Accurate detection of partial discharges (PDs) in medium-voltage overhead transmission lines is critical for preemptive maintenance and avoiding costly outages, yet it is challenged by scarce labeled data and pervasive electromagnetic interference. This paper investigates a hybrid simulation-and-data-driven framework in which synthetically generated PD signals are used to pretrain deep neural networks and are subsequently fine-tuned on a limited set of real overhead-line measurements. The synthetic pipeline systematically varies PD repetition rates, amplitude distributions, vegetation-contact scenarios, and noise conditions, producing diverse time-series and spectrogram-like representations that approximate real operating environments. We conduct a comprehensive ablation study across multiple architectures-Convolutional Neural Networks (CNNs), a Vision Transformer (ViT), and a Long Short-Term Memory (LSTM) network-and analyze their sensitivity to granular sweeps of synthetic-data parameters. CNN-based models decisively outperform ViT and LSTM counterparts on the spectrogram-based classification task, while ViT and LSTM fail to learn meaningful representation. For the successful CNNs, pretraining on carefully parameterized synthetic datasets-particularly those reflecting higher PD activity, such as our Datasets 3 and 4-consistently improves downstream performance on real data, boosting the Matthews Correlation Coefficient (MCC) on imbalanced, cost-sensitive test sets by roughly 10-20% compared with training from scratch. At the same time, we show that poorly aligned synthetic data can degrade generalization, underscoring the need for accurate noise calibration and domain-aligned simulation. Overall, the results confirm that (i) architectural choice is pivotal for PD detection in overhead lines and (ii) well-designed synthetic data is a powerful, practical lever for achieving reliable and cost-effective PD monitoring when real labeled data are limited.

Department of Computer Science VSB Technical University of Ostrava Ostrava Czech Republic

ENET Centre CEET VSB Technical University of Ostrava Ostrava Czech Republic

Zobrazit více v PubMed

Stone, G. Partial discharge diagnostics and electrical equipment insulation condition assessment. DOI

Pakonen, P.

Kabot, O., Fulneček, J., Mišák, S., Prokop, L. & Vaculík, J. Partial discharges pattern analysis of various covered conductors. In

Chiu, B., Roy, R. & Tran, T. Wildfire resiliency: California case for change. DOI

Talaat, M., El-Shaarawy, Z., Tayseer, M. & El-Zein, A. An economic study concerning the cost reduction of the covered transmission conductors based on different optimization techniques. DOI

Fulnecek, J. & Misak, S. A simple method for tree fall detection on medium voltage overhead lines with covered conductors. DOI

Misak, S., Fulnecek, J., Jezowicz, T., Vantuch, T. & Burianek, T. Usage of antenna for detection of tree falls on overhead lines with covered conductors. DOI

Kaziz, S. et al. Radiometric partial discharge detection: A review. DOI

Chan, J. Q., Raymond, W. J. K., Illias, H. A. & Othman, M. Partial discharge localization techniques: A review of recent progress. DOI

Uwiringiyimana, J. P. et al. Comparative analysis of partial discharge detection features using a UHF antenna and conventional HFCT sensor. DOI

Yang, Y. et al. Detection of partial discharge patterns in hybrid high voltage power transmission lines based on parallel recognition method. PubMed DOI PMC

Rauscher, A., Kaiser, J., Devaraju, M. & Endisch, C. Deep learning and data augmentation for partial discharge detection in electrical machines. DOI

Klein, L. et al. A data set of signals from an antenna for detection of partial discharges in overhead insulated power line. PubMed DOI PMC

Kabot, O., Klein, L., Prokop, L., Mišák, S. & Slanina, Z. Dataset for antenna-based detection of fault types in covered conductors for 22 kv voltage power lines. PubMed DOI PMC

Kabot, O., Klein, L., Prokop, L. & Walendziuk, W. Enhanced fault type detection in covered conductors using a stacked ensemble and novel algorithm combination. PubMed DOI PMC

Song, Y. et al. Online multi-parameter sensing and condition assessment technology for power cables: A review. DOI

Long, L. et al. On LLMs-driven synthetic data generation, curation, and evaluation: A survey. In (eds. Ku, L.-W., Martins, A. & Srikumar, V.)

Zhang, C., Kuppannagari, S. R., Kannan, R. & Prasanna, V. K. Generative adversarial network for synthetic time series data generation in smart grids. In

Rather, I. H. & Kumar, S. Generative adversarial network based synthetic data training model for lightweight convolutional neural networks. PubMed DOI PMC

Kahr, M., Kovács, G., Loinig, M. & Brückl, H. Condition monitoring of ball bearings based on machine learning with synthetically generated data. PubMed DOI PMC

Selçuk, ŞY., Ünal, P., Albayrak, Ö. & Jomâa, M. A workflow for synthetic data generation and predictive maintenance for vibration data. DOI

Khan, M. A. et al. Improved fault classification and localization in power transmission networks using VAE-generated synthetic data and machine learning algorithms. DOI

Ahang, M. et al. Synthesizing rolling bearing fault samples in new conditions: A framework based on a modified CGAN. PubMed DOI PMC

Li, S. et al. Partial discharge data enhancement and pattern recognition method based on a CAE-ACGAN and ResNet. DOI

Klein, L., Dvorsk , J. & Nagi, L. Usability of cGAN for partial discharge detection in covered conductors 246–260 (Springer Nature Switzerland, 2024).

Pinceti, A., Sankar, L. & Kosut, O. Synthetic time-series load data via conditional generative adversarial networks. 10.48550/ARXIV.2107.03545 (2021).

Ibarrola, F. J., Ravikumar, N. & Frangi, A. F. Partially conditioned generative adversarial networks. 10.48550/ARXIV.2007.02845 (2020).

Ganin, Y. et al. Domain-adversarial training of neural networks.

Chen, Z., Fu, H. & Zeng, Z. A domain adaptation neural network for digital twin-supported fault diagnosis. In

Azuma, C., Ito, T. & Shimobaba, T. Adversarial domain adaptation using contrastive learning. DOI

Imbusch, B. T., Schwarz, M. & Behnke, S. Synthetic-to-real domain adaptation using contrastive unpaired translation. In:

Cruz, J. D. S. et al. Partial discharges monitoring for electric machines diagnosis: A review. DOI

Hussain, G. A. et al. Review on partial discharge diagnostic techniques for high voltage equipment in power systems. DOI

Klein, L., Kabot, O., Mimra, T. & Slanina, Z. Spectrogram-based fault detection in covered conductors using ResNet50V2 with SHAP and Grad-CAM analysis. In

Klein, L., Žmij, P. & Krömer, P. Partial discharge detection by edge computing. IEEE Access 1–1. 10.1109/ACCESS.2023.3268763 (2023).

He, K., Zhang, X., Ren, S. & Sun, J. Identity mappings in deep residual networks. In (eds Leibe, B., Matas, J., Sebe, N. & Welling, M.) Computer Vision—ECCV 2016, 630–645 (Springer Int. Publ., 2016).

Kingma, D. P. & Ba, J. Adam: A method for stochastic optimization. 10.48550/ARXIV.1412.6980 (2014).

Chicco, D. & Jurman, G. The advantages of the Matthews correlation coefficient (mcc) over f1 score and accuracy in binary classification evaluation. PubMed DOI PMC