Finite Time State-of-Charge Observer With Robustness Against Unknown Parameter Uncertainties

Accurate state of charge (SOC) is of great engineering significance to relieve the range anxiety, avoid safety accidents, and prolong the battery lifetime. This brief proposes an SOC observer with finite-time stability and robustness against unknown and time-varying parameter uncertainties. First, an equivalent circuit model (ECM) is employed to describe the battery dynamic characteristic with full consideration of time-varying parameter uncertainties and unknown noises, which are only assumed to be bounded. Second, a hybrid method combined with off-line and on-line identifications is proposed to estimate the battery model parameters. It means that the change of the battery parameters can be well traced and the estimation error in the early stage can be effectively diminished. Third, a finite-time observer is designed to estimate SOC, which can achieve faster error convergence rate. Finally, sufficient criteria and rigorous proof are given to ensure the convergence of the estimation errors in finite time. Extensive simulations and experiments are conducted to assess the effectiveness of the proposed method as well. Results illustrate the high accuracy, strong robustness, and fast convergence speed of the proposed method with the estimation errors bounded by 1% and the convergence time limited in 20 s.