Cell-Wise Fault Warning for Lithium-Ion Battery Packs Using Local Outlier Factor and Validity-Gated Extreme-Value Thresholds
- 1Hangzhou Xili Intelligent Technology Co., Ltd., Hangzhou 310024, China
- 2School of Materials and Energy, University of Electronic Science and Technology of China, Chengdu 611731, China
- 3School of Computer Science and Engineering, University of Electronic Science and Technology of China, Chengdu 611731, China
Abstract
Early fault warning in lithium-ion battery packs is challenged by weak labels and changing operating conditions. This study combines cell-wise Local Outlier Factor (LOF) scoring with validity-gated Streaming Peaks-Over-Threshold (SPOT). Five descriptors are used to summarize each seven-sample median-centered voltage trajectory, and the maximum cell LOF is logarithmically transformed into a pack score. A generalized Pareto distribution (GPD) supplies the threshold when calibration diagnostics support the fitted tail; otherwise, an empirical quantile is used. Across nine real-world vehicles and six AI-generated CH-BatteryGen cases, 12 streams retain GPD-SPOT and three use the empirical fallback. At a target risk parameter q = 0.01, aggregate false-alarm rates are 1.85% for pre-event baselines and 1.64% for independent normal vehicles. Seven recorded field events have an alarm within the preceding 24 h, with a median lead time of 17.39 h. However, car10 provides only 0.053 h of notice, and window-level recall is 0.029 under broad ±24 h event labels. The supplied faulty cell ranks first in all three labeled lithium iron phosphate cases. The workflow provides statistical fault warnings and suspicious-cell rankings without fault-labeled model training.
Related
- Acoustic Emission to Provide Early Warning Detection of Lithium-Ion Battery Thermal Runaway
- Critical Evaluation of Electrochemical Impedance Spectroscopy for State-of-Health Estimation of Lithium-Ion Cells Against Battery Management System Baselines
- Evaluation of α-V2O5 lithium ion battery cathode performance by means of cavity microelectrode: Effect of the particle size
- Surface Spin‐State Manipulation via a Strong Electronegative Ligand Field Enables Direct Regeneration of Spent Lithium‐Ion Battery Cathodes
- An Improved SegRNN-Based Hierarchical Warning Method for Power Battery Thermal Runaway
BibTeX
@article{yang2026cellwise,
title = {{Cell-Wise Fault Warning for Lithium-Ion Battery Packs Using Local Outlier Factor and Validity-Gated Extreme-Value Thresholds}},
author = {Xing Yang and Ziqiang Zhang and Zhili Qin and Baihai Li},
journal = {Batteries},
year = {2026},
month = oct,
doi = {10.3390/batteries12100415},
url = {https://www.mdpi.com/2313-0105/12/10/415},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Cell-Wise Fault Warning for Lithium-Ion Battery Packs Using Local Outlier Factor and Validity-Gated Extreme-Value Thresholds
AU - Xing Yang
AU - Ziqiang Zhang
AU - Zhili Qin
AU - Baihai Li
PY - 2026
DA - 2026/10/10
JO - Batteries
DO - 10.3390/batteries12100415
UR - https://www.mdpi.com/2313-0105/12/10/415
AB - Early fault warning in lithium-ion battery packs is challenged by weak labels and changing operating conditions. This study combines cell-wise Local Outlier Factor (LOF) scoring with validity-gated Streaming Peaks-Over-Threshold (SPOT). Five descriptors are used to summarize each seven-sample median-centered voltage trajectory, and the maximum cell LOF is logarithmically transformed into a pack score. A generalized Pareto distribution (GPD) supplies the threshold when calibration diagnostics support the fitted tail; otherwise, an empirical quantile is used. Across nine real-world vehicles and six AI-generated CH-BatteryGen cases, 12 streams retain GPD-SPOT and three use the empirical fallback. At a target risk parameter q = 0.01, aggregate false-alarm rates are 1.85% for pre-event baselines and 1.64% for independent normal vehicles. Seven recorded field events have an alarm within the preceding 24 h, with a median lead time of 17.39 h. However, car10 provides only 0.053 h of notice, and window-level recall is 0.029 under broad ±24 h event labels. The supplied faulty cell ranks first in all three labeled lithium iron phosphate cases. The workflow provides statistical fault warnings and suspicious-cell rankings without fault-labeled model training.
SN - 2313-0105
ER -