Options
2013
Journal Article
Title
Online parameterization of a function describing the open-circuit voltage by a least square method with adaptive forgetting factor
Abstract
Adapting the model employed by stochastic filters for state estimation of a specific battery type or during aging of the battery in the application is time consuming or never done leading to decreased accuracy of both state of charge and state of health estimation. The most important characteristic of a battery for state estimation is the open-circuit voltage since this voltage is correlated to the state of charge. This paper introduces a procedure for identifying online a function describing the open circuit voltage while running the battery in any application avoiding thereby laboratory testing of the battery. The future aim is a plug-and-play state estimation by stochastic filters for lithium-ion batteries. The procedure first uses a recursive least square method with adaptive forgetting factor which employs an impedance model for identifying the current open circuit voltage of the battery, then correlates this value to the current state of charge and finally uses a least square method without forgetting factor to estimate the function for the open-circuit voltage characteristic based on the Nernst equation. The method reaches accuracies for new and aged batteries in photovoltaic and electric vehicle applications of better than 1%.