SOC level estimation of lithium-ion battery based on time series forecasting algorithms for battery management system

Abstract

To fulfill a reliable battery management system, a precise state of charge (SOC) estimation method for a battery energy storage system should be developed. This study makes two contributions to the battery management system. First, a combined electro-thermal battery model is proposed. To identify the electrical and thermal battery parameters, constant current -constant voltage (CC-CV) charge, constant current (CC) discharge, and pulse discharge tests should be performed on the lithium-ion battery cells and each of the above experiments, battery SOC level should be estimated precisely. The second study of this research is the development of the SOC level estimation method by using time series forecasting algorithms. In this study, six kinds of models are used in real-time, and each of the models is evaluated with the performance indices and the computational time, and finally, forecast diagrams are graphically represented for each of the experiments.