VRLA蓄电池SOC估算策略的研究

发布时间：2018-07-21 11:13

【摘要】：蓄电池以其容量大、成本低、安全性好、技术成熟、原料丰富以及免维护等优点,广泛应用于光伏发电系统、风力发电系统、不间断电源系统(UPS)、照明以及电动汽车等领域,是应用最为广泛的二次电池,至今尚未有任何一种电池能够完全取代它。为了提高蓄电池的工作效率,有效延长使用寿命,就必须对蓄电池的剩余容量或荷电状态(SOC)进行准确的估算。而SOC的估算需建立电池等效电路模型,选择合适的估算方法,因此本文针对阀控式密封铅酸(VRLA)蓄电池,重点研究SOC的估算并围绕以下几个方面展开：首先分析了VRLA蓄电池的工作原理及特性,并阐述了充放电倍率、环境温度、电池健康状态等因素对SOC的影响：重点介绍了国内外常见的SOC估算方法,主要有：安时法、电动势法、神经网络法、模糊法以及卡尔曼滤波算法；重点分析比较了几种常用的电池等效电路模型,选取改进的PNGV模型作为蓄电池SOC估算的模型,对该模型采用HPPC脉冲实验进行参数辨识,并通过cftool工具拟合出模型的各个参数。通过对比分析采用两种方法来估算SOC：一种是改进的安时法结合电动势法。由于安时法只受到蓄电池工作电流以及充放电倍率的影响,可以弥补电动势法对电动势模型依赖的缺陷；而安时法存在的积累误差问题也可以通过电动势法得以校正。因此,本文是将电动势法与经过修正的安时法通过并联加权的结构来对SOC进行估算,从而两种方法可以优势互补,提高SOC的估算精度。另一种方法是扩展卡尔曼滤波(EKF)算法,根据改进的PNGV模型,建立电池系统的状态方程和观测方程,确定EKF估算SOC的步骤,实现SOC最小均方差估计。采用MATLAB进行离线仿真,其估算结果接近于理论值。EKF估算精度高,在蓄电池SOC估算领域有着广阔的应用前景,因此进一步实现基于EKF估算SOC的工程化很有必要。
[Abstract]:Battery is widely used in photovoltaic power system, wind power system, uninterruptible power supply system (UPS), lighting and electric vehicle for its advantages of large capacity, low cost, good safety, mature technology, abundant raw materials and no maintenance. Is the most widely used secondary battery, so far no battery can completely replace it. In order to improve the working efficiency and prolong the service life of the battery, it is necessary to estimate the residual capacity or the state of charge (SOC) accurately. But the SOC estimation needs to establish the battery equivalent circuit model, selects the suitable estimation method, therefore this article aims at the valve type sealed lead acid (VRLA) battery, This paper focuses on the estimation of SOC and focuses on the following aspects: firstly, the working principle and characteristics of VRLA battery are analyzed, and the charge-discharge rate and ambient temperature are described. The influence of battery health on SOC: the SOC estimation methods at home and abroad are introduced, such as ampere-hour method, electromotive force method, neural network method, fuzzy method and Kalman filter algorithm. Several commonly used equivalent circuit models of battery are analyzed and compared emphatically. The improved PNGV model is selected as the model of battery SOC estimation. The parameters of the model are identified by HPPC pulse experiment, and the parameters of the model are fitted by cftool tool. Two methods are used to estimate SOC through comparative analysis: one is an improved ampere-hour method combined with electromotive force method. Because the amperage method is only affected by the operating current of the battery and the charge / discharge rate, it can make up for the defect of the electromotive force method which depends on the electromotive force model, and the accumulated error problem of the ampere-hour method can also be corrected by the electromotive force method. Therefore, in this paper, the EMF method and the modified ampere-hour method are used to estimate SOC through parallel weighted structure, so that the two methods can complement each other and improve the precision of SOC estimation. Another method is extended Kalman filter (EKF) algorithm. According to the improved PNGV model, the state equation and observation equation of battery system are established, and the steps of estimating SOC by EKF are determined, and the minimum mean-variance estimation of SOC is realized. The off-line simulation with MATLAB shows that the estimation results are close to the theoretical value. EKF has high precision and has a broad application prospect in the field of SOC estimation of batteries. Therefore, it is necessary to further realize the engineering of SOC estimation based on EKF.
【学位授予单位】：宁夏大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TM912

【相似文献】