纯电动汽车动力电池荷电状态估计的研究

发布时间：2018-07-15 17:27

【摘要】：随着全世界能源危机和环境污染问题的日益凸显,完全以电池为动力来源的纯电动汽车在节能减排上的优点表现得越来越突出。为此,动力电池也逐渐成为企业和高校的研究热点。在动力电池研究中,荷电状态(State of Charge,SOC)估计方法关系到纯电动汽车的使用性能,是电动汽车控制及策略优化的重要研究基础。本文主要是以磷酸铁锂动力电池为研究对象,建立电池模型,研究电池的SOC估算方法。本文建立了电池充放电测试实验平台,对电池实际容量的测定方法及温度修正方法进行了研究。通过脉冲放电实验,对电池的脉冲响应特征进行分析,为电池模型的建立及模型中的辨识参数提供理论和实验依据。对比分析了常用几种电池等效电路模型的优劣,选定PNGV等效电路模型作为本文SOC估算的电池模型,并建立模型方程。通过不同温度下的开路电压(Open Circuit Voltage,OCV)测定实验获取了OCV与SOC之间的函数关系,并分析了温度因素对OCV的影响。对电池进行不同温度下的HPPC脉冲测试,结合脉冲放电电压曲线并运用最小二乘法对模型参数辨识,获得电池模型参数与温度、SOC之间的关系。搭建了电池仿真模型,通过对自定义混合脉冲充放电实验验证了模型准确性,为SOC估算提供了模型基础。对PNGV电池模型建立了引入温度修正系数的空间状态方程,并在MATLAB环境下实现了基于无迹卡尔曼滤波算法的SOC估算程序。对电池分别进行了恒流放电实验和DST工况循环放电实验,分别从已知算法初始值和未知初始值两种条件下验证了本文SOC估计算法的准确性,并分析了初始值选取对算法收敛性的影响。最后通过与传统的安时积分法的SOC估算结果进行比较,表明无论在恒流模式还是工况模式放电下,基于无迹卡尔曼滤波的SOC估算误差都小于5%,相对于安时积分法具有较高的估算精度,并且能够消除初始值的误差,具有更快的收敛性。
[Abstract]:With the worldwide energy crisis and environmental pollution problems becoming increasingly prominent, the advantages of pure electric vehicles (EVs), which are entirely powered by batteries, have become more and more prominent in terms of energy saving and emission reduction. Therefore, power battery has gradually become the research hotspot of enterprises and universities. In the research of power battery, the method of state of charge (SOC) estimation is related to the performance of pure electric vehicle (EV), and it is an important research foundation of EV control and strategy optimization. In this paper, the battery model is established and the SOC estimation method of the battery is studied. In this paper, a battery charge and discharge test platform is established, and the method of measuring the actual capacity of the battery and the method of temperature correction are studied. The impulse response characteristics of the battery are analyzed through the pulse discharge experiment, which provides theoretical and experimental basis for the establishment of the battery model and the identification parameters in the model. The advantages and disadvantages of several kinds of battery equivalent circuit models are compared and analyzed. PNGV equivalent circuit model is selected as the battery model for SOC estimation in this paper, and the model equation is established. The functional relationship between OCV and SOC was obtained by measuring the open circuit voltage (OCV) at different temperatures, and the influence of temperature factors on OCV was analyzed. The HPPC pulse was measured at different temperatures. The model parameters were identified by using the least square method combined with the pulse discharge voltage curve, and the relationship between the model parameters and SOC was obtained. The battery simulation model is built, and the accuracy of the model is verified by the self-defined mixed pulse charge-discharge experiment, which provides the model basis for SOC estimation. The spatial state equation of PNGV battery model with temperature correction coefficient is established, and the SOC estimation program based on unscented Kalman filter algorithm is implemented in MATLAB environment. Constant current discharge experiments and DST cyclic discharge experiments are carried out for the battery. The accuracy of the proposed SOC estimation algorithm is verified by two conditions: the initial value of the known algorithm and the unknown initial value, respectively. The influence of initial value selection on the convergence of the algorithm is analyzed. Finally, by comparing the SOC estimation results with the traditional amp-hour integration method, it is shown that both the constant current mode and the operating mode discharge, The SOC estimation error based on unscented Kalman filter is less than 5, which has higher estimation accuracy than Anchor integration method, and it can eliminate the error of initial value and has faster convergence.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：U469.72

【相似文献】