Reliable state of charge estimation of battery packs using fuzzy adaptive federated filtering

Inconsistencies among battery cells and measurement errors have significant impacts on the accuracy and reliability of state of charge (SOC) estimations for series-connected battery packs. The aim of this paper is to propose a novel SOC estimation method for series-connected battery packs based on the fuzzy adaptive federated filtering. The mean-plus-difference model is employed to characterize the inconsistencies among battery cells. The fuzzy system is designed to improve the accuracy and adaptability of SOC estimation under cell inconsistencies. The SOC estimation value from a cell mean model and the standard deviation of SOC estimation are combined with a fuzzy system to determine their fusion weights. The master filter adaptively adjusts the information distribution coefficient according to the local filter estimation accuracy to improve reliability. Through simulation and experimentation on series-connected battery packs with different SOC distributions, the estimation accuracy of the proposed method is compared between the proposed method and the conventional methods. The SOC estimation accuracy of each battery cell is evaluated. The results show that, over the full SOC range, the root-mean-square error (RMSE) of the battery pack SOC estimation is less than 0.6% and 1.5% using online and offline parameters, respectively. The SOC estimation RMSEs of the battery cells using online and offline parameters are less than 0.4% and 1%, respectively. The fault tolerance is verified by artificially adding measurement errors. These accurate and reliable results show a strong prospect for the design and optimization of future clean and sustainable mobility.