SOC estimation for traction battery pack of electric vehicles with battery degradation tracking

Abstract

In the electric vehicle industry, a good estimation of a traction battery pack or the state of charge (SOC) is crucial as it reflects how far a vehicle travel before recharging. As the battery degrades, its behavior and the associated parameters such as internal resistance, capacity and SOC-OCV (open circuit voltage) mapping changes. Thus, a battery model has to take into account the changes in the battery parameters for it to be accurate throughout the battery lifetime. For such a model to be computational intensive, it requires powerful processors. With limited calculation performance processors found in vehicles, the model fidelity is normally compromised. In this paper, two battery models are used to accurately estimate traction battery SOC; The Ohmic resistance model is used to sense changes in battery internal resistance, when the change is significant, the resistor-capacitor (RC) model is used to update the battery SOC-OCV curve which is used to estimate the battery initial SOC. Hence, the coulomb counting method is used to update the battery SOC. The real operational battery data from PEA Ze-Bus (Zero-Emission bus of the Provincial Electricity Authority of Thailand) are used in this study. The proposed algorithm used to test the state of charge of the battery has been verified and illustrates the error of SOC estimation at 3.31%, less than the unadaptable model.