锂动力电池健康度评价与估算方法的研究

发布时间：2018-01-14 05:28

本文关键词：锂动力电池健康度评价与估算方法的研究　出处：《哈尔滨理工大学》2016年博士论文　论文类型：学位论文

【摘要】：锂动力电池能量高、污染低、安全性高,目前已成为新能源汽车的主要动力来源之一。动力电池的健康度(State of health,SOH)是目前电动车动力电池系统中最重要的一项性能指标。电池的SOH与电压、电流、内阻、温度等其他电气参数有别,无法通过设备或仪器采用直接测量法获得,但这些电气参数随着电池的SOH的变化而变化。因此,准确的锂电池健康度状态在线估算问题也成为了电动车动力电池管理系统的关键技术之一。本文针对锂电池健康度估算研究、扩展卡尔曼滤波(Extended Kalman Filter,EKF)算法研究、双卡尔曼滤波(Dual Extended Kalman Filter,D-EKF)在线估算策略及联合卡尔曼滤波(Federated Extended Kalman Filter,F-EKF)综合评价等问题提进行深入研究,具体如下:在基于自放电电阻的电池等效电路模型研究基础上,通过电池循环寿命测试实验结果,得到电池健康度受到内阻、容量、自放电电阻和荷电状态(State of Charge,SOC)等参数影响的结果。结合电池状态参数的特点,以电池戴维南等效电路模型为原型,构建带有电池自放电因子的等效电路改进模型。通过对模型的系统参数辨识,得到电池参数中荷电状态初始值选取的不同会带来模型参数差异的结论。在结合改进的戴维南等效电路模型的基础上,采用安时积分法建立锂电池SOC的状态方程和输出方程,采用扩展卡尔曼滤波算法对系统中的各个参数进行观测,实现在线观测锂电池荷电状态的目标。系统仿真观测结果表明,扩展卡尔曼滤波算法的锂电池荷电状态在线估算具有较好的精度。在扩展卡尔曼滤波算法的锂电池荷电状态在线估算研究的基础上,展开电池容量健康度、内阻健康度和自放电健康度估算研究。根据电池内部状态参数的多样性和复杂性,进行针对内阻、容量和自放电特性的相关实验。实验结果表明,基于EKF算法的电池SOH在线估算方法具有较好精度,可以基本满足电动公交车实际使用工况的需要。针对电池SOH在线估算时电池模型的初始参数无法事先确定导致误差增大的问题,引入D-EKF算法,通过建立两个独立的扩展卡尔曼滤波器,分别估算电池系统状态和参数,并相互更新状态和参数。实际路况试验和数据仿真结果验证D-EKF算法的有效性和精度。针对电池健康度标准和描述不完善的问题,从电池的实际应用和测量学角度,提出“电池广义健康度(Generalized State Of Health,GSOH)”概念,用来评价电池的性能参数。建立带有信息分配因子的电池广义健康度函数结构,设计欧姆内阻健康度、自放电健康度和容量健康度的局部滤波器,并对信息分配因子对主滤波器及子滤波器的分配影响进行研究,并给出一种自适应式概率分配因子的最优分配算法。针对单体电池组成动力电池模块后寿命下降的问题,通过基于失效模式及效应分析法(Failure Mode Effect Analysis,FMEA)对锂电池进行故障诊断和失效分析,提出一种聚类分析法的电池分选办法,得到与电池成组寿命相关的主要因素和电性能参数,提出“电池动态一致性分选”概念,有效提高动力电池的健康度。根据对电动汽车锂动力电池广义健康度的理论分析和大量的实验室及现场试验结果,结合电动公交车实际使用工况,对研究成果进行了验证。实验结果表明,基于F-EKF的锂动力电池在线估算方法够满足锂动力电池SOH估算的精确度要求。
[Abstract]:Lithium battery with high energy, low pollution, high safety, and has become one of the main sources of new energy vehicles. The health of the battery (State of health, SOH) is one of the most important performance indexes of power battery system of electric vehicles at present. SOH and battery voltage, current, temperature resistance, etc. other electrical parameters are different, not through equipment or instrument by direct measurement method, but these electrical parameters change with the SOH of battery change. Therefore, accurate lithium battery health status online estimation problem has also become one of the key technology of the power battery management system of electric vehicle. According to the estimation of the health of the lithium battery. The extended Calman filter (Extended Kalman Filter EKF) algorithm, double Calman filter (Dual Extended Kalman Filter, D-EKF) online estimation strategy and combined filter Calman (Federated Extended Kalman Filter, F-EKF) of comprehensive evaluation provided in-depth study as follows: Based on the study of equivalent circuit model of battery self discharge resistance, through the battery cycle life test results obtained by the health of the battery internal resistance, capacity, self discharge resistance and the state of charge (State of, Charge, SOC etc.) parameters affect the results. According to the characteristics of parameters of batteries, battery Thevenin equivalent circuit model for the prototype, build the equivalent circuit with a battery self discharge factor improved model. Through system parameter identification of the model, get the parameters of the battery charged state of the selection of initial values for different model parameters will bring different conclusions on the basis of Thevenin. The improved equivalent circuit model, the state equation and output equation is established by using lithium battery SOC ah integration method, using extended Calman filter algorithm in the system All of the parameters were observed. The implementation of state of charge of online observation of lithium battery target. Observation results of simulation results show that the extended Calman filter algorithm of lithium battery state of charge of online estimation has good accuracy. In the state of charge of battery extended Calman filter algorithm in line is estimated on the basis of the study, the health of the battery capacity expansion resistance, health and health of self discharge. According to the estimation of battery internal state parameters of the diversity and complexity of the internal resistance, capacity and the self discharge characteristics. The experimental results show that EKF algorithm has better battery SOH online estimation method based on accuracy, can basically meet the needs of electric bus in actual working conditions. The initial parameters of the battery when the battery SOH online estimation model can't confirm cause error problem, D-EKF algorithm is introduced, through the establishment of two Independent of the extended Calman filter, estimated battery system status and parameters, and each update status and parameters. The actual traffic data test and simulation results verify the validity of D-EKF algorithm and accuracy. According to the battery health standards and describe the problems, the perspective of the practical application and measurement from the battery, the battery general Health (Generalized State Of Health, GSOH) "concept, to establish the evaluation of battery performance parameters. The generalized degree function structure with battery health information distribution factor, design the ohmic resistance of health, health and local self discharge filter capacity of health, and also to study the effects on the distribution of information distribution factor on the main filter and the sub filter, optimal allocation algorithm and presents an adaptive probability distribution factor. According to the single battery life down power battery module The problem, through the failure mode and effect analysis method based on (Failure Mode Effect Analysis, FMEA) for fault diagnosis and failure analysis of lithium battery, this battery sorting method of cluster analysis, the main factors associated with the battery life and the electrical performance parameters, put forward the "dynamic consistency of battery sorting" concept. Improve battery health. According to the lithium battery electric vehicle theory analysis of generalized health degree and a large number of laboratory and field test results, combined with the actual conditions of the electric bus, the research results are verified. The experimental results show that the F-EKF lithium battery on-line estimation method based on estimation of lithium power enough to meet the requirements the accuracy of SOH battery.

【学位授予单位】：哈尔滨理工大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：U469.72

【相似文献】