A closed-loop voltage prognosis for lithium-ion batteries under dynamic loads using an improved equivalent circuit model

Abstract

Discharge voltage is an essential indicator to suggest the remaining energy of a lithium-ion battery. Thus, the prediction of discharge voltage is a suitable way to alarm power exhaustion. In this paper, an improved equivalent circuit model is proposed to describe the voltage variation of lithium-ion batteries under dynamic loads. Based on this model, a closed-loop voltage prognosis is presented to compensate for the error caused by the state of charge recovery occurring when loads change. In order for the model to closely follow dynamic loads, the model parameters are continuously updated by a particle filter technique combined with a kernel smoothing-based approach, which ensures that parameters quickly converge to the actual values. Furthermore, a real dataset is used to demonstrate the effectiveness of the proposed method. The results show that the closed-loop prognosis with the improved equivalent circuit model performs well in long-term predictions under dynamic loads.