节点文献
基于改进单粒子模型的锂离子电池组充电策略研究
Research on Charging Strategy for Lithium-ion Battery Pack Based on Improved Single Particle Model
【作者】 刘璇;
【作者基本信息】 哈尔滨工业大学 , 电气工程, 2016, 硕士
【摘要】 锂离子电池以其优越的性能逐步成为新能源领域的核心储能部件,与此同时对锂离子电池管理系统也提出了更高的要求,高效、健康的电池充电方法受到越来越多的关注。本文开展了锂离子电池组充电策略的研究,提出基于电池模型参数的长寿命快速充电方法,结合机理模型实现电池组均衡控制,为电化学模型应用于管理系统提供了技术支持。首先,详细叙述了锂离子电池的改进单粒子模型,给出了机理模型中各个物理、化学反应过程的相关表达式。在此基础上,分析模型参数与电池外在行为之间的联系,利用激励响应分析的方法,通过不同的辨识工况实现模型参数的分步获取。通过商品钴酸锂电池的参数获取实验,验证参数获取的有效性和精度。实验表明参数获取结果准确,可用于电池状态参量的估计以及电池管理的研究中。为实现锂离子电池快速且安全的充电,详细分析了锂离子电池充电老化机理,在传统充电方式的基础上,基于改进单粒子模型并结合模型参数,从电池健康管理的角度出发,设计出一种监测并控制负极活性粒子表面嵌锂率的充电方式。搭建充电实验平台实现锂离子电池长寿命快速充电,实验数据指出该充电方式具备较高充电速度和深度的同时,能够延长电池寿命。针对当前电池组均衡技术中出现的问题,结合改进单粒子模型确定了电池组均衡控制方案。基于改进单粒子模型并利用扩展卡尔曼滤波算法估计电池SOC,分别利用恒流工况和动态工况对SOC估计的精确度进行评估,讨论不确定因素如初始偏差、量测噪声对SOC估计的影响。结果表明该估计方法具有较高的精度和较强的抗干扰能力,该估计结果用来作为电池组均衡的数据依据。与此同时,设计出一种能量非耗散型电感均衡电路拓扑结构,对电路的工作原理进行了详细的分析,确定SOC均衡的具体实现方法。最后,采用聚类算法筛选出性能相近的电池,对于本文设计的长寿命快速充电方法、电池组均衡控制方案,在搭建的电池组充电实验系统平台上进行实验,验证整个充电策略的可行性和有效性,实验结果表明该充电策略能够有效地延缓电池组容量衰减。
【Abstract】 Lithium-ion battery has been widely used in the energy industry as a kind of core energy storage device due to its advantages. The battery management system has been put forward higher demands, and the efficient and healthy charging method of lithium-ion battery has drawn more and more attention. This research focus on charging strategy of lithium-ion battery pack, and a long-life and quick charging mode and a equalization control scheme based on the improved single particle model are proposed, which provide ideas for the electrochemistry model in the application of the management system.Firstly, the internal physical or chemical process of the improved single particle model is analyzed in detail. Afterwards, the relationship between model parameters and the cell terminal voltage is deeply excavated. Different charge-discharge conditions are adopted to identify model parameters step by step using the excitation response analysis. The speed and result accuracy of identification method can be verified via the experiment of LiCoO2 batteries. And the precise results can be used in study of state estimation and health management.In order to achieve the lithium-ion battery charging quickly and safely, the ageing mechanism in the process of charging is expounded systematically in this paper. On the foundation of the traditional charging mode, a long-life and quick charging mode is proposed by monitoring solid phase lithium-ion concentration ratio in the surface of negative electrode particles, which combines with the improved single particle model and parameters. Then the charging experiment is carried out on the experimental platform. The result shows that the method improves the charging efficiency and depth, and prolongs the battery useful cycle life evidently.According to the disadvantages of previous schemes in battery pack equalization technology, an equalization control scheme on the basis of battery SOC estimation is developed in this paper. After acquiring the parameters of the improved single particle model, the extended Kalman filter algorithm is apply to estimate SOC. The estimated SOC shows good agreement with the real ones under various operating scenarios. Applicability and robustness of the proposed method are also assessed considering the influencing factors such as initial errors and measured noise. Meanwhile, the non-dissipative equalization topology based on inductor is designed. With the analysis of working principle of the circuit, the equalization process is determined concretely.At last, the experiment system platform is built for conducting charging test including the long-life and quick charging mode and the equalization control scheme for lithium-ion battery pack. The batteries with higher consistency are selected by clustering algorithm to guarantee satisfactory reliability and precision of the experiment. Research results provide feasibility and validity of the charging strategy for lithium-ion battery pack, and weaken the battery pack usage capacity degeneration.
【Key words】 Lithium-ion Battery Pack; Improved Single Particle Model; Long-life and Quick Charging; SOC Estimation; Equalization;