基于物理失效模型和在线阻抗监测的锂电池状态估计与寿命预测的研究

Lithium batteries are widely used in Electrical Vehicles (EVs), Energy Storage and Consumer Electronics. Safety and reliability are considered as the critical concerns in battery applications. This is especially the case for EVs, where lithium battery safety is the primary obstacle in popularizing EVs. However, current battery management technology lacks the capability for direct monitoring of the internal parameters of battery health. Moreover, most of the models used to estimate the battery state and life are empirical, lacking physical meaning and self-adaptive capacity. This project will investigate the electrochemical and thermo-mechanical fusion process in battery failures, and establish a multiple aspect Physics-of-Failure model. By using both theoretical analysis and experimentation, this study will develop an online impedance monitoring method to extract the equal characteristic impedance between the electrodes of the battery couple and electrolyte and the solid-state electrolyte. The Physics-of-Failure model and characteristic impedance will then be used within a nonlinear filter framework to estimate the battery state and predict the Remaining Useful Life (RUL). To solve the uncertainties in actual operating conditions, the improved incremental learning and a historical model database will be used online to adjust the model parameters. The new multiple aspect Physics-of-Failure battery model developed will be used to investigate the relationship between battery failure and battery material, the micro/macro structure, and the operating conditions.

锂电池已经被广泛的应用到电动车、储能电站和消费电子中。电池的安全性和可靠性是电池应用中的关键性问题。特别对于电动车而言，锂电池的安全性问题是阻碍电池车普及的首要问题。然而目前的电池管理技术缺乏直接监测反应电池健康状况的内部参数的能力，电池的状态估计和寿命预测模型大多都是经验模型，缺乏物理意义和对不同工况的自适应能力。本研究将对电池失效的电化学和热物理的耦合过程进行研究，并建立一个多重物理失效模型。通过理论分析与实验相结合的方法，开发一种在线阻抗监测技术，用于提取并解耦电池内部正负电极以及电解质和固体电解质中间相等的特征阻抗。物理失效模型和特征阻抗将结合在非线性滤波的框架下对电池进行高精度的寿命预测和状态估计。改进的增量学习法和历史模型数据库将用来在线调节模型参数以解决工况的不确定性对模型的影响。研究成果提供全新的电池多重物理模型，揭示电池失效与电池材料、微/宏观结构、以及工况之间的联系。

锂电池已经被广泛的应用到电动车、储能电站和消费电子中。电池的安全性和可靠性是电池应用中的关键性问题。特别对于电动车而言，锂电池的安全性问题是阻碍电池车普及的首要问题。本项目首先提出一种新的商用锂电池电极分析策略，基于三电极体系对电池内部电极特性和电池外部特征参数进行建模分析，从而提取有效的全电池特征，以在线辨识电极退化行为。基于LiFePO4电池过充退化实验分析，表明了所提方法可有效建立电池外部特征和内部电极退化行为的关系模型，为电池在线退化分析提供了有效途径。其次，提出一种基于计算机断层扫描的电池无损检测方法，用于评估电池质量、识别电池制造缺陷和结构变形。通过实际案例分析表明了该方法的有效性。最后，还开发了一种基于超声传感的锂离子电池监测技术，再通过融合超声信号实现了电池健康状态评估，通过对商用锂电池的实验研究验证了该技术的可行性，从而扩展了传统电池管理系统的功能。