State of Charge Estimation for Li-Ion Battery under Non-Gaussian Uncertainties with Experimental Validation

Accurate estimation of state of charge (SoC) for batteries is essential for successful operation of battery management systems. In this work, SoC estimation under a complete non Gaussian framework is considered, where it is assumed that non Gaussian noise is present in the current as well as the terminal voltage measurements. First of all, the second-order resistance-capacitance model is established and its parameters are identified through the variable forgetting factor recursive least squares (VFFRLS) algorithm. After model validation, a maximum correntropy unscented Kalman filter ($M C_{C}-U K F$) based on Cauchy kernel was derived such that the non Gaussian uncertainties in both process and measurements are accounted for. Finally, SoC estimates were generated and validated with the acquired data set. Moreover, a comparative analysis was done with respect to the UKF in terms of root mean square error (RMSE) and its average value. Index Terms-lithium-ion battery, state of charge (SoC), parameter identification.