A Discussion on a Feedback Mechanism of Estimated OCV in RLS Algorithm Based on Battery Equivalent Circuit Model

Abstract

Open circuit voltage (OCV) is a critical value in the models of Lithium ion batteries and also widely used in the battery modeling as well as state estimation in the battery management system (BMS) of electric vehicles, hybrid electric vehicles and plug-in hybrid electric vehicles (EVs/HEVs/PHEVs). Till now various methods of obtaining the value of OCV have been presented, during which the method of online estimating OCV through the equivalent circuit model (ECM) and reclusive least square (RLS) algorithm is especially typical and common. Based on this approach OCV is considered as an unknown quantity and solved recursively at each time step without high computational burden. However, although a lot of researchers focused on the applications of obtained OCV through this method, there has not been any discussion or analysis about the accuracy or convergence of OCV in the recursion considering co-relevant uncertainties between OCV and other model parameters which is deemed crucial and necessary. In this paper, the authors attempt to put forward a feedback mechanism of coefficients existing in the difference equation of battery ECM which guarantees the error of OCV will not change in a single direction or diverge in each recursion. To validate this explanation, the experiment profile of New European Driving Cycle (NEDC) is used to obtain and analyze the change of OCV.