A hybrid two-stage machine learning pipeline that decouples detection from classification is proposed, and the direction of the zero-sequence signature is found to be system-dependent, motivating a learned decision boundary in place of a fixed relay threshold.
Abstract
Rapid and accurate fault detection in high-voltage transmission networks is essential for grid reliability and equipment protection. Transmission fault datasets are frequently imbalanced, and certain fault types produce electrical signatures that fall within the normal operating envelope, causing single-model classifiers to fail on safety-critical cases. This paper proposes a hybrid two-stage machine learning pipeline that decouples detection from classification. Stage 1 combines an Isolation Forest anomaly detector with an optional supervised binary detector through an OR-fusion rule; the supervised branch is allocated automatically during training for any fault class the anomaly detector cannot resolve, and is omitted when no such class exists. Stage 2 applies a Random Forest multiclass classifier only to samples flagged by Stage 1. Feature engineering is expressed as a per-measurement-point operator mapping six raw channels to eighteen features, including zero-sequence symmetrical components derived from Fortescue's theorem, yielding 18L features for L measurement points. On the TLFaultDataset, the pipeline raises Line-fault end-to-end accuracy from 31.3% to 95.8%. On an independent single-point dataset, the same framework attains 97.25% end-to-end accuracy across all classes including normal operation, exceeding the TLFed federated benchmark of 94.84% without GPU or federated infrastructure, at 0.05 ms per sample on CPU. Ablation on both datasets shows zero-sequence features resolving the three-phase versus three-phase-to-ground ambiguity, raising the F1-score of that class pair from 0.39 to 0.997. The direction of the zero-sequence signature is found to be system-dependent, motivating a learned decision boundary in place of a fixed relay threshold.
The detection of fault and its resolution are crucial in a power transmission line for ensuring unhampered and efficient power supply. These lines are often exposed to unpredictable environmental conditions and therefore encounter several challenges. Most failures in the power system are attributed to these, thereby necessitating the need for quick fault detection and resolution procedures. Research on hybridizing ANN and SVM is limited to fault detection and classification. Work on hybridizing ANN and SVM on IEE 39-bus for three simultaneous diagnostic tasks of fault type classification (LG, LL, LLG, LLL), faulted-line identification, and protection zoning defined as near-end versus far-end fault discrimination in a model is rare. This study bridges this gap by presenting a hybrid ANN-SVM model onfault type classification, fault line identification and protection zone identification in power transmission lines. The proposed framework employs ANN as a nonlinear feature embedding and a Radial Basis Function SVM subsequently classifies using an Error-Correcting Output Codes (ECOC) strategy. Evaluated on fault scenarios from the IEEE 39-bus New England test system simulated in MATLAB/Simulink R2025b, the hybrid model achieves fault type classification accuracy of 97.2%, protection zoning accuracy of 95.8%, and faulted-line identification accuracy of 97.7%, with ROC Area Under the Curve (AUC) values exceeding 0.90. These results consistently outperform standalone nd SVM baselines by 4% to 6% in fault type, 2.9% to 15.9% in protection zoning and 5% to 10% in fault line classification, validating the effectiveness of the proposed hybridisation strategy for intelligent transmission system protection.
Kudu Abubakar Mohammed, M. Balogun, Adesina M. Lambe et al.· Communication in Physical Sc...· 0 citations
A progressive three-stage time–frequency learning framework that identifies series arc faults directly from normalized current waveforms and provides robust discrimination across unseen measurement sessions within the evaluated load categories and operating conditions is presented.
Seoyoung Jeon, Won-Kyu Choi, Sungsoo Kwon et al.· Italian National Conference...· 0 citations
Experimental results on stator inter-turn faults across multiple low-load operating conditions demonstrate superior reconstruction performance compared with representative recurrent, transformer-based, and graph-based autoencoder models while maintaining computational efficiency suitable for low-latency deployment.
Chibuzo Nwabufo Okwuosa, J. Hur· IEEE Access· 0 citations
The results confirm that the proposed framework provides a comprehensive and efficient solution for real-time fault analysis by combining classification, localization, temporal analysis, and stability-aware decision support within a single model.
Nazmun Nahar Karima, M. Hazari, Shameem Ahmad et al.· Energies· 0 citations
Supervised fault classifiers frequently encounter difficulties when dealing with faults that are not present in the training data, which is a prevalent issue in industrial contexts. To mitigate this limitation, we introduce a two-stage hybrid framework for diagnosing bearing fault. In the first stage, a supervised classifier assesses the posterior probability of each Known Fault class and predicts a Known Fault when its confidence exceeds a calibrated threshold $\tau $ . Samples with confidence below $\tau $ proceed to the second stage, where a one-class anomaly detector, trained solely on Normal Operation data, evaluates whether the sample aligns with normal behavior (Normal Operation) or deviates from it (Unknown Anomaly). This confidence-based routing produces three diagnostic outcomes: Normal Operation, Known Fault, and Unknown Anomaly, without necessitating retraining of the anomaly detector when new fault types are identified. Using the MAFAULDA dataset, which includes 41 experimental blocks covering closed-set, partial open-set, and fully open-set conditions, we evaluated supervised classifiers, deep learning architectures, and one-class detectors as potential candidates for each stage. Machine learning classifiers demonstrated superior performance compared to the deep learning architectures assessed, with the Isolation Forest proving to be the most effective one-class detector. Integrating the most robust classifier with the Isolation Forest under a calibrated confidence threshold enhanced the identification of previously unseen fault types compared to a baseline relying solely on supervised methods.
Ana Caroline Mendes Costa, P. D. S. De Melo, Augusto Wohlgemuth Fleury Veloso da Silveira et al.· IEEE Access· 0 citations
A new hybrid approach to fault detection in power systems based on S-Transform feature extraction and SVM-based intelligent classification is proposed that can clearly distinguish between power swings and actual faults, including symmetrical three-phase faults that have characteristics similar to power swings.
P. Sharma, M. Silas, P. Roy et al.· African Journal Of Applied R...· 0 citations