Optimization of battery state of charge estimation method by correcting available capacity

Abstract

Accurately and efficiently estimating the state of charge (SOC) of lithium-ion batteries remains a challenging task due to the complexity of their operating conditions, such as variations in current and temperature. Previous research often overlooks the effects of current and temperature on the batteries' available capacity, making it difficult to estimate the SOC with accuracy and efficiency. To enhance the accuracy and efficiency of SOC estimation for lithium-ion batteries in real situations, this article proposes an optimization approach based on the Thin Plate Splines (TPS) method for correcting the available capacity. Initially, the battery capacity response surface is formulated using the Thin Plate Splines (TPS) method, considering current and temperature changes. The model's accuracy has been verified through constant current discharge experiments, which control temperature and discharge rate. Furthermore, the average current model is enhanced by incorporating a forgetting factor, enabling its application beyond constant-current and constant-temperature conditions to more complex conditions. During experiments, we also compared the performance of the traditional A-h integration method against the Peukert's law-based method and the TPS method. The results showed that: the optimized method based on Peukert's law reduces RMSE and MAE substantially, and the TPS method further enhances these metrics, particularly under varying temperature conditions.