摘要
单体锂电池经一致性筛选后以不同串并联方式构成大规模模组以满足动力或储能系统的高电压及大容量应用要求,但现有电池管理技术尚难以检测因电池状态差异引起的模组内过充电及过放电故障,为大规模储能电池安全稳定运行带来了隐患。对此,该文提出了一种基于电化学阻抗谱的并联电池状态差异辨识方法。该方法首先通过设计0~18%健康状态(SOH)差异电池组成模组,对锂电池模组的阻抗特性进行电化学阻抗谱测试;然后基于电化学阻抗谱(EIS)测试结果搭建Randles等效电路模型及拟合模型参数,结合弛豫时间分析确认模型参数准确性并获取等效电路参数,进一步利用Spearman分析其与电池荷电状态(SOC)和SOH相关性,从而提取可用于辩识状态差异的特征参量;最后构建支持向量机模型以实现模组内电池状态差异有效辨识。结果表明,模组内单体电池状态差异会使EIS的实部阻抗在0.01~1 Hz频段差异显著,阻抗虚部在1~100 Hz频段差异明显;Randles等效电路模型具有良好的仿真效果,其中电阻R_(s)、R_(p1)、R_(p2)与常相位元件导纳Y_(1)等四参数与电池SOH具有较强相关性,可作为特征参量,经支持向量机辩识模型可对并联电池状态差异进行快速辨识,准确性可达94.12%。该文通过研究电池阻抗特性,结合模型构建与数据驱动方法实现了对电池并联结构下状态差异的准确判别。该方法可为大规模电池模组内电池状态检测提供理论基础及技术路线,具有较强的理论及工程意义。
Single lithium is used in the form of a battery module in various series and parallel systems after consistency screening to meet high-voltage and significant capacity application demands of power or energy storage systems.Still,during the working process,the battery's solid electrolyte interface(SEI)continuously forms and thickens,contributing to the battery's aging.The existing battery management technology is challenging to identify the module charging and discharging promptly,which can cause thermal runaway accidents,undoubtedly endangering the safe and stable operation of large-scale energy storage.This paper employs a state difference identification method for the parallel battery module based on electrochemical impedance spectroscopy technology.Firstly,this paper aims to design a 0%to 18%state of health difference in the parallel battery modules.The electrochemical impedance spectroscopy of the battery modules is tested.The impedance characteristics of the parallel battery modules are analyzed under various states of charge and state of health.Secondly,Randles'equivalent circuit models are built.Model parameters are fitted based on EIS test results,and relaxation time analysis confirms the accuracy of the model parameters.The suitable equivalent circuit parameters are extracted.Further,Spearman's correlation analysis examines the correlation between the parameters and SOC and SOH.The parameters of strong correlation with SOH and weak correlation with SOC are identified,determining the characteristic parameters for distinguishing the state differences of the parallel battery module.Finally,a support vector machine training model is developed to distinguish between battery states.Measuring experiments were conducted using an electrochemical impedance spectroscopy testing system and a battery charging and discharging cycle system.The batteries were placed in a thermostat to ensure constant temperature conditions.An 18650 ternary lithium-ion battery was selected due to its high energy density and endurance.The samples were pretreated to achieve the best overall performance,using a 0.5C constant current and constant voltage to charge and a 0.2C constant current to discharge five times.The samples were placed in the thermostat for 5 minutes to stabilize the temperature.Then,the samples were measured,and data were collected from 100%SOC to 10%SOC.The conclusions of this study are as follows.(1)The difference in the module can make the real impedance of the EIS in the 0.01~1 Hz band and the imaginary impedance in the 1~100 Hz band show a significant difference.(2)The simulation effect of the Randles equivalent circuit is good,in which R_(s),R_(p1),R_(p2),and the guide Y_(1)have high correlation with SOH,which can be used as feature parameters.(3)The support vector machine identification model can effectively identify the state difference of the parallel battery module,and the accuracy reaches 94.12%.The EIS test of the battery module can reflect the state of the internal single battery.This paper develops a battery equivalent circuit model,calculates and determines the characteristic parameters,and constructs an identification model,providing a theoretical basis and technical route for detecting battery states in large-scale battery modules after work extension.
作者
董明
罗阳
雷万钧
任明
郭安祥
Dong Ming;Luo Yang;Lei Wanjun;Ren Ming;Guo Anxiang(State Key Laboratory of Electrical Insulation and Power Equipment,Xi’an Jiaotong University,Xi’an,710048,China;State Grid Shaanxi Electric Power Co.Ltd,Electric power Research Institute,Xi’an,710005,China)
出处
《电工技术学报》
北大核心
2025年第20期6744-6756,共13页
Transactions of China Electrotechnical Society
基金
国家电网公司科学技术资助项目(4000-202499063A-1-1-ZN)。
关键词
锂离子电池
电化学阻抗谱
阻抗特性
等效电路
特征参量筛选
支持向量机
Lithium-ion battery
electrochemical impedance spectroscopy
impedance characteristics
equivalent circuit
feature parameter screening
support vector machine
