Results indicate that the proposed model provides an effective representation-learning approach for simulated transmission-line fault classification, and independent validation under a retraining protocol further supports the reproducibility of the framework on another simulated power-line fault dataset.
Abstract
Accurate transmission-line fault diagnosis is important for reliable power-system protection. Existing deep learning methods often rely on phase-domain voltages and currents, which may be insufficient to distinguish severe three-phase faults with and without a ground path. This paper proposes a domain-knowledge-enhanced 7-channel MTF-ResNet18-MSA framework for electrical fault diagnosis. The input consists of three-phase voltages, three-phase currents, and zero-sequence current. The zero-sequence current is introduced as an explicit grounding-related variable, while Markov Transition Field (MTF) encoding maps local voltage-current sample groups into two-dimensional transition maps. A ResNet18 backbone is adapted to 7-channel inputs, and a multi-head self-attention module is inserted after global average pooling to refine high-level feature dependencies. To avoid overlap-induced train-test leakage, the original sample records are split before local window generation. The proposed model achieves 100.00% accuracy under noiseless conditions and maintains 98.04% accuracy at 10 dB Gaussian noise. Compared with 1D-CNN, 1D-ResNet, 1D-Transformer, MTF-CNN, MTF-ResNet18, GAF-ResNet18-MSA, and STFT-ResNet18-MSA, it obtains the highest average accuracy over the tested noise levels. Channel ablation demonstrates that the 7-channel input is more effective than the 3-channel, 6-channel, and 8-channel configurations in this dataset. Sensitivity analysis shows stable performance under different MTF window lengths and bin numbers. Independent validation under a retraining protocol further supports the reproducibility of the framework on another simulated power-line fault dataset. The results indicate that the proposed model provides an effective representation-learning approach for simulated transmission-line fault classification. Further event-level validation, field-data testing, and lightweight deployment remain necessary.
In renewable-powered distribution systems and microgrids, reliable cable condition monitoring is essential for operational security and early fault detection. Sheath circulating current signals provide valuable information for identifying grounding abnormalities and incipient faults, but their diagnosis is difficult because the signals exhibit strong inter-phase coupling, and fault samples are limited in practice. This study proposes a dual-view Mixup-ResNet framework for fault diagnosis of cable sheath circulating current signals. Specifically, physical-range normalization is employed to retain magnitude-related fault information and inter-phase proportional relationships, while sample-wise z-score normalization is used to emphasize waveform morphology. These two complementary views are concatenated to form a six-channel input for a lightweight one-dimensional residual network, which is trained with Mixup, label smoothing, dropout, and cosine annealing. On an ATP-EMTP-generated eight-class dataset, the proposed method achieves an average accuracy of 91.50%, a weighted F1-score of 91.69%, and a macro F1-score of 91.69% under a unified 5 × 5 repeated stratified cross-validation protocol. Additional tests under 12% relative Gaussian noise show that the method maintains competitive Gaussian-noise tolerance within the tested condition, although broader field disturbances remain to be further validated. These findings suggest that the proposed method has potential for small-sample fault diagnosis of cable sheath circulating current signals and provides a preliminary basis for intelligent cable condition monitoring.
Haiqi Yang, Jinwei Mao, Bo Zhang et al.· Energies· 0 citations
High-voltage transmission lines are critical carriers of electric power, and rapid and accurate fault location is essential for secure and stable power-system operation. This paper proposes a fault-location method that integrates distributed-parameter physics-based features with a convolutional block attention module-based convolutional neural network (CBAM-CNN). Voltages and currents measured at both line terminals are transformed into modal quantities, after which a distributed-parameter line model is used to derive compensated-voltage waveforms along the line and construct a physically meaningful feature matrix. The CBAM-CNN is trained offline using multiple waveform samples, and the fault location is determined from the minimum similarity value among the observation points. Under identical test settings, the proposed model achieves an overall mean absolute fault-location error of 0.0746 km across four simulated test locations, representing reductions of 79.23%, 85.15%, and 68.44% relative to CNN, SE-CNN, and ECA-CNN, respectively. For a field record whose operation and maintenance record places the fault 51.0 km from the M terminal, the proposed model estimates 51.2089 km, corresponding to an absolute error of 0.2089 km. These results demonstrate high fault-location accuracy within the scope tested and provide preliminary evidence of applicability to field recordings.
Faguang Chen, Mengzhou Li, Shibin Fan et al.· Electronics· 0 citations
Rapid and accurate identification of various faults occurring in transmission lines is essential for restoring normal line operation. However, existing transmission line fault diagnosis methods still face challenges in terms of noise immunity and diagnostic accuracy. To address these issues, this paper proposes a deep learning method based on recurrence plots and a convolutional neural network–bidirectional gated recurrent unit–attention mechanism model. The voltage and current signals of transmission lines are transformed into recurrence plots in both the time and frequency domains. Parallel convolutional neural networks are then employed to extract local features from the two domains, while bidirectional gated recurrent units are used to capture temporal dependencies. Furthermore, multi-head self-attention and cross-attention mechanisms are introduced to enhance key features within each domain and achieve adaptive fusion of inter-domain feature information. A transmission line model is established in Simulink to collect data under various fault conditions and influencing factors, thereby verifying the effectiveness and adaptability of the proposed method. Experimental results show that the proposed method achieves fault recognition accuracies of 99.63%, 96.68%, and 75.38% under NL1, NL2, and NL3 Gaussian-noise conditions, respectively, and maintains accuracies of 99.02%, 95.93%, and 72.43% under mixed-noise conditions. Compared with other deep learning models, the proposed method demonstrates higher diagnostic accuracy and stronger robustness.
