The LiFePO4 battery lifespan SoC estimation using Ham-Informer and internal pressure

The goal of this study is to use the Ham-Informer to accurately and reliably estimate state of charge (SoC) throughout the LiFePO 4 battery lifespan. First, Pearson heat map is used to study the correlation between different parameters and SoC, and found that the correlation between internal pressure is as high as 0.97. Hence , internal pressure is introduced to improve the accuracy of SoC estimation. The Informer model is refined using Hampel filter and Multi-head ProbeHamSparse to reduce the computational complexity and improve its feature extraction ability. The use of internal pressure and Ham-Informer can reduce the MAE (Mean Absolute Error) of the estimated SoC values of LFP batteries with light, mild, and severe degradation by 75 %, 74 %, and 86 %, respectively, compared to the absence of internal pressure. Compared with the unimproved Informer, the MAE of SoC estimates at different life stages decreased by 76 %, 68 %, and 84 %. The model exhibiting an average decrease of 89 % in MAE when compared to the LSTM model, and a 77 % decrease when compared to the GRU model. The incorporation of internal pressure and Ham-Informer can enhance the accuracy and reliability of battery lifespan SoC estimation. Compared with other methods, the computational complexity of Ham-Informer is smaller. Therefore, it can be applied to battery management systems, enabling accurate and dependable SoC estimation throughout the battery's lifespan.