A Data Augmentation and Deep Residual Network–Based Method for Water Detection in Cable Elbow Terminations
Main Article Content
Abstract
This paper uses an acoustic fingerprint recognition technique to solve the problems posed by lack of samples and degraded recognition performance in low signal-to-noise ratio (SNR) situations when detecting water ingress in cable elbow termination. This technique incorporates data augmentation and deep residual networks. Two unique data augmentation techniques, Speed Perturbation (SP) and Spectrogram Masking (SM), are proposed to extend sample diversity in the time and time-frequency domains. The classification backbone here is ResNet-18, which uses skip connections in order to address the issue of vanishing gradients, enabling the system to perform the recognition task from Mel spectrograms to states of water ingress in an end-to-end manner. Experimental evidence illustrates that ResNet-18 minimizes the total error rate from 9.68% (CNN baseline) to 3.46%. The SP+SM strategy showed the best results in the 0–5 dB high-noise region. In the 5–15 dB moderate-noise regions, SM alone showed the best results, while SP alone is better suited to clean, high-SNR scenarios. These findings provide a rationale for dynamically implementing augmentation strategies in practical settings. The non-contact and easy-to-deploy nature of this method presents a promising approach for the intelligent detection of water ingress hazards in power distribution equipment.
Downloads
Article Details
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
References
H. Yang, J. Wang, P. Zhao, C. Yu, H. You, and J. Dou, “On-Line Insulation Monitoring Method of Substation Power Cable Based on Distributed Current Principal Component Analysis,” Energies, vol. 18, no. 3, pp. 20, 2025. DOI: 10.3390/en18030688.
Y. Li, L. Jiang, M. Xie, J. Yu, L. Qian, K. Xu, M. Chen, and Y. Wang, “Advancements and Challenges in Power Cable Laying,” Energies, vol. 17, no. 12, pp. 26, 2024. DOI: 10.3390/en17122905.
J. Tao, S. U. Rehman, R. Ali, and S. A. Raza, “Advancement and challenges: A review of power cable aging monitoring and diagnostic techniques,” Renew. Sust. Energ. Rev., vol. 222, pp. 20, 2025. DOI: 10.1016/j.rser.2025.115970.
P. Meng, T. Li, K. Zhou, Z. Tang, G. Zhu, Z. Li, Y. Cao, H. Zhang, Y. Yin, and J. Guo, “A Novel Time-Frequency-Domain Reflectometry Location Method for Power Cable Defects Based on Synchrosqueezing Transform,” IEEE Trans. Instrum. Meas., vol. 73, pp. 9, 2024. DOI: 10.1109/TIM.2024.3418110.
C. C. Uydur and O. Arikan, “Dielectric performance analysis of laboratory aged power cable under harmonic voltages,” Electr. Eng., vol. 104, no. 6, pp. 4197– 4212, 2022. DOI: 10.1007/s00202-022-01614-4.
G. Zhu, Z. Liu, K. Zhou, Z. Wang, L. Lu, Y. Li, and P. Meng, “A Novel Dampness Diagnosis Method for Distribution Power Cables Based on Time-Frequency Domain Conversion,” IEEE Trans. Instrum. Meas., vol. 71, pp. 9, 2022. DOI: 10.1109/TIM.2022.3160840.
G. Zhu, Z. Liu, S. Pan, P. Meng, K. Zhou, X. Wang, Q. Shen, and H. Zhou, “A Dampness Discrimination Method for MV Power Cable Joints Based on PDC Testing Under Thermal Excitation Conditions,” IEEE Trns. Dielectr. Electr. Insul., vol. 32, no. 1, pp. 581–588, 2025. DOI: 10.1109/TDEI.2024.3487957.
L. Wang, B. Wang, H. Ma, J. Zhang, Y. He, H. Wang, and H. Zhang, “Power Cable Status Evaluation Method Based on Electrical Tree Growth and Data Association Rules,” Processes, vol. 12, no. 9, pp. 1, 2024. DOI: 10.3390/pr 12092026.
T. G. Bade, J. Roudet, A. Derbey, A. de Andrade, and L. Sadi-Haddad, “Adapter for the Impedance Measurement of Power Cable With an Impedance Analyzer,” IEEE Trans. Electromagn. Compat., vol. 65, no. 3, pp. 705–715, 2023. DOI: 10.1109/TEMC.2023.3250107.
G. Li, J. Chen, H. Li, L. Hu, W. Zhou, C. Zhou, and M. Li, “Diagnosis and Location of Power Cable Faults Based on Characteristic Frequencies of Impedance Spectroscopy,” Energies, vol. 15, no. 15, pp. 18, 2022. DOI: 10.3390/en15155617.
