基于连接片约束的电动汽车动力电池组多重寿命衰退机理及辨识研究

The cycle-life and credible capacity of a battery module are influenced by the uneven current flowing through in-parallel battery cells caused by the battery dynamic inconsistency. In view of this problem, the degradation patterns of the battery cells under different operating conditions are firstly analyzed. The classification criterion based on the partial least squares method is then designed to judge the specific degradation pattern. Secondly, the time-frequency response of the in-parallel battery is studied. A balanced expression is further deduced to describe the relationship between the battery module model complexity and accuracy. Then a battery module equivalent circuit model that describes the whole performance of the battery is built. Thirdly, the matching relation of the dynamic inconsistency caused by the inter-cell connecting plates (ICCP) and the original inconsistency through the screening process is then investigated. And the influence of the ICCP on the evolution of the battery module inconsistency is revealed. Fourthly, the current uneven degree which is induced by the internal resistance and capacity inconsistency is inspected. And the relationship between the credible capacity and voltage curve of a battery module is studied. Lastly, the mean-difference distribution concept is applied to acquire the weight of different characteristic value in different modes. Then a multi-mode estimation method is further proposed to estimate the credible capacity of the battery module and to identify the cycle-life state of the battery pack. Furthermore, the accuracy and efficiency of the proposed algorithm are verified in the hardware-in-the-loop platform, and the reliability and real-time property of the algorithm are assessed by the data from the road condition. The research results of this project provide new theoretical foundation to research the degradation mechanism of a battery pack, and provide new idea and method to evaluate the battery pack cycle-life state.

连接片约束下的并联电池动态差异将引起支路电流不平衡，进而影响单节电池循环寿命及可信容量。针对此问题，对不同工况下电池循环寿命衰退模式进行聚类分析，基于偏最小二乘法构建衰退模式类别判定准则；分析并联电池时-频特性，制定评估单节电池模型复杂度与预测精度综合性能函数，建立全寿命周期单节电池等效电路模型；研究连接片引起的电池动态差异与初始差异等价性，得到连接片约束下并联单体电池差异化状态演变规律定量描述；分析电池参数差异对支路电流不平衡的敏感性，探明连接片约束下的单节电池可信容量与电压曲线间内在关联机理；引入“平均及离散分布”理念求解不同模式下的电压曲线特征量权重，提出多层次特征量融合的单节电池可信容量估算方法，实现动力电池组寿命状态在线辨识，并基于硬件在环仿真平台及实车道路工况数据评估算法可靠性及实时性。研究成果将为电池组寿命衰退机理研究奠定理论基础，给电池组寿命状态辨识提供新的思路与方法。

本项目重点围绕并联电池建模、差异状态分析、状态估计等开展研究。主要工作有：（1）建立了电化学-热耦合模型分析了电池电化学特性影响因素，研究了电池差异性产生机理；揭示了内部温度与交流阻抗相值映射关系，实现了无传感器的电池内部温度估计；（2）提出了考虑电压及电流双约束的并联电池模型阶次确定方法，建立了常规及差异状态下并联电池最优模型；在此基础上，研究了连接片引起的动态差异与初始差异等价性，得到了并联电池差异状态演变规律；（3）考虑电压及电流测量噪声问题，实现了基于充电电压融合双卡尔曼滤波的荷电状态估计；引入协方差矩阵的对角化分解和状态噪声矩阵自调节，提出了基于改进容积卡尔曼滤波算法的以最小荷电状态为包络线的并联电池荷电状态估计方法；（4）探明了电池老化及温度对容量增量曲线的耦合机制，结合内部温度估计实现了宽温度范围下电池容量估计；进一步，揭示了并联电池容量与电压曲线关联机理，实现了基于特征量的并联电池容量在线估计。