An EIS–DRT Fusion-Based Method for Internal Resistance Anomaly Detection in Energy Storage Cells and Its Engineering Application

To ensure the safe operation of large-scale energy storage systems, internal resistance degradation induced by cell swelling is recognized as a critical precursor to thermal runaway. This study proposes an abnormal internal resistance identification method by integrating electrochemical impedance spectroscopy (EIS), distribution of relaxation times (DRT), and hybrid pulse power characterization (HPPC) for early detection of swelling faults in large-capacity LiFePO? cells. First, DRT inversion of EIS data is employed to quantitatively determine the RC order of the equivalent circuit, and a five-order RC model is constructed and validated to accurately characterize multi-timescale impedance behavior. Then, cells with different swelling degrees are fabricated via high-temperature overcharge experiments, and their resistance evolution is systematically analyzed under various states of charge using multi-rate charge–discharge and HPPC tests. The results indicate that swelling defects lead to consistent resistance increases across all time constants, with more pronounced effects in high-time-constant branches. Furthermore, based on operational data from an actual energy storage station, resistance thresholds for anomaly detection are established. A hardware-in-the-loop (HIL) platform is utilized to perform co-simulation of the battery management system (BMS). By embedding abnormal cell data into real operating profiles, the proposed method demonstrates reliable identification and localization capability under complex conditions. In addition, a multi-MCU parallel computing architecture enables millisecond-level online resistance estimation. The proposed method provides an effective engineering solution for early detection of swelling-induced abnormalities and supports safety warning and operation decision-making in large-scale energy storage systems.