【作者】 杨涛；

【导师】 刘清河；

【作者基本信息】 哈尔滨工业大学 ， 车辆工程， 2015， 硕士

【摘要】 大学生方程式纯电动赛车依托国家新能源汽车发展政策,受到各大高校的青睐。纯电动赛车对轻量化、可靠性和安全性要求很高,但是电池和电池管理技术制约着其发展。通过对纯电动赛车用三元聚合物锂电池(又称镍钴锰酸锂电池,以下简称三元电池)进行研究,设计电池管理系统,能够实时监测电池组的状态信息,掌握电池组的剩余容量,进行有效的均衡管理并对危险状况及时处理,使电池安全、可靠、高效的运行,延长使用寿命。本文以三元电池为研究对象,首先通过充放电试验对电池的开路电压、回弹电压特性进行分析,建立了二阶RC网络电池模型,采用指数拟合的方法对模型中的参数进行辨识。在MATLAB/Simulink软件中实现电池模型并对模型进行脉冲及自定义充放电工况下的仿真。仿真结果与试验数据拟合较好,验证了电池模型及参数辨识的准确性。在对常用的SOC(State of Charge)估算方法进行分析的基础上,选择扩展卡尔曼滤波算法作为本文SOC估算方法。由于扩展卡尔曼滤波算法忽略了Taylor展开式的高阶项造成一定误差,采用有限差分方法修正其高阶项。将有限差分扩展卡尔曼滤波算法与电池模型相结合,选定各主要变量及方程的具体形式,在算法递推作用下进行SOC估算。在MATLAB/Simulink软件中对两种估算算法进行仿真,仿真结果与试验数据进行对比,结果表明基于有限差分扩展卡尔曼滤波算法的SOC估算误差不超过3%,估算精度更高。针对基于端电压中心均衡或基于SOC均衡方法的不足,提出一种将端电压均衡与SOC均衡结合的均衡方法。根据电池充放电特性在不同阶段制定了不同的均衡评判标准,确定均衡的阈值并制定均衡策略。在MATLAB/Simulink软件中建立了基于电池模型的均衡电路,并进行了均衡策略的仿真。对电池的仿真和均衡试验结果表明该方案能有效防止电池过充和过放,提高了能量利用率。为纯电动赛车锂电池组设计了配套的硬件和软件。硬件部分为主从分布式结构,分1个主控模块和6个从控模块。主控模块完成对SOC的估算,总电压、电流的采集,模块间的CAN通信功能。每个从控模块管理16节电池单体,负责单体电压检测、温度检测和均衡控制。软件设计给出了系统各部分的软件流程图。对电池组进行系统性能测试,检测系统采样和估算精度,试验结果表明所设计的电池管理系统能够达到预期的设计目标。更多还原

【Abstract】 Under the guidance of the country’s policy of new energy vehicles, Formula Student pure electric racing car is gaining popularity among colleges and universities. Pure electric racing car requires lightweight design, high reliability and safety, but the battery and battery management technology restrict its development. Through the research of lithium polymer battery, one kind of battery management system was designed to real-time monitor status of the battery pack, grasp the remaining capacity of the battery pack, implementing effective equilibrium management and timely handling of hazardous conditions. Therefore, the batteries can be safe, reliable and efficient operation.In this paper, lithium polymer battery was regarded as the object and its open circuit voltage and rebound voltage were analyzed through charging and discharging experiments. Then one battery model based on second-order RC network was established and the model parameters were identified by using exponential fitting method. The model simulation test was carried out with MATLAB/Simulink by applying pulse current and user-defined current. The simulation results fit the experimental data quite well and the accuracy of the battery model and parameters has been verified.Through the analysis of estimation methods of State of Charge(SOC), the extended Kalman filter(EKF) algorithm was selected as the paper’s SOC estimation method. The EKF algorithm ignores the high-order terms of Taylor expansion, which leads to some errors. The finite difference method was adopted to amend its high-order terms. With the combination of finite difference EKF algorithm(FDEKF) and the battery model, the estimation of SOC was performed. The simulation of EKF algorithm and FDEKF algorithm was carried out with MATLAB/Simulink. Compared to the experimental data, the simulation results show that the estimation error of FDEKF algorithm does not exceed 3% and the FDEKF algorithm has better estimation accuracy.The equalization strategy simply based on cell voltage or SOC has some deficiency. One new equalization strategy combined cell voltage with SOC was studied. Different equalization criteria and thresholds were formulated at different stages of charging or discharging and the equalization strategy was also developed. The equalization circuit was established based on the battery model and the simulation of equalization strategy was carried out with MATLAB/Simulink. The simulation and experiment results show that the strategy can improve energy utilization, prevent battery overcharge and over discharge effectively.Hardware and software were designed for lithium polymer battery management system. The part of hardware is one kind of master-slave distributed structure, which has a master control unit(BMU) and six distributed units(BDU). The functions of estimation of SOC, detection of total voltage and current of the battery pack, CAN communication between modules were completed with BMU. Each BDU manages 16 cells, responsible for detection of cell voltage, temperature and equalization control. The part of software was listed flow charts of the system. Finally, the performance test for the battery pack was conducted to detect the sampling and estimation accuracy of the system. Experiment results show that the battery management system can achieve the desired goals.更多还原