State of charge estimation of lithium-ion batteries using a fractional-order multi-dimensional Taylor network with adaptive Kalman filter

In the study, a fractional-order multi-dimensional Taylor network (FMTN) structure was introduced to accurately estimate the state of charge (SOC) of lithium-ion batteries (LiBs). By combining multi-dimensional Taylor expansion with fractional calculus, the fractional power function was utilized as the activation function of the middle layer node in the FMTN to improve the fineness of the network. Besides, to solve the fluctuation of the SOC estimation of LiBs caused by drastic changes in the measured data and the noise in the actual driving environment of electric vehicles, the adaptive Kalman filter (AKF) algorithm was combined with the FMTN model. With an open dataset, the accuracy and robustness of the SOC estimation method based on the FMTN model with the AKF algorithm (FMTN-AKF) under various temperatures and operating conditions were evaluated. The results show that the accuracy of the SOC estimation based on the FMTN-AKF method is significantly improved. The average value of the root-mean-square error (RMSE) of the FMTN-AKF method is decreased by 51 %, 29 %, and 42.5 % compared with that of the FMTN, FMTN-KF, and MTN-AKF, respectively. In addition, there is no significant increase in the estimation time of the FMTN-AKF method in comparison with other methods.