摘要
电池健康状态(SOH)评估是电池管理系统(BMS)中的关键一环,传统的SOH估计方法通常是基于锂离子电池的开路电压、容量或内阻等静态测量参数,然而由于测试过程耗时长、测试环境特殊,在线容量或电阻测量在BMS中很少能实现。从CC放电模式下100%~60%SOC区间内的温度变化速率曲线中提取出一种新的SOH在线评估健康因子(dT/dt)mean。为验证健康因子的合理性,基于集总电池模型,研究电池的生热率发现(dT/dt)mean反应了电池正极熵变特性,以此建立了(dT/dt)mean、可逆反应热与SOH关系的健康寿命预测模型。并基于6组18650 LCO锂离子电池共计706组容量衰退循环数据给出了Pearson相关系数法以及神经网络下的多数据融合法有效性验证,验证结果表明(dT/dt)mean作为SOH在线评估健康因子可以有效提升模型系统预测精度。
State of health(SOH) assessment is a key part of the battery management system(BMS). Traditional SOH estimation methods usually utilize static measurement parameters such as the open circuit voltage, capacity or internal resistance of lithium-ion batteries. However, due to the time-consuming test process and the special test environment, online capacity or resistance measurement is rarely achieved in BMS. In this article, a new SOH online assessment health indicator(dT/dt)mean is extracted from the temperature change rate curve in the 100%~60% SOC interval under CC discharge mode. To verify its reasonability, based on the lumped battery model, the heat generation rate of the battery is studied. Results show that(dT/dt)mean reflects the positive entropy change characteristic of the battery. In this way, a health span prediction model of SOH related to(dT/dt)mean and reversible reaction heat can be formulated. Based on a total of 706 sets of six 18650 LCO lithium-ion battery capacity decline cycle data, this article verifies the effectiveness of Pearson correlation coefficient method and the multi-data fusion method under neural network. Experimental results show that(dT/dt)mean as an online diagnosis health indicator of SOH can effectively improve the prediction accuracy.
作者
石伟杰
王海民
Shi Weijie;Wang Haimin(School of Energy and Power Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China;Shanghai Key Laboratory of Multiphase Flow and Heat Transfer in Power Engineering,Shanghai 200093,China)
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2020年第8期206-216,共11页
Chinese Journal of Scientific Instrument
基金
国家重点研发计划(2018YFB0104400)项目资助
关键词
锂离子电池
健康状态
温度变化速率
可逆反应的熵变
在线诊断
神经网络
lithium-ion batteries
state of health
temperature variation rate
entropy change of reversible reaction
on-line diagnosis
neural network
