锂离子动力电池作为当前新能源汽车的核心部件，其安全性问题受到广泛关注，尤其是热失控这一核心安全问题急需解决。新能源汽车行业需要有效的热失控防控方案；而制定相关行业标准需要有效的测试和模拟手段。本课题即面向新能源汽车行业与政府制定相关行业标准的双重需求，在车用锂离子动力电池系统热失控诱发与扩展机理、建模与防控研究方面开展了大量的工作。 基于事故调查以及文献综述，指出热失控事故主要分为“热失控诱因”、“热失控发生”、“热失控扩展”三个阶段，为保证动力电池系统的安全性，必须对以上三个阶段进行逐级防控。 在热失控发生机理研究方面，针对大容量动力电池绝热热失控特性的测试难题，基于大型加速绝热量热仪的工作原理，提出了一种仪器校准方案，保证了绝热测试环境，获得了大容量动力电池的绝热热失控测试结果。进一步地，设计了“突然停止”的绝热热失控实验，冷冻了锂离子动力电池在绝热热失控测试过程中的状态，对冷却后的电池进行寿命衰减机理分析，揭示了锂离子动力电池绝热热失控测试过程中的热电耦合机制，并建立了锂离子动力电池绝热热失控的热电耦合模型。基于所建立的动力电池绝热热失控模型，提出了一种基于模型的热失控安全等级划分方法，可用于热失控的监控与预警。 在热失控扩展研究方面，针对大容量动力电池内部温度分布不均匀，热失控扩展实验结果可重复性差的问题，基于动力电池内部温度场测试与重构技术，设计并完成了针刺诱发的大容量动力电池串联模块的热失控扩展实验，获得了热失控扩展的可重复测试结果，揭示了串联电池模块的热失控扩展动力学机制，建立了热失控扩展的集总参数热阻模型以及3D热失控扩展模型。基于模型仿真结果，提出了4种抑制热失控扩展的方法，并进行了实验验证。 在热失控诱因方面，针对内短路这一热失控诱因的共性环节，提出了一种内短路替代实验方案，建立了内短路的3D电化学-产热模型。基于“平均+差异”假设的电池系统故障诊断方法，开发了基于模型的内短路检测算法，经模块内短路替代实验测试，所开发的内短路检测算法可以提前将可能造成严重热失控的内短路故障检测出来。 研究成果为宝马汽车集团、宁德时代新能源科技有限公司提供了技术指导，为制定国内外电动汽车安全技术标准提供了技术支撑。

Lithium ion traction battery is the core power source of the current renewable energy vehicles. The safety of lithium ion traction battery arouses wide concerns throughout the world. Thermal runaway is a critical problem for the battery safety issues. To guarantee the safety of the passengers and customers, the industry is calling for effective prevention approaches to control the potential hazard caused by thermal runaway, whereas the government is updating test standards, of which the test procedures require technical support. The purpose of this Ph.D. Dissertation is to serve the dual-requirement of the thermal runaway prevention and standard update from the industry and the government. The Thesis has conducted massive work on the mechanism, modeling and prevention methodologies of the battery thermal runaway. Based on the accident investigations and literature review, the accident can always be divided into three stages: thermal runaway initiation, thermal runaway occurrence, and thermal runaway propagation. The safety management should focus on the stage-by-stage prevention of thermal runaway. On the mechanisms of thermal runaway occurrence, the research solves the problem of adiabatic testing of thermal runaway for large format lithium ion battery. A calibration approach has been established based on the working mode of the extended volume-accelerating rate calorimetry. The good calibration leads to successful adiabatic tests with correct results for the thermal runaway features of large format lithium ion battery. Furthermore, adiabatic test with early termination is proposed to freeze the status of the battery during thermal runaway testing. The electrochemical-thermal coupled mechanisms during thermal runaway testing has been revealed, and a coupled thermal runaway model is built. Furthermore, a model-based algorithm to evaluate the level of battery safety has been proposed to provide early warning of thermal runaway. On the thermal runaway propagation, the research solves the test problem of the temperature distribution within large format lithium ion battery by a method with techniques on the reconstruction of the internal temperature. Experiments have been conducted for penetration induced thermal runaway propagation within large format lithium ion battery module. The thermal runaway propagation mechanism has been revealed. Thermal runaway propagation models, including a lumped model with thermal resistance and a 3D model, have been built. 4 possible prevention approaches for thermal runaway propagation have been proposed based on the modeling analysis. On the thermal runaway initiations, the research focuses on the internal short circuit, which is one of the most common characteristics among varies kinds of thermal runaway intiations. A substitute test approach is proposed to simulate the internal short circuit by experiment. A 3D electrochemical-thermal coupled model for the internal short circuit simulation has been built and can be validated by the experimental data. The modeling analysis provides guidance to develop the detection algorithm of the internal short circuit. Moreover, a model-based internal short circuit detection algorithm has been proposed based on the “mean+difference” model, which is a useful approach for the fault diagnosis of system. The internal short circuit detection algorithm can detect internal short circuit fault before the fault develops into thermal runaway. The research provides guidance for the pack design of BMW, and for developing the battery management algorithm of CATL. The techniques of testing and modeling developed in the Dissertation support the establishment of correlated safety standards for electric vehicles.