conditions,a state-of-health estimation method based on support vector regression machine was designed.The voltage,current,temperature and state of charge that can be monitored while the battery is running were extracted and fused into a new health factor.The state-of-health estimation model was trained by support vector regression machine,and the parameters of the model were optimized by grid optimization algorithm to realize the state-of-health estimation of lithium battery under dynamic condition based on the monitoring parameters.The simulation results show that the health factors selected can accurately reflect the state-of-health of batteries,and the average estimation accuracy of state-of-health is less than 1%.

关 键 词：锂电池 健康因子 支持向量回归 参数优化 健康状态估计 

分 类 号：TM912[电气工程—电力电子与电力传动]