Page Header Logo Applied Science and Engineering Progress

Home > ONLINE FIRST > Mohammed Merza

Towards Smart Substations: A Cloud-Based Hybrid Deep Learning Model for Fault Detection and Predictive Maintenance of GIS Equipment

Ahmed Mohammed Merza, Ali Abdulhasan Rasool Al‐Karaawi, Assel Ali Hussein, Ahmed Shaker Abdulaah, Mohammed Abdul Azeez Yousif Alkhafaji, Ali Najim Abdullah, Wael Abdulhasan Atiyah

Abstract


Gas-insulated substations (GIS) have become an important constituent of extra-high voltage (EHV) electrical systems because they are smaller and more efficient. Conversely, the conventional maintenance method, which usually relies on planned maintenance, can lead to unnecessary maintenance or sudden equipment failure. This research work is a holistic framework that would increase fault detection and proactive protection of GIS equipment. In the proposed design, a hybrid CNN–LSTM model processes multimodal sensor data (PD, temperature, SF₃, and environmental measures) for fault identification and predictive maintenance, while the cloud platform provides expandable storage, real-time analysis, and digital twin-based monitoring. Additionally, before switching operations, the DC main contact validation subsystem acts as a physical verification layer to verify conductor-path integrity. The updated manuscript highlights these components' complementary roles in improving GIS dependability, safety, and maintenance efficiency by clearly describing how they interact within a unified end-to-end architecture. The proposed framework achieves an average classification accuracy of 94.7%, with a macro-F1 score of 94.3% and an AUC of 0.965, demonstrating strong performance across multiple modalities. It also achieves a TPR of 0.88 at 1% FPR and an ECE of 2.7%, indicating reliable and well-calibrated detection. A real-world case study of a 132 kV GIS explosion in the Hilla West substation confirms that early warning indicators and pre-operational DC contact verification can effectively prevent failures, supporting the practicality and scalability of intelligent GIS monitoring systems.

Keywords



[1] N. Zebouchi and M. A. Haddad, “A review on real-size epoxy cast resin insulators for compact high voltage direct current gas insulated switchgears (GIS) and gas insulated transmission lines (GIL)—Current achievements and envisaged research and development,” Energies, vol. 13, no. 23, p. 6416, Dec. 2020, doi: 10.3390/en13236416.

[2] A. Ymeri, N. Krasniqi, G. Kicaj, and D. Morina, “Advantages of a gas insulated substation (GIS),” in Proceeding of UBT International Conference, 2023, doi: 10.33107/ubt-ic.2023.53.

[3] T. Li, Q. Xu, K. Gao, Z. Yuan, J. Chen, and C. Li, “Review of insulation defect detection methods for a gas-insulated switchgear,” Energies, vol. 19, no. 6, Art. no. 1491, Mar. 2026, doi: 10.3390/en19061491.

[4] Q. Wang, Z. He, S. Lin, and Z. Li, “Failure modeling and maintenance decision for GIS equipment subject to degradation and shocks,” IEEE Transactions on Power Delivery, vol. 32, no. 2, pp. 1079–1088, Apr. 2017, doi: 10.1109/TPWRD.2017.2655010.

[5] D. Olojede, S. King, and I. Jennions, “Application of machine learning in power grid fault detection and maintenance,” Energy Informatics, vol. 8, no. 1, Sep. 2025, doi: 10.1186/s42162-025-00574-w.

[6] Y. A. M. Alsumaidaee, C. T. Yaw, S. P. Koh, S. K. Tiong, C. P. Chen, and K. Ali, “Review of medium-voltage switchgear fault detection in a condition-based monitoring system by using deep learning,” Energies, vol. 15, no. 18, p. 6762, Sep. 2022, doi: 10.3390/en15186762.

[7] K. Sundsgaard, “Predictive maintenance of medium voltage cable networks using machine learning,” M.S. thesis, Technical University of Denmark, 2025, doi: 10.11581/e39dacea-e69a-4398-a8ed-68b3e0cfc96e.

[8] J. Xu, “GIS equipment status assessment method for substations based on multi-frequency signal processing,” International Journal for Housing Science and Its Applications, vol. 46, no. 4, pp. 5978–5990, 2025. doi: 10.70517/ijhsa464435.

[9] P. D. Rocha, F. J. P. Lopes, and L. A. Da Silva Cruz, “Automating electrical grid asset inspection: From current challenges to future directions,” IEEE Access, vol. 13, pp. 201392–201438, 2025, doi: 10.1109/ACCESS.2025.363 6718.

[10] W. Chi, T. Wang, J. Zhang, Z. Wang, and C. Zhang, “Full-life-cycle management of high-voltage bushings based on digital twin: Typical scenarios, core technologies, and research prospects,” Energies, vol. 18, no. 23, p. 6343, Dec. 2025, doi: 10.3390/en18236343.

