| Title |
PCA-Based Fusion of Electrochemical Impedance Spectroscopy and Distribution of Relaxation Times Features for Early Battery Fault Diagnosis |
| Authors |
Minwoo Song ; Jaea Lee ; Taebin Ha ; Eunjin Kang ; Jonghoon Kim |
| DOI |
https://doi.org/10.6113/TKPE.2026.31.3.256 |
| ISSN |
1229-2214(pISSN), 2288-6281(eISSN) |
| Keywords |
Lithium-ion battery; Electrochemical Impedance Spectroscopy(EIS); Distribution of Relaxation Times(DRT); Principal Component Analysis(PCA) |
| Abstract |
This study proposes an early fault diagnosis method that combines indicators from electrochemical impedance spectroscopy (EIS) and distribution of relaxation times (DRT). Cells were cycled at 1C rate under high temperature; failure onset was defined as the loss of charge/discharge capability. Cycle-wise EIS data passed linear Kramers?Kronig validation, with mean real/imaginary residuals within 1%, ensuring physical consistency. From EIS, 12 indicators were extracted via equivalent-circuit and geometric analyses, and five fault-relevant indicators were selected. DRT analysis provided five additional indicators, of which two were retained, yielding seven features in total. These indicators were combined using principal component analysis. A model trained on normal data and applied for validation showed that the EIS+DRT fusion offered early warnings approximately 20 cycles before failure, outperforming EIS-only approaches. |