Parameter identification of lithium battery pack based on novel cooperatively coevolving differential evolution algorithm

Parameter identification is of great importance for lithium battery. In this study, the parameter identification problem for a lithium battery pack is addressed, and the efficient parameter identification model and algorithm are developed by using the cooperatively coevolving theory. Firstly, the offline optimization model for battery parameter identification is established by defining the identification time-window and ultra-high dimensional optimization vector. Secondly, the variable-coupling relationship is comprehensively analysed and the developed model is proved to be a partial-separate problem. Then, by introducing the dynamic-decomposition based variable-grouping mechanism, adaptive multi-mutation mechanism and deep extended archive mechanism, a novel identification algorithm is developed to improve the performance on optimizing high-dimensional models. Finally, the developed model and algorithm are verified by a comprehensive set of case studies. Experimental results show that the aforementioned algorithmic mechanisms can significantly improve the global optimization performance. In addition, the developed algorithm can obtain accurate performance for identifying more than 8000 parameters, and can also significantly outperform the compared state-of-the-art algorithms on both accuracy and robustness.