基于无迹卡尔曼滤波的动力锂电池SOC估计与实现

发布时间：2018-07-14 19:41

【摘要】：随着能源危机和环境污染的问题日益严重,各国政府对零排放和新能源的电动汽车的研发越来越重视。对电池状态进行控制和管理的电池管理系统是电动汽车发展需要突破的关键技术之一,准确的电池荷电状态(State Of Charge, SOC)估算是电池管理系统运行的前提和关键,对电池使用寿命的提高和整车性能的提升都具有重要意义。为此本文开展动力锂电池SOC的估计研究,主要内容如下：首先,介绍了锂电池SOC估计的背景及意义,分析了SOC的估计现状、定义以及影响因素。在了解动力锂电池的工作原理基础上,考虑工程实现的难易以及数学算法可以弥补等效模型的精确性,选择将内阻等效电路模型作为锂离子电池的动力模型,此后进行开路电压和SOC关系标定以及内阻辨识实验获得电池模型参数并验证该模型能较好的模拟电池特性。其次,由于电池等效模型的开路电压与SOC关系是高度非线性的函数,无迹卡尔曼滤波算法相比扩展卡尔曼滤波在解决非线性非高斯随机系统的状态问题有更好的估计精度。为此本文基于电池的内阻模型,采用基于无迹卡尔曼滤波算法实现非线性条件下锂电池SOC的估算。该算法将电池模型的内阻和SOC作为状态参数,通过无迹变换来处理均值和协方差的非线性传递,在此基础上采用卡尔曼滤波的框架,完成锂电池SOC的估算的方法。通过对自定义充放电工况的SOC变化进行了MATLAB估算仿真实验,结果证明无迹卡尔曼滤波在该模型下能很好估算电池SOC,同时弥补模型的误差。最后搭建系统硬件平台,该平台主要有STM32最小系统、充放电保护电路、数据采集电路与CAN通讯的硬件电路设计。在IAR编译环境下设计系统软件程序,完成了电池组电压、电流、温度、以及SOC的估算各个模块的软件编程。通过实验验证系统的采集数据测量精度以及SOC估算精度。图41幅,表4个,参考文献60篇。
[Abstract]:With the problem of energy crisis and environmental pollution becoming more and more serious, governments pay more and more attention to the research and development of electric vehicles with zero emissions and new energy sources. The battery management system which controls and manages the battery state is one of the key technologies that need to be broken through in the development of electric vehicle. Accurate estimation of the state of charge (SOC) is the premise and key to the operation of the battery management system. It is of great significance for the improvement of battery life and the improvement of vehicle performance. The main contents of this paper are as follows: firstly, the background and significance of SOC estimation for lithium batteries are introduced, and the status quo, definition and influencing factors of SOC estimation are analyzed. On the basis of understanding the working principle of power lithium-ion battery and considering the difficulty of engineering and the mathematical algorithm which can make up for the accuracy of equivalent model, the equivalent circuit model of internal resistance is chosen as the dynamic model of lithium ion battery. Then the open circuit voltage and SOC relationship calibration and internal resistance identification experiments were carried out to obtain the parameters of the battery model and verify that the model can better simulate the characteristics of the battery. Secondly, because the open circuit voltage of the battery equivalent model is a highly nonlinear function, the unscented Kalman filter has better estimation accuracy than the extended Kalman filter in solving the state problem of nonlinear non-Gao Si stochastic systems. In this paper, based on the internal resistance model of the battery, the unscented Kalman filter algorithm is used to estimate the SOC of the lithium battery under the nonlinear condition. In this algorithm, the internal resistance and SOC of the battery model are taken as state parameters, and the nonlinear transfer of mean value and covariance is processed by unscented transformation. Based on this, the estimation method of SOC of lithium battery is completed by using Kalman filter framework. Based on the simulation experiment of SOC estimation based on MATLAB, the results show that the unscented Kalman filter can estimate the SOC of the battery well under the model, and make up the error of the model at the same time. Finally, the hardware platform of the system is built. The platform mainly includes STM32 minimum system, charge-discharge protection circuit, data acquisition circuit and can communication hardware circuit design. The software program of the system is designed under IAR compiling environment, and the software programming of each module of battery pack voltage, current, temperature and SOC estimation is completed. The measurement accuracy and SOC estimation accuracy of the system are verified by experiments. There are 41 figures, 4 tables and 60 references.
【学位授予单位】：中南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM912

【参考文献】