Unknown input observer design for lithium-ion batteries SOC estimation based on a differential-algebraic model

An accurate battery management strategy is a crucial need in the developing of reliable and viable plug in and hybrid electric vehicles. This on-board algorithm has the advantages to protect the battery from critical operating conditions and improve its lifetime. However, the effectiveness of the battery management strategy mainly depends on the accuracy of its state of charge (SOC). In this context, this paper proposes a novel technique for the SOC estimation based on the unknown input observer and a new differential-algebraic model of a lithium iron phosphate battery. The proposed observer aims to overcome the unknown value of the initial SOC for on-board batteries using only current and terminal voltage measurements. A reduced-order based unknown input observer is developed to estimate the open circuit voltage and the SOC using the OCV-SOC characteristic offline-determined. The unbiasedness of the estimation error is guaranteed by the parameterization of a set of Sylvester constraints. The performance of the proposed observer is verified by simulations and experiments and the accuracy of the obtained results is analyzed and assessed.