面向安全高效运行的动力电池智能管理系统研究

Battery management system (BMS) is the core guarantee for safe and efficient operation of power batteries. However, there are still many key scientific and technical bottlenecks which need to be broken. Improper or failure management of BMS is a key factor which makes the battery pack life reduced greatly or even runaway. This project focuses on developing a set of intelligent BMS theoretical method for safe and efficient operation. Firstly, this project aims at establishing the multi-time scale thermoelectric coupling model and the aging model of the battery, and realizing on-line data driven identification of their parameters to improve accuracy and robustness. Secondly, multi-state joint accurate estimation will be realized based on the strong tracking and adaptive unscented Kalman filters. Thirdly, the safe and efficient charge-discharge control strategy and the equalizer will be built based on battery microcosmic states, and the effect of battery equalization will be evaluated scientifically. Fourthly, this project plans to achieve the battery incipient fault diagnosis on-line and to establish the multi-level active safety early warning mechanism based on the multi-source information fusion theory. Fifthly, the intelligent battery module will be designed with the ability of perception, diagnosis, decision, communication, and so on. And this project will realize dynamic coordinated optimization control of the intelligent BMS based on multi-agent theory. Finally, a comprehensive evaluation method considering battery performance, economy and safety will be proposed, and all the results will be verified by experiments. This project belongs to multi-interdisciplinary fields such as control theory, electrochemistry and so on, which is of great theory value and practical significance to promote the relevant theoretical research and application, and the healthy development of battery-related industries, as well as to solve the environmental energy crisis.

电池管理系统（BMS）是动力电池安全、高效运行的核心保障，尚存在诸多关键科学与技术瓶颈亟待突破。BMS管理不当或失效是导致电池组寿命加速衰减甚至失控的关键。本项目着眼于发展一套面向安全高效运行的智能BMS理论方法，重点建立电池多时间尺度热电耦合模型和老化模型，实现其参数在线数据驱动辨识，提高模型精度和鲁棒性；基于强跟踪无迹卡尔曼滤波方法实现电池多状态自适应精准估计；研究计及电池微观状态的安全高效充放电及均衡管理，并科学评价均衡效果；基于多源信息融合理论实现电池微小故障在线诊断，建立多级主动安全预警机制；设计具有感知、诊断决策及通信等能力的智能模组，进而基于多Agent理论方法实现BMS协调优化控制，并提出考虑电池性能、经济性和安全性的综合评价方法，最终完成试验验证。本项目属控制、电化学等多学科交叉领域，对促进相关理论研究及应用、推动电池相关产业健康发展、解决环境能源危机具有重大的现实意义。

电池管理系统BMS是管好用好动力电池组，提高续驶里程的核心。针对BMS核心算法不成熟，电池管理质量难以满足电动汽车安全高效运行的需求，本项目产学研协同攻关，融合电化学、信息科学、控制科学等相关交叉学科先进理论与技术创新突破动力电池“建模-估计-均衡-诊断”新算法，发展完善了BMS设计理论方法，取得了丰富的研究成果，主要包括：①提出了动力电池分数阶融合建模与参数辨识系列方法，有效提高了动力电池模型精度和适应性；②构建了机理与数据融合的多尺度多状态自适应估计方法，有效解决了因噪声特性未知、多状态深度耦合严重影响估计精度的难题；③深入分析了电池多参数不一致性的演化发展规律，建立了自适应的参数耦合不一致性评估方法，进而提出了一系列高效紧凑型主动均衡拓扑，有效提高了动力电池组使用寿命；④构建了动力电池多故障综合诊断方法及系统，有效保障了电池系统安全高效稳定运行。应用上述理论方法研制成功高性能电池管理系统，并与合作单位北汽新能源等推广应用于电动车辆、高铁等领域。项目执行期间，共发表学术论文68篇，其中SCI收录57篇、EI收录11篇；授权中国发明专利21件、美国发明专利1件，计算机软件著作权5件；牵头和参与起草团体标准3项；培养教育部青年长江学者1名、国家优青1名、博士10名、硕士27名。相关研究成果获得中国自动化学会自然科学一等奖和山东省专利奖一等奖各1项。本项目研究成功对促进电池管理系统产业发展和科学研究，保障动力电池及其供电装备安全运行具有重要意义。