灰色新陈代谢预测模型与车用动力锂离子电池组SoC准确在线估计

In order to extend the life of vehicle power battery and to improve the efficient management of energy and vehicle safety, it is necessary to accurately estimate the power batteries' stage of charge (SoC), which is the core and the difficulty of the vehicle battery management system. The project aims to explore the application of grey metabolic prediction model theory to solve the problem in the SoC online estimates of power lithium-ion batteries for vehicles. The proposed research will be carried out in the following aspects. Firstly, a prediction model of SoC, that does not dependent on the cell model parameters, is intended to be built based on grey metabolic prediction model. Secondly, quantitative analysis of the influence factors of SoC estimates are undertook based on the grey relation principle, and then the SoC of multidimensional grey prediction model of cell/battery are established. Thirdly, the SoC estimates prediction models of batteries will be established under constant discharge C rate and varied discharge C rate. Here, the grey Verhulst metabolic model or the gray metabolic model combined with the Markov chain model is adopted to build these prediction models. Finally, the ultimate practical embedded of SoC accurate online estimate models are achieved for lithium-ion battery of vehicles, through theoretical analysis and simulation, experimental verification and optimization of relevant experiment models and algorithms. These researches maybe provide a new approach for the online SoC estimates of power lithium-ion battery of vehicles. The study of this project not only has important scientific research significance, but also has a very high application value in engineering.

为延长车用动力电池寿命，提高能量有效管理及车辆行驶安全，有必要准确估计动力电池组荷电状态(SoC)，这也是整车电池管理系统的核心和难点。本项目旨在探究灰色新陈代谢预测模型理论，解决目前车用动力锂离子电池组SoC在线估计中的难题。拟在以下几个方面开展研究：（1）基于灰色新陈代谢模型，建立一种不依赖电池模型参数的SoC灰色预测模型；（2）基于灰色关联原理，定量分析影响SoC估计要素，建立电池单体/电池组的SoC多维灰色预测模型；（3）采用灰色Verhulst新陈代谢模型或灰色新陈代谢模型与马尔可夫链预测结合探究电池组SoC估计在恒放电倍率和变放电倍率下的预测模型。通过理论分析与仿真、实验验证对相关模型和算法进行优化，最终得到实用的嵌入式车用锂离子电池组的SoC准确在线估计模型，为车用动力锂离子电池组SoC的在线估计提供新思路。本项目的研究不但具有重要的科学研究意义，而且具有极高的工程应用价值。

车用动力电池组荷电状态（State of Charge, SoC）的在线准确估算是电动汽车电池管理系统的核心和难点，对延长电池组使用寿命和提高整车性能具有至关重要的作用。本项目用新陈代谢灰色模型理论解决车用动力锂离子电池组SoC在线估算问题，在以下几方面取得预期进展。首先，深入研究灰色系统理论和动力电池放电特性，采用GM(1,1)、Verhulst GM(1,1)、新陈代谢灰色模型及滑窗灰色预测模型等建立不依赖电池模型参数的SoC灰色预测模型实现电池SoC预测，并进行精度检测和预测模型修正提高电池单体SoC估算精度；其次，用灰色关联分析原理定量分析影响电池SoC估算因素，确定电池放电过程电池SoC变化与多种影响因素（负载电流大小、电池端电压、环境温度等）灰色关联度，奠定电池组SoC估算研究基础；最后，针对电池最大可用容量随电池老化降低而导致SoC在线估算精度下降问题，采用新陈代谢灰色模型、新陈代谢残差灰色模型及新陈代谢马尔科夫残差灰色模型建立电池容量灰色模型预测电池随老化的容量变化，在恒放电倍率和变放电倍率下的预测模型，容量预测最大误差不超过3%，可有效提高电池全生命周期SoC估算精度。项目同时还系统研究单体电池筛选及成组策略，构建电池组“大单体”模型；并将滑窗灰色预测模型引入扩展卡尔曼(Extended Kalman Filter, EKF)中替代Jacobian矩阵进行先验估算，有效解决传统EKF估算电池SoC引入较大截断误差问题；同时提出分段三次埃尔米特插值法建立开路电压(Open Circuit Voltage, OCV)模型进行SoC后验更新，最终实现动态工况下SoC在线准确估算，误差不超过0.5%。通过上述理论分析与仿真、实验验证对相关模型和算法进行优化；搭建多通道动力电池嵌入式测试实验平台，运行自主开发的车用动力电池管理系统验证上述SoC在线估算算法的正确性及合理性，为车用动力锂离子电池组SoC的在线估计提供新思路。项目在车用动力电池/电池组SoC在线估算、容量在线预测、OCV建模、单体电池筛选及成组、电池管理系统等方面取得较好进展。发表/录用论文共8篇，其中被SCI收录2篇（Energy, IF:4.282; IET Power Electronics. IF:1.68），被EI收录4篇，申请发明专利8项，授权软件著作权2项。