A. Ghaforian, P. Duggan, and L. Lu, “A Comprehensive Review of Cable Monitoring Techniques for Nuclear Power Plants,” Energies, vol. 18, no. 9, pp. 26, 2025. DOI: 10.3390/en18092333.
C. S. Subudhi and S. K. Sen, “Non-Intrusive Online Time Domain Reflectometry Technique for Power Cables,” IEEE Trans. Power Deliv., vol. 39, no. 1, pp. 101– 110, 2024. DOI: 10.1109/TPWRD.2023.3324655.
J. Wang, Z. Zhang, R. Zhou, A. Fan, T. Zhai, T. Guo, and Y. Zhao, “A Novel Fiber-Optically Powered Partial Discharge Monitoring System for Cable Joints Based on Acoustic Emission,” IEEE Sens. J., vol. 25, no. 20, pp. 38809–38819, 2025. DOI: 10.1109/JSEN.2025.3605966.
D. F. J. Jabha, R. Joselin, and R. Sowmya, “Examining the Use of Acoustic Emission Technique for Evaluating Partial Discharge in Power Cables: A Review,” J. Fail. Anal. Prev., vol. 24, no. 5, pp. 2316–2326, 2024. DOI: 10.1007/s11668-024-02006-5.
W. Tang, K. Brown, D. Mitchell, J. Blanche, and D. Flynn, “Subsea Power Cable Health Management Using Machine Learning Analysis of Low-Frequency Wide-Band Sonar Data,” Energies, vol. 16, no. 17, pp. 17, 2023. DOI: 10.3390/en16176172.
J. Lee, J. Park, and H. Park, “Optimizing Cable Replacement Schedules in Power Grids: A Data-Driven Approach with Machine Learning and Power Flow Analysis,” IEEJ Trans. Electr. Electron. Eng., vol. 20, no. 9, pp. 1497–1502, 2025. DOI: 10.1002/tee.70056.
H.-W. Sian, C.-C. Kuo, S.-D. Lu, and M.-H. Wang, “A novel fault diagnosis method of power cable based on convolutional probabilistic neural network with discrete wavelet transform and symmetrized dot pattern,” IET Sci. Meas. Technol., vol. 17, no. 2, pp. 58–70, 2023. DOI: 10.1049/smt2.12130.
M.-H. Wang, S.-D. Lu, and R.-M. Liao, “Fault Diagnosis for Power Cables Based on Convolutional Neural Network With Chaotic System and Discrete Wavelet Transform,” IEEE Trans. Power Deliv., vol. 37, no. 1, pp. 582–590, 2022. DOI: 10.1109/TPWRD.2021.3065342.
J. Zhao, X. Han, Z. Niu, G. Zhang, J. Yang, B. Li, and Y. Wu, “Power System Operation State Identification Considering the Process of Transient Stability,” Proceedings of the Chinese Society of Electrical Engineering, vol. 44, no. 20, pp. 7970–7982, 2024. DOI: 10.13334/j.0258-8013.pcsee.231219.
L. Xi, D. Peng, W. Cao, H. Chen, F. Bai, and W. Wang, “FDIA Location Detection for Data-driven Algorithms in Cyber-physical Power Systems,” Proceedings of the Chinese Society of Electrical Engineering, vol. 45, no. 18, pp. 7110–7122, 2025. DOI: 10.13334/j.0258-8013.pcsee.240412.
B. Zhao, X. Liu, L. Zhang, Z. Chen, L. Zhang, and C. Xie, “Fault Diagnosis of Proton Exchange Membrane Fuel Cell Integrated System Based on GoogleNet and Transfer Learning,” Proceedings of the Chinese Society of Electrical Engineering, vol. 44, no. 13, pp. 5147–5157, 2024. DOI: 10.13334/j.0258-8013.pc see.230210.
L. Gao, D. Wang, L. Zhuang, X. Sun, M. Huang, and A. Plaza, “BS3LNet: A New Blind-Spot Self-Supervised Learning Network for Hyperspectral Anomaly Detection,” IEEE Trans. Geosci. Remote Sensing, vol. 61, pp. 18, 2023. DOI: 10.1109/TGRS.2023.3246565.
J. Ircio, A. Lojo, U. Mori, S. Malinowski, and J. A. Lozano, “Minimum Recall-Based Loss Function for Imbalanced Time Series Classification,” IEEE Trans. Knowl. Data Eng., vol. 35, no. 10, pp. 10024–10034, 2023. DOI: 10.1109/TK DE.2023.3268994.