动力电池状态观测及管理系统研究

发布时间：2018-03-29 15:11

  本文选题：电池建模　切入点：荷电状态观测　出处：《中国矿业大学》2017年博士论文

【摘要】：近年来,新能源汽车得到了迅猛发展,磷酸铁锂电池以其优异的特性成为新能源汽车的首选动力源。为了提高复杂工况下新能源汽车的续航里程及载荷安全性能,研究磷酸铁锂电池的状态观测及可靠的管理系统已成为促进新能源汽车发展的关键,具有重要意义。本课题以磷酸铁锂电池管理系统为研究中心,对电池的状态观测及管理控制两个方面的关键技术进行了深入的研究。本文首先介绍了课题的选题背景及意义,综述了与课题研究相关的国内外现状,给出了课题研究的主要内容及安排。分析了磷酸铁锂电池的运行机理,指出影响电池建模准确性的因素;基于电化学思想推导了一种通用非线性等效电气电路模型,揭示了极化及扩散现象的本质表现;同时,对现有数学模型进行了改进,提出了一种考虑库伦效率、温度、劣化、松弛效应、迟滞效应等影响因素在内的优化的数学模型,为电池荷电状态精确观测打下基础。为了降低传统扩展卡尔曼滤波算法的计算复杂度,提高算法稳定性,提出一种基于最优高斯近似卡尔曼滤波器的自适应Sigma卡尔曼滤波荷电状态观测算法,引入状态估计误差的协方差的平方根,提高了状态协方差的半正定性,构建了Sigma采样序列,基于迭代型最小均方误差估测对状态估计及观测估计进行更新,实现了对电池荷电状态的精确估计。针对磷酸铁锂电池的劣化及其剩余寿命预测展开研究,设计了电池容量衰减实验,推导了用于观测的电池容量衰减模型,提出了一种基于磷酸铁锂电池容量衰减模型的自适应Sigma卡尔曼粒子滤波电池剩余寿命预测方法,克服了粒子匮乏的问题,方法简单且有效;为了提高电池剩余寿命的预测准确性,构建了容量变化的数据驱动型自回归模型,提出了一种基于数据驱动型融合模型的自适应Sigma卡尔曼粒子滤波电池剩余寿命预测方法,实验验证了所述方法的有效性及可行性。为了实现电池组内各单体电池的一致性均衡控制,课题从能量的角度展开均衡电路及控制研究,设计了一种基于储能电感的非耗散型二级均衡电路并进行了数学建模,结构简单、便于级联控制,同时给出了MOSFET控制时序,提出了一种自适应均衡电流的控制方案,空载工况、充电工况及效率实验下的测试验证了所提均衡电路及方案的有效性与可行性。为解决动力电池组充电问题,课题从V2G综合性能量变换系统从发,提出一种双极DC-BUS充电系统。针对前级PWM主变换器,提出了一种基于V-S平衡特解规划的通用PWM调制算法,在单个载波周期内,设计了一种通用的开关序列,物理意义更明确,普适性更强;通过特解规划、脉冲分配思想,提出了九段式低频脉动抑制PWM特解规划算法和中点电位闭环控制策略,无需扇区判定及矢量合成分析,算法更简洁,可实现全范围下消除中点电位低频振荡和中点电位直流不平衡功能。同时展开后级充放一体式DC/DC系统拓扑与控制研究,设计了多级恒流与恒压涓流的充电策略,均衡充电与不均衡充电实验验证了所提充电系统及策略的有效性与可行性。最后,以飞思卡尔汽车级嵌入式处理器为核心,高精度电池信息采集芯片为采集单元,设计了一种基于MSCAN总线、Modbus_RTU协议等架构的多通讯网络分布式主从电池管理系统,分别给出硬件电路及软件的模块化设计方法,提高了系统的可靠性,缩短了开发周期。
[Abstract]:In recent years, new energy vehicles has been rapid development of lithium iron phosphate batteries become the preferred power source of new energy vehicles because of their excellent properties. In order to load the mileage and safety performance of new energy vehicles to improve the complex condition, the study of lithium iron phosphate battery state observation and reliable management system has become the key to promote new energy vehicles the development is of great significance. In this paper the lithium iron phosphate battery management system as the research center, state observation and control of the battery management key technology two aspects are studied. This paper firstly introduces the background and significance of the topic, the current situation of the research at home and abroad are reviewed and the topic is given the main content of the research and arrangement. Analyzing the operation mechanism of lithium iron phosphate battery, factors that affect the accuracy of the electrochemical cell modeling is derived based on a thought; General nonlinear equivalent electrical circuit model, reveals the essence of polarization and diffusion phenomena; at the same time, the existing mathematical model was improved, presents an account of Kulun efficiency, temperature, deterioration, relaxation effect, mathematical model of optimization, influence factors such as hysteresis, the battery state of charge make accurate observation under the traditional foundation. In order to reduce the computational complexity of the extended Calman filter algorithm, improve the stability of the algorithm, we propose an approximate optimal adaptive filter based on the Calman Gauss Sigma Calman filter charge state observation algorithm, introducing the state estimation error covariance square root, improve the positive semi definiteness of the state covariance, constructed Sigma sampling series iterative minimum mean square error estimation of state estimation and the observation estimates were updated based on the accurate estimation of battery state of charge. For P Degradation of acid lithium iron battery and residual life prediction research, design of battery capacity attenuation, attenuation model is derived for observation of the battery capacity, provides a method to predict the residual life of Sigma Calman adaptive particle filter cell attenuation model of lithium iron phosphate battery capacity based on particle overcomes the problem of lack of method simple and effective; in order to improve the prediction accuracy of the remaining life of the battery, constructed the capacity change of data driven autoregressive model, proposes a method for predicting remaining life of Sigma adaptive particle filter Calman cell data driven fusion model based on the experiments verify the effectiveness and feasibility of the method. In order to achieve balanced consistency each single cell battery control topic equalization circuit and control from the angle of energy, design a energy storage based on The two sense of non dissipative equalization circuit and mathematical modeling, simple structure, convenient for cascade control, and gives the MOSFET control timing control scheme was proposed, an adaptive equalization charging current of no-load condition, under the experimental condition and the efficiency test verified the feasibility and effectiveness of the equalization circuit and program. To solve the power battery charging problem, research from V2G comprehensive energy conversion system from hair, put forward a kind of DC-BUS bipolar charging system. Aiming at the level PWM converter, presents a general PWM modulation algorithm based on V-S equilibrium solution of planning, in a single period of carrier, designed a switching sequence in general, the physical meaning is more clear, more applicable; special solution of the planning, pulse distribution thought, proposed nine stage low frequency ripple suppression PWM solution planning algorithm and the midpoint potential of closed-loop control strategy without sector Determination of synthetic analysis and vector algorithm, more concise, can realize the full range of the elimination of neutral point potential oscillations and neutral point DC unbalance function. At the same time after the commencement of discharge level research on topology and control of integrated DC/DC system, design the charging strategy multistage constant current and constant voltage trickle, equalizing charge and unbalanced charging experiment the feasibility and effectiveness of the proposed charging system and strategy. Finally, by Freescale Carle automotive embedded processor as the core, high precision acquisition chip for battery information acquisition unit, based on the design of a MSCAN bus, Modbus_RTU protocol architecture of multi communication network distributed battery management system, modular design methods are given the hardware circuit and software, improve system reliability, shorten the development cycle.

【学位授予单位】：中国矿业大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TM912;U469.7
，

本文编号：1681672