State of charge estimation of a LiFePO4 battery: A dual estimation approach incorporating open circuit voltage hysteresis

Abstract

In recent years, lithium iron phosphate batteries have become widespread energy sources for high-power automotive and storage applications. The efficient use of such applications require an accurate information about the battery's state of charge. Unfortunately, the battery's open circuit voltage shows a nonlinear and in wide areas flat dependence on the state of charge and is significantly influenced by hysteresis phenomena. As a result, the unique determination of the state of charge using the open circuit voltage is hampered. This work deals with a dual estimation approach utilizing a sequential Monte-Carlo method in connection with a marginalization technique to provide an accurate estimation of the battery's state of charge. In order to circumvent the hysteresis induced ambiguous mapping between the state of charge and the open circuit voltage, a nonlinear hysteresis state is introduced.