[11] T. Ji, Z. Liu, X. Zhuang, Q. Li, L. Zhang, and Q. H. Wu, “Advancing GIS operational monitoring: A novel voiceprint recognition method using Grassmann manifold and multi-kernel functions,” IEEE Transactions on Power Delivery, vol. 39, no. 5, pp. 2894–2907, Oct. 2024, doi: 10.1109/TPWRD.2024.3448354.

[12] K. Sekar, M. S. Nattar, P. Muthukamatchi, N. Ranganathan, and S. Srithar, “Integrating machine learning and IoT for real-time PdM in industrial ecosystems: A case study analysis citation,” International Journal of Research in Industrial Engineering, vol. 14, no. 2, pp. 385–409, 2025, doi: 10.22105/riej.2025.502596.1531.

[13] H. Koch et al., “Operation and maintenance,” in Gas Insulated Substations, H. Koch, Ed. Hoboken, NJ, USA: Wiley, Dec. 2021, pp. 325–370, doi: 10.1002/9781119623632.ch7.

[14] W. Hu, J. Li, X. Liu, and G. Li, “Partial discharge fault identification method for GIS equipment based on improved deep learning,” The Journal of Engineering, vol. 2024, no. 5, May 2024, doi: 10.1049/tje2.12386.

[15] Y. Sun et al., “Partial discharge pattern recognition of transformers based on MobileNets convolutional neural network,” Applied Sciences, vol. 11, no. 15, p. 6984, Jul. 2021, doi: 10.3390/app11156984.

[16] A. Jafari, M. K. Azad, A. M. Zanjani, and M. A. Fatideh, “GIS-based predictive maintenance for electrical insulators,” in Proceeding of 2025 29th International Electrical Power Distribution Conference (EPDC), Tehran, Iran, 2025, pp. 1–6, doi: 10.1109/EPDC67173.2025.11278249.

[17] M. W. Hoffmann et al., “Integration of novel sensors and machine learning for predictive maintenance in medium voltage switchgear to enable the energy and mobility revolutions,” Sensors, vol. 20, no. 7, p. 2099, Apr. 2020, doi: 10.3390/s20072099.

[18] L. Yunyang, W. Lixiang, and Z. Yang, “Cloud-based monitoring and management platform for electrical equipment,” in Proceeding of 2025 IEEE 3rd International Conference on Image Processing and Computer Applications (ICIPCA), Shenyang, China, 2025, pp. 1995–1998, doi: 10.1109/ICIPCA65645.2025.11138 900.

[19] S. Pan, W. Zhang, Y. Ma, Y. Su, and W. Huang, “Pattern recognition of GIS partial discharge based on UHF signal characteristics,” Electronics, vol. 15, no. 5, p. 1096, Mar. 2026, doi: 10.3390/electronics15051096.

[20] A. Moradzadeh, H. Teimourzadeh, B. Mohammadi-Ivatloo, and K. Pourhossein, “Hybrid CNN-LSTM approaches for identification of type and locations of transmission line faults,” International Journal of Electrical Power & Energy Systems, vol. 135, p. 107563, Feb. 2022, doi: 10.1016/j.ijepes.2021.107563.

[21] Y. A. M. Alsumaidaee et al., “Fault detection for medium voltage switchgear using a deep learning hybrid 1D-CNN-LSTM model,” IEEE Access, vol. 11, pp. 97574–97589, 2023, doi: 10.1109/ACCESS.2023.3294093.

[22] T. Liu, J. Yan, Y. Wang, Y. Xu, and Y. Zhao, “GIS partial discharge pattern recognition based on a novel convolutional neural networks and long short-term memory,” Entropy, vol. 23, no. 6, p. 774, Jun. 2021, doi: 10.3390/e23060774.

[23] A. Aleran, H. Almukhalfi, A. Noor, R. Alluhaibi, A. Hafez, and T. H. Noor, “An IoT-based predictive maintenance framework using a hybrid deep learning model for smart industrial systems,” Computers, Materials and Continua, vol. 86, no. 3, Jan. 2026, doi: 10.32604/cmc.2025.070741.

[24] S. Rana, “AI-driven fault detection and predictive maintenance in electrical power systems: A systematic review of data-driven approaches, digital twins, and self-healing grids,” American Journal of Advanced Technology and Engineering Solutions, vol. 1, no. 1, pp. 258–289, Feb. 2025, doi: 10.63125/4p25x993.

[25] N. Q. Minh, N. T. Khiem, and V. H. Giang, “Fault classification and localization in power transmission line based on machine learning and combined CNN-LSTM models,” Energy Reports, vol. 12, pp. 5610–5622, Nov. 2024, doi: 10.1016/j.egyr.2024.11.061.

[26] F. M. Almasoudi, “Grid distribution fault occurrence and remedial measures prediction/forecasting