Abstract

    Determining the state of charge (SoC) of a battery is a crucial factor for its proper functioning. Due to its strongly non-linear character, this is a real challenge for LiFePO4 (LFP) batteries. In this paper we present a technique for estimating the SoC, which is based on a third order RC network model and an adaptive observer. The observer gets, as input values, the voltage at the terminals of the battery and the discharge current, and provides the SoC of the battery as an output. Testing the performance of the adaptive observer is done by direct comparison of data from its output with the data obtained experimentally by discharging an LFP battery in keeping with the Urban Dynamometer Driving Schedule (UDDS) discharge cycle.