A Novel Modeling Methodology Based on Dynamic Hysteresis Characters for Lithium-Ion Battery

Battery manage systems (BMS) rely heavily on sets of models to describe the lithium-ion cells. However, the hysteresis voltage response of the batteries, which is caused by the polarization of the electrodes, will inevitably cause a great influence on the accuracy and real-time performance of the model's output. In this work, a dynamic Prandtl–Ishlinskii model of hysteresis is established to present a certain magnitude of hysteresis not only regarding the major open circuit voltage (OCV) boundaries but also the different minor loops of partial charge and discharge at the different lifetime stages. The relationship between state of charge (SOC), ambient temperature, cycling C-rate, aging state and hysteresis behavior of the cell's voltage is investigated through a self-designed pulsed-current test over a range of operating temperatures. Finally, the dynamic performance of the model is verified using the results obtained from the electric vehicle (EV) under dynamic driving cycle. The results show that better OCV prediction accuracy can be achieved when the hysteresis voltage is a function with various parameters rather than a hypothesis constant. The proposed model can significantly improve the accuracy of the model-based SOC estimator as well as the robustness in terms of initial voltage error and current error.