Onboard Estimation of State of Charge for Li-Ion Cells Using Fractional Order Model with Hysteresis

One of the main functions of the Battery Management System (BMS) is to estimate the State of Charge (SoC), as this gives not only an indication of the remaining charge but also an idea of the system dynamics. The efficiency of a BMS is solely reliant on the accuracy of the cell model it utilises. This paper attempts to implement the Fractional Order Model with Hysteresis on a microcontroller and investigate the efficiency of the SoC estimator with respect to the computational time along with the estimation accuracy. Lithium Iron Phosphate cells are considered for the modelling as these are by far the popular choice for Electric Vehicle (EV) batteries. Also, these batteries possess unique modelling challenges caused by flat SoC-OCV characteristics and high hysteresis. A dual-extended Kalman filter is utilised to estimate the SoC. The SoC estimator is implemented on the Hercules TMS570LS1224 board. The validation was conducted using the NEDC drive cycle at different ambient temperatures, and the Mean of Absolute Error in SoC estimation is less than 2.5% for all the different ambient temperatures.