Fei Long, Long Hong, Zhenman Gao· Processes· 0 citations
The operational reliability of rotating machinery is critical for modern industrial systems. However, existing large language model (LLM)-based fault diagnosis approaches face challenges in processing high-frequency, strong-noise vibration signals, including modal misalignment, cross-modal negative fusion, and limited learning capacity for early-stage weak faults. To address these issues, this paper proposes a gate-free multi-physical field fusion LLM (GMPF-LLM) for fault diagnosis. The method constructs a parallel multi-physical feature extraction and alignment mechanism to effectively decouple steady-state harmonics, background residuals, and transient impulses. A continuous multimodal projector with a gate-free residual fusion structure is employed to mitigate information distortion and noise interference during cross-modal fusion. Furthermore, a difficulty-aware joint optimization strategy using low-rank adaptation (LoRA) and multi-class focal loss enhances recognition of early weak faults under low signal-to-noise ratio conditions. Experiments on the XJTU-SY and GDUPT bearing datasets demonstrate mean diagnostic accuracies of 99.25% and 90.55% with minimal standard deviation, outperforming mainstream methods. Ablation studies further confirm the contributions of each core module. This work provides an effective framework for integrating multi-physical field data with LLMs in industrial fault diagnosis.
Quansi Huang, Guanhua Zhu, Renhui Yu et al.· Engineering Research Express· 0 citations
This paper presents a compact physics-informed convolutional neural network (CNN) for post-fault classification and feeder-section localization in medium-voltage distribution networks using only three-phase voltages and currents measured at the primary substation. Instead of relying on raw waveforms, each post-fault record is represented by features extracted from a one-cycle window at 50 Hz and sampled at 10 kHz. Fundamental-frequency phasors are estimated through a single-cycle discrete Fourier transform, and three feature families are evaluated: phasor magnitudes, phasor magnitudes with phase-difference features, and an extended representation including symmetrical-component indicators. The proposed multi-task CNN jointly predicts the faulted feeder section among six candidate line sections and the fault category among four classes: line-to-ground (LG), line-to-line (LL), double line-to-ground (LLG), and three-phase (LLL). The method is validated on a simulated 20 kV two-feeder radial distribution network under an unseen fault resistance of 15Ω, which is excluded from training. Results show that phase information improves single-window localization, while the inclusion of symmetrical-component indicators further enhances performance, reaching 94.00% location accuracy and 98.65% fault-type accuracy. A record-level decision strategy based on the fusion of consecutive post-fault windows improves event-level robustness, achieving 96.44% Top-1 location accuracy, 100.00% fault-type accuracy, and 99.92% Top-2 location accuracy. Residual localization errors are mainly associated with electrically adjacent sections, which is consistent with the observability limits of single-ended measurements.
Eduardo Salazar, Verónica Rosero, F. Gonzalez-Longatt· 2026 6th International Confe...· 0 citations
Existing transmission line component recognition methods mostly adopt closed-set detection paradigms and rely on large-scale annotated data. They fail to accommodate diverse equipment models and emerging defects in complex industrial park scenarios, and their fault recognition results cannot support comprehensive operation and maintenance decision-making. To address these issues, this paper proposes a two-stage open-world component recognition method for industrial park microgrid inspection. In the first stage, a lightweight Transformer detector coupled with permutation attention and deformable convolution is constructed to enhance the feature extraction capability of small-scale components and suppress complex industrial background interference. In the second stage, a lightweight vision-language large model (VLM) is introduced to build a multi-task inference framework of "component recognition—defect diagnosis—public health risk assessment—financial analysis". It performs zero-shot recognition of unseen components and novel defects, and combines industrial park scene semantics to analyze personnel health and safety risks, power outage impacts, maintenance costs and economic losses that may be caused by transmission line faults, generating structured fault diagnosis and decision-making reports. Furthermore, a contrastive learning-based domain-adaptive feature alignment strategy is designed to mitigate domain shifts in industrial park scenarios and reduce annotation requirements for target domains. Experimental results show that the proposed method achieves 96.2% mAP@0.5 for known classes and 82.5% zero-shot recognition accuracy on the self-built industrial park transmission line dataset, outperforming mainstream real-time and zero-shot detection methods. It can provide technical support for intelligent inspection and comprehensive operation and maintenance decision-making of industrial park microgrids.
Chaoyuan Hu, Jiadong Zhang, Sifei Wang et al.· International Conference on...· 0 citations