Improved joint prediction strategy for state of charge and peak power of lithium-ion batteries by considering hysteresis characteristics-current measurement deviation correction

The peak power and state of charge of lithium-ion batteries are closely related to the safety of electric vehicles. Accurate peak power and state of charge prediction can extend battery life while ensuring safe driving. In this paper, a modeling strategy for the joint estimation of the battery state of charge and peak power is proposed to consider the effect of current measurement deviation. First, a modified Thevenin model of the battery considering the internal polarization reaction process and the open-circuit voltage hysteresis effect is developed to improve the physical significance of the parameter identification results. On this basis, a current measurement deviation correction strategy based on the double-layer forgetting factor recursive least squares algorithm is proposed. To solve the nonlinearity and noise disturbance problems of the battery system, an Unscented Kalman filter-based multi-parameter constrained adaptive dynamic state observer is developed and used for the joint estimation of the state of charge and peak power. In particular, multiple parameters such as current, voltage, and state of charge are selected for the prediction of the battery peak power. Experimental results for different complex dynamic conditions at different temperatures show the excellent performance of the proposed modeling method in predicting the validity and accuracy verification of the state of charge and peak power. The proposed method provides a viable theoretical basis for the manufacturing technology of advanced battery management systems.