Digital Twin Model for Lithium-Ion Battery SOC Estimation in Battery Swapping Station

As electric vehicles rapidly gain popularity, battery swapping stations have emerged as key infrastructure to enhance the convenience of electric vehicles. Accurately estimating the State of Charge (SOC) of batteries during the battery swapping process is critical for ensuring efficient operation and optimizing battery management. This paper proposes a digital twin framework and establishes a data-driven model for SOC estimation. The model, Ensemble Weighted Network (EWN), is embedded in the digital twin framework. It has three remarkable processes: multi-model ensemble, accuracy weight scaling, and time weight scaling. The multi-model ensemble combines five popular SOC estimating learners, to avoid the accuracy limitations of a single estimation model. The accuracy weight scaling and time weight scaling strategies are put forward to assign appropriate weights to each model, addressing significant fluctuations in estimation capability among different models across various periods. The results show the root mean square error, mean absolute error, and mean absolute percentage error of the model are less than 1.10%, 0.96%, and 1.92% respectively. Furthermore, we conduct experiments on public datasets to prove the proposed model has high estimation accuracy and robustness for different temperatures and battery types. Finally, the visualization of the digital twin is built, allowing operators to monitor and manage the batteries.