Deep transfer learning-enabled battery health prognosis using impedance spectrum data

Lithium-ion battery state of health (SOH) and remaining useful life (RUL) are crucial for maintaining their safety and reliability. Electrochemical impedance spectroscopy (EIS) provides extensive information about the battery's status, but practical challenges in obtaining full-frequency EIS data include long measurement times, low accuracy due to noise, and high costs. Given the dominant role of mid and low-frequency EIS in battery aging processes, this paper proposes an elliptical fitting feature extraction method leveraging these frequency-specific impedance data. A transfer learning-based predictive model is developed under varying temperature conditions to enable battery state of health estimation and remaining useful life prediction. This method significantly improves the accuracy and efficiency of state of health estimation and remaining useful life prediction, addresses the limitations of traditional methods, and reduces the dependence on extensive measurement data. This advancement enhances the safety, reliability, and cost-effectiveness of battery applications.