Performance evaluation strategy for battery pack of electric vehicles: Online estimation and offline evaluation

Abstract

Electric vehicles are powered by battery packs, which are usually composed of hundreds of units in series or in parallel. For power battery pack performance evaluation, many literatures have been published, including online performance assessment, life prediction and offline performance evaluation. Nevertheless, performance evaluation strategies that include both online and offline have not attracted sufficient attention. In this paper, we propose a performance evaluation method based on MCPE-DEKF, which can solve the problem of consistency analysis and sort of battery cells offline, as well as, implementing battery pack state estimation online. MCPE-DEKF is designed to enhanced the accuracy and adaptability for power battery pack SOC estimation. The pack SOC online estimation value from cells means model and the standard deviation of SOC estimation are combined with MCPE to determine their aggregation weights. An offline evaluation framework for cells grading and sorting problem is solved by MCPE methods. As the results show that, the RMSE of the online estimation for the battery pack is less than 2% and 7 mV for the SOC and the terminal voltage, respectively. The RMSE of the cell estimation is less than 0.3% and 6 mV, respectively. In terms of offline evaluation approaches for cells, we propose an approach with a fusion method to improve the reliability of sorting by integrating method 1 and method 4 (the second-best sorting method).