高效紧凑车载动力电池均衡与低温加热拓扑及其自学习控制研究

Battery balancing and low-temperature heating management systems are the cores to ensure high-efficiency and safe operations of lithium-ion power battery packs of electric vehicles. However, lithium-ion power batteries have unknown internal mechanism, bad inconsistency, poor low-temperature performances, high nonlinearity, and strong coupling, leading to difficult design and control of battery balancing and low-temperature heating topologies. Therefore, according to the balancing issues in and among battery modules, following the integrated design of battery equalizers and heaters, this project will mainly study the evolution law and evaluation method of battery inconsistency based on probability statistics; study the balancing topology in battery modules based on coupled half-bridge converters and propose the multi-variable self-adaption balancing strategy based on voltage, SOC, and remaining capacity; study the integrated topology of the active equalizer between battery modules and the low-temperature heater, and propose a heating method based on iterative learning control; analyze the damage of battery balancing and heating to batteries based on reinforcement learning, and ascertain the optimum operating ranges of battery balancing and heating. Finally, this project will form a set of methods of design and control of active balancing and low-temperature heating for battery strings. This project is a cross-disciplinary field of control, power electronics, computers, which is of great practical significance to improve the available capacity, cycle life, and economy of battery strings, and ensure the safe, reliable, and high-efficiency operations of electric vehicles and power batteries.

电池均衡与低温加热管理系统是保障电动汽车锂离子动力电池组安全高效运行的核心。锂离子动力电池因具有内部机理不明、不一致性差、低温性能差、高度非线性和强耦合性，致使电池均衡、低温加热拓扑设计与控制难度极大。为此，本项目拟从电池模组内和模组间均衡问题入手，以均衡－加热一体化设计为指导，重点研究基于概率统计的电池单体不一致性演化规律及其评估方法；研究基于耦合半桥变换的模组内均衡拓扑，提出基于电压、SOC、剩余容量的多变量自适应均衡方法；研究模组间均衡－低温加热一体化拓扑，提出基于迭代学习控制的加热方法；基于深度学习建立均衡和加热对电池的损伤模型，探明均衡、加热最优工作区间；最终形成一套电池组主动均衡、低温加热拓扑设计与控制方法。本项目属控制、电力电子、计算机等多学科交叉领域，对提升动力电池组可用容量、使用寿命和经济性，保障电动汽车及动力电池安全、可靠、高效运行具有重要意义。

均衡与预热技术是改善电池组一致性、提高低温电池性能的必备手段。电池均衡和预热系统作为有限能源车载装置，对体积、能效、成本要求极为严苛。然而，现有主动均衡拓扑结构冗余、效率低、成本高；传统交流预热频率低，不仅导致电流源体积大，车载应用难，而且易损伤电池。本项目针对这两个难题，经3年攻关，取得如下创新成果。.首先，基于概率统计研究了不同充电倍率和温度下电池容量的不一致性分布规律，建立了基于标准差的电池组不一致性评估模型。然后，发明了基于同步整流和正反激变换的高效紧凑均衡拓扑，降低了电路体积和成本，提高了均衡效率和速度；提出了基于博弈论的动力电池剩余电量、健康状态SOH一致性控制策略，有效提高了电池系统容量、延长使用寿命。其次，推演出只由电池供电的LC谐振式正弦波交流预热拓扑，建立了交流预热电流和效率的数学模型，揭示了开关频率、电感L、电容C等参数对预热能效的影响机制；构造出基于Buck-Boost变换的“均衡-预热”一体化拓扑，提出了基于高增益控制的快速柔性预热控制策略，解决了预热速度与电池损伤之间的矛盾。最后，探明了电池阻抗随频率、温度的变化规律，揭示了高频交流预热下电池欧姆和电化学产热机理，建立了电池热电耦合模型；研究了交流预热电流频率、幅值、波形对预热速度、效率的影响；形成了交流预热对电池损伤评估方法；给出了最优预热电流频率和波形，即高频方波。.项目执行期间，发表学术论文13篇，其中SCI检索论文7篇，EI检索论文6篇，入选ESI高被引和热点论文1篇。授权中国发明专利6件、PCT国际检索报告1件，申请中国发明专利5件、美国发明专利1件、欧洲发明专利1件，5件专利实现转化应用。项目负责人商云龙入选国家优青、山东省杰青、泰山学者青年专家、美国斯坦福大学发布的全球前2%顶尖科学家。研究成果获中国自动化学会CAA自然科学一等奖、日内瓦国际发明展金奖、山东省专利奖一等奖、中国专利优秀奖以及国际顶刊IEEE TIE杰出论文奖等。.本项目攻克了车载动力电池高效均衡与柔性预热的难题，对提升动力电池可用容量、延长使用寿命、保障安全可靠运行具有重要意义。