Forecasting the state-of-charge of Li-ion batteries using fuzzy inference system and fuzzy identification

Abstract

This study proposes a method to forecast the state of charge (SOC) of Li-ion batteries using Fuzzy inference system and Fuzzy identification. In this study, 5 pieces of Li-Co batteries were used in this research for the life-cycle testing. The cycle testing includes CC (0.5C)/CV (4.2V) charge, CC (0.2, 0.4, 0.6, 0.8, 1C) discharge, and the rest time (one minute). The life-cycle testing indicates the relations of the voltage, the discharging time and the SOC with various life-cycles and various discharging currents. This study forecast the SOC with the data of the above, Fuzzy inference system and Fuzzy identification. This study also compares the SOC forecast accuracy using Fuzzy inference system, Fuzzy identification, and Fuzzy inference system combined with Fuzzy identification. The testing results reveal that the average error, the standard deviation, the maximum error, and the minimum error of the forecasted SOC was -0.4%, 6%, 18% and 25.1%, respectively. The 81.48% of the forecasted SOC error is within ± 5%.