VRLA电池的SOC估计与其模型参数辨识研究

发布时间：2018-02-02 05:30

本文关键词： VRLA电池电荷状态系统辨识　出处：《北京工业大学》2016年硕士论文　论文类型：学位论文

【摘要】：电池的剩余容量是蓄电池的重要性能指标之一,也是判断电池,是否发生故障以及失效的最可靠的参数,通常用电荷状态SOC来表示(state of charge)。然而由于蓄电池的SOC,不能通过直接测量得到,因此在工程上,常通过监测蓄电池的内阻或测量开路电压,来对蓄电池进行日常维护。但由于蓄电池在不同的工作环境和不同的失效模式下,其内阻和电压的表现形式也不一样。本文重点研究了VRLA电池的电荷状态SOC与内阻和电压之间的关系,以及通过系统辨识来估计VRLA电池的参数。首先介绍蓄电池和VRLA电池的发展及其工作原理,以及到目前为止,人们关于VRLA电池SOC估计研究现状,研究VRLA电池的常见失效模式,及其相应的应对措施,同时比较VRLA电池的几种常用的物理模型以及各模型在电池性能研究方面的优劣。之后,研究系统辨识的相关知识,以及通过系统辨识来间接估计VRLA电池的SOC的方法。其次,主要针对公达数码公司提供的VRLA电池设计了一系列的充放电实验来获取VRLA电池的充放电特性曲线,同时证明了单独通过电池的电压或单独通过电池的内阻来判断电池是否失效是不可靠的。利用这些实验数据通过MATLAB仿真并拟合出VRLA电池的电荷状态SOC分别与电池的端电压,开路电压,和内阻之间的关系,以及电池的SOC与内阻和端电压的关系,电池的SOC与内阻和开路电压之间的关系。通过拟合优度筛选出最优电池的SOC与各参数数学模型,为VRLA电池的失效性诊断和日常维护提供理论依据。同时为搭建合理的VRLA等效电池模型提供参考。再次,考虑到在实际工程中的应用,建立了VRLA电池的二阶RC电路模型,设计了VRLA电池模型的辨识输入信号,基于模型推导出了电池的开路电压,内阻与模型参数之间的关系,采用基于遗忘因子最小递推二乘算法,估计了电池模型的相关参数,最后通过电压验证模型的正确性。其优点在于避免了以往测量开路电压时需要静止很长时间,而且此方法也可以应用在线监测,在检测仪器方面来说,系统辨识所需的设备成本比利用内阻测试仪来测量电池参数的成本低。最后,开发了VRLA电池监测系统软件,开发工具使用的是Microsoft Visual Studio 2012,结合SQL SERVER2012数据库,实现对VRLA电池的性能监测。软件主要是通过测量的数据分析并拟合出可以判断VRLA电池失效的数学公式,为VRLA电池的失效性判断提供依据。
[Abstract]:The residual capacity of battery is one of the most important performance indexes of battery, and it is also the most reliable parameter to judge whether the battery is in trouble or not. The charge state SOC is usually used to denote the state of charge. However, since the SOC of the battery cannot be obtained by direct measurement, it is therefore in engineering. The battery is usually maintained by monitoring the internal resistance of the battery or measuring the open circuit voltage. However, the battery is in different working environment and different failure mode. The forms of internal resistance and voltage are also different. This paper focuses on the relationship between charge state SOC and internal resistance and voltage of VRLA cells. The parameters of VRLA battery are estimated by system identification. Firstly, the development and working principle of battery and VRLA battery are introduced. The current situation of SOC estimation of VRLA batteries, the common failure modes of VRLA batteries and the corresponding countermeasures are studied. At the same time, several common physical models of VRLA battery and the advantages and disadvantages of each model in battery performance research are compared. Then, the related knowledge of system identification is studied. And the method of indirectly estimating the SOC of VRLA battery by system identification. Secondly. A series of charge-discharge experiments are designed to obtain the charge-discharge characteristic curve of VRLA battery. It is also proved that it is not reliable to judge the failure of the battery by the voltage of the battery alone or by the internal resistance of the battery alone. The charge of the VRLA battery is simulated and fitted by MATLAB using these experimental data. The status SOC is associated with the terminal voltage of the battery. The relationship between open circuit voltage and internal resistance, and the relationship between SOC of battery and internal resistance and terminal voltage. The relationship between SOC and internal resistance and open circuit voltage. The mathematical models of SOC and each parameter of the optimal battery were selected by the goodness of fit. It provides a theoretical basis for the failure diagnosis and daily maintenance of VRLA batteries, and provides a reference for setting up a reasonable VRLA equivalent battery model. Thirdly, considering the application in practical engineering. The second-order RC circuit model of VRLA battery is established, and the identification input signal of VRLA battery model is designed. Based on the model, the relationship between open circuit voltage, internal resistance and model parameters is derived. The parameters of the battery model are estimated by using the least recursive two-multiplication algorithm based on the forgetting factor. Finally, the correctness of the model is verified by voltage. The advantage of the model is that it avoids the need for a long time of static when measuring open circuit voltage in the past, and this method can also be used in on-line monitoring, in terms of measuring instruments. The equipment cost of system identification is lower than that of using internal resistance tester to measure battery parameters. Finally, the software of VRLA battery monitoring system is developed. The development tool uses Microsoft Visual Studio 2012, combined with the SQL SERVER2012 database. The software is mainly based on the analysis of the measured data and fitting out the mathematical formula for judging the failure of VRLA battery, which provides the basis for judging the failure of VRLA battery.
【学位授予单位】：北京工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TM912

【相似文献】