The End of Discharge prediction of the Lithium-Ion battery with improved model based on Particle Filtering

Abstract

Determining the end of discharge in relation to time for Li-ion batteries is possibly one of the most important problems in the field of battery health management. This paper proposes the improvement of a physical model taking into account the relationship between state-of-charge (SoC), open circuit voltage (OCV) and end of discharge (EOD). We try to predict EOD by describing the SoC-OCV relationship in the circuit models. There are many models of Lion batteries to solve the reaction on batteries and a mathematics equation is picked and utilized to explain lithium-ion chemical reaction. This paper will show two different models of Li-ion batteries and the second model is used as an experiment for making the prediction. Parameters are identified based on experimental data and supplemented by Prognostics Center of Excellence (PCoE), NASA batteries data-set. Moreover, this paper enhances the SoCOCV relationship with Nernst’s theorem and the Randle’s circuit model for improving the model. Finally, the EOD curve is good fitness in fitting Li-ion batteries discharge of the real data curve. This paper also presents EOD predictions based on particle filtering.