LPV Estimation of SOC Based on Electricity Conversion and Hysteresis Characteristic

Abstract

This paper proposes the concept of an electricity conversion coefficient. The variation of the electricity value, with different charging and discharging currents, is unified into discharging electricity at the standard discharging current through the electricity conversion coefficient. An experimental study is performed on the hysteresis characteristics of a LiFePO4 battery. The results reveal whether or not the battery hysteresis characteristics are considered. Simultaneously, the linear model replaces the charging and discharging hysteresis curve of the state of charge (SOC) at different starting points, and the maximum error between the model value and the test value is <3%. On this basis, using the linear parameter-varying (LPV) theory and the extended Kalman filter (EKF) principle, a SOC estimator that considers the hysteresis characteristics is designed. A simulation model is constructed to estimate the SOC of the battery. The disparity between the simulation and experimental results shows that the corresponding SOC–open-circuit-voltage relationship, which is in the curve of the battery hysteresis in main circulation, can be expressed more accurately by considering the battery hysteresis characteristic in secondary circulation, and that the real-time estimation error of the SOC can be more accurate.