基于UKF的锂离子电池SOC估算方法

发布时间：2018-09-12 20:45

【摘要】：锂离子电池以其端电压高、自放电率低、能量体积比大、无记忆效应等优良性能得到了广泛应用。本文主要研究锂电池，重点研究锂电池的工作状态，并通过电池荷电状态(SOC, State of Charge)来描述。电池荷电状态是电池管理系统的重要指标，对电池管理系统有重要意义。根据电池荷电状态采取调控措施，就可以在充放电过程中保护电池组，避免发生过充、过放而对电池造成损害。一般电池放电较为平稳，电池荷电状态的计算容易且准确，但在特殊环境下，如电流、环境温度变化较大的场合，会使得电池工作过程中电压波动剧烈，难以获取准确的电池荷电状态。目前为止电池荷电状态的获取已有多种估算方法，估算结果优缺点各异。为提高估算准确性，估算过程中滤波方法的选取已成为当前电池荷电状态研究的一个重要分支。研究国内外现状可知，获得准确的SOC表达式、建立合适的电池模型、选择合适的SOC估算算法是准确预测电池荷电状态的三个关键因素。基于以上分析，本文针对SOC估算的主要研究工作如下： ①研究一种动力学电池模型（KiBaM模型），该模型能形象描述电池的可用电荷和束缚电荷，通过对模型进行仿真，验证电池容量的非线性特性和电池的自恢复效应特性。 ②研究几种常用的电池等效模型并进行对比，综合考虑模型的计算量、精度、适用范围后选取Thevenin模型。将Thevenin模型与上述KiBaM模型结合得到一种综合Thevenin模型，通过对模型参数辨识后进行仿真，验证模型精度。 ③研究电池剩余容量SOC常用的几种估算方法及各自优缺点，通过分析电池SOC的影响因素，分别研究环境温度、充放电倍率、循环次数、电池内阻与荷电状态的关系，得到修正后的安时计量法。 ④研究无迹卡尔曼滤波（UKF）法，，及其核心思想UT变换，利用UKF滤波原理建立SOC估算模型，并在Matlab中仿真，以上述修正后的安时计量法估算结果为参考值，验证UKF估算电池SOC的有效性。本文以KiBaM模型结合Thevenin模型得到的综合模型为基础，利用UKF滤波算法估算电池SOC，以改进后的安时计量法为对比，仿真验证UKF能有效估算电池SOC。
[Abstract]:Lithium ion batteries have been widely used for their high terminal voltage, low self-discharge rate, large energy volume ratio and no memory effect. In this paper, we mainly study lithium battery, especially the working state of lithium battery, and describe it by (SOC, State of Charge). The charged state of battery is an important index of battery management system, which is of great significance to battery management system. According to the charge state of the battery, the battery can be protected from overcharge and overdischarge during the charging and discharging process. In general, the discharge of the battery is relatively stable, and the calculation of the charged state of the battery is easy and accurate. However, in the special environment, such as the large change of the current and the ambient temperature, the voltage fluctuates sharply during the operation of the battery. It is difficult to obtain accurate battery charge state. Up to now, there are many methods to estimate the charged state of the battery, and the advantages and disadvantages of the estimation results are different. In order to improve the accuracy of estimation, the selection of filtering methods in the estimation process has become an important branch of the current battery charge state research. By studying the current situation at home and abroad, it is known that accurate SOC expression, establishment of appropriate battery model and selection of appropriate SOC estimation algorithm are three key factors to accurately predict the charged state of the battery. Based on the above analysis, the main research work of SOC estimation in this paper is as follows: 1 A dynamic battery model (KiBaM model), which can describe the available and bound charges of the battery vividly, is simulated. The nonlinear characteristics of the battery capacity and the characteristics of the self-recovery effect of the battery are verified. 2 several commonly used battery equivalent models are studied and compared. The Thevenin model is selected after considering the calculation amount and accuracy of the model. A comprehensive Thevenin model is obtained by combining the Thevenin model with the KiBaM model mentioned above. The model accuracy is verified by simulation after the model parameter identification. 3 several methods of estimating the battery residual capacity SOC and their respective advantages and disadvantages are studied. By analyzing the influence factors of battery SOC, studying the relationship between ambient temperature, charge / discharge rate, cycle times, battery resistance and charged state, the modified amperometric method is obtained. 4 the unscented Kalman filter (UKF) method is studied. And its core idea UT transform, using UKF filter principle to establish SOC estimation model, and simulation in Matlab, using the modified amperometric method as the reference value, to verify the effectiveness of UKF estimation of cell SOC. Based on the synthetic model of KiBaM model combined with Thevenin model, this paper uses UKF filter algorithm to estimate the SOC, of battery, compared with the improved amperometric method. The simulation results show that UKF can effectively estimate the SOC. of battery.
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM912

【参考文献】