Parameter Identification and SOC Estimation of a Battery Under the Hysteresis Effect

Abstract

This article concerns the estimation algorithms of battery's parameters and state of charge (SOC) and it is twofold. First, we present a simple variable length block wise least square estimation algorithm by considering locally linear SOC and open circuit voltage (OCV) relation on the run. The proposed algorithm can estimate accurately the parameters and follow the parameter changes. After estimating the battery parameters, SOC is directly computed from the combination of estimated parameters and SOC-OCV relation. This estimation scheme, which assumes hysteresis effect is negligible, is an online method. Second, if the hysteresis effect is considered in the model, we present an additional least square optimization problem to estimate SOC accurately. In this case, some parameters are tuned up in an offline way. Our approach compared with the extended Kalman filter based estimation method. The algorithms are validated by experiments with real data obtained from lab tests.