A multifeature fusion approach for Lithium-ion battery state of charge estimation based on mechanical stress via the BiMamba-X model

In electric vehicles and energy storage systems, accurate estimation of the state of charge (SoC) of lithium-ion batteries is crucial for ensuring system safety and enhancing performance. However, existing battery charge and discharge processes involve volume changes, external pressure, and material structure modifications due to mechanical stress over time. These factors disturb the battery's state of health, and the data sampling intervals tend to be wide (e.g., 10 s). Existing SoC estimation techniques fail to adequately account for these factors, making it difficult to reflect the battery's state in real-world electric vehicle operating scenarios. This study investigated the influence of mechanical stress on SoC estimation in pouch lithium-ion batteries. A novel method that integrates mechanical stress with multidimensional features, such as current, voltage, and temperature, is proposed. A homemade mechanical stress test device is used for stress data acquisition to increase the perception of the internal physical state of the battery. The data are then integrated with a model named BiMamba-X, which improves the robustness, accuracy, and generalizability of SoC estimation. The research model is experimentally verified to exhibit a lower estimation error and greater goodness-of-fit at different ambient temperatures, discharge rates, and data sampling intervals. The results indicate that incorporating mechanical stress as a key input feature into the BiMamba-X model can effectively improve the SoC estimation accuracy and reliability; it compensates for the need for different data sampling intervals and has broad applicability.