电工技术学报  2021, Vol. 36 Issue (24): 5201-5212    DOI: 10.19595/j.cnki.1000-6753.tces.210385
“新能源汽车电驱动系统与充放电技术”专题(特约主编:崔淑梅 教授 程 远 教授) |
一种基于数据驱动与经验模型组合的锂电池在线健康状态预测方法
王萍, 弓清瑞, 张吉昂, 程泽
天津大学电气自动化与信息工程学院 天津 300072
An Online State of Health Prediction Method for Lithium Batteries Based on Combination of Data-Driven and Empirical Model
Wang Ping, Gong Qingrui, Zhang Ji’ang, Cheng Ze
School of Electrical and Information Engineering Tianjin University Tianjin 300072 China
全文: PDF (2109 KB)   HTML
输出: BibTeX | EndNote (RIS)      
摘要 锂离子电池健康状态(SOH)的准确估计是电池管理系统(BMS)的关键技术。该文提出一种基于数据驱动与经验模型组合的在线SOH预测方法。通过电池容量增量分析(ICA),找出与SOH相关性较高的两个电压升片段下所耗时间作为电池外部健康特征(HF),并使用高斯过程回归(GPR)的方法建立电池老化的数据驱动模型。利用数据驱动模型对电池工作初期的SOH进行预测,并使用预测值拟合指数经验模型。之后,电池各循环下的SOH用指数经验模型来预测,并且每隔固定循环次数使用观测器对指数模型参数进行一次修正,以保证电池SOH预测的准确性。实验结果表明,该文所提的方法可以在减轻电池监测设备负担的前提下将预测精度保持在较高水平。
服务
把本文推荐给朋友
加入我的书架
加入引用管理器
E-mail Alert
RSS
作者相关文章
关键词 锂离子电池健康状态指数模型高斯过程回归观测器    
Abstract:The accurate estimation of the state of health (SOH) of lithium-ion batteries is a key technology in battery management system (BMS). This paper presents an online SOH prediction method combined with data-driven and empirical models. Through incremental capacity analysis (ICA), the time spent under two voltage rise segments with high correlation with SOH was identified as the external health factor (HF) of the battery, and a data-driven model of battery aging was established using gaussian process regression (GPR). The data-driven model was used to predict the SOH of the battery in the initial working period, and the exponential empirical model was fitted with the predicted values. Then the SOH of the subsequent cycle was predicted by the exponential model. And the parameters of the exponential model were modified by the observer every fixed number of cycles to ensure the accuracy of SOH prediction. The experimental results show that the proposed method can keep the prediction accuracy at a high level while reducing the burden of battery monitoring equipment.
Key wordsLithium-ion battery    state of health    exponential model    Gaussian process regression    observer   
收稿日期: 2021-03-22     
PACS: TM912  
基金资助:国家自然科学基金资助项目(61873180)
通讯作者: 程 泽 男,1959年生,教授,博士生导师,研究方向为现代电力电子技术等。E-mail: chengze@tju.edu.cn   
作者简介: 王 萍 女,1959年生,教授,博士生导师,研究方向为电能质量、智能检测与控制等。E-mail: pingW@tju.edu.cn
引用本文:   
王萍, 弓清瑞, 张吉昂, 程泽. 一种基于数据驱动与经验模型组合的锂电池在线健康状态预测方法[J]. 电工技术学报, 2021, 36(24): 5201-5212. Wang Ping, Gong Qingrui, Zhang Ji’ang, Cheng Ze. An Online State of Health Prediction Method for Lithium Batteries Based on Combination of Data-Driven and Empirical Model. Transactions of China Electrotechnical Society, 2021, 36(24): 5201-5212.
链接本文:  
https://dgjsxb.ces-transaction.com/CN/10.19595/j.cnki.1000-6753.tces.210385          https://dgjsxb.ces-transaction.com/CN/Y2021/V36/I24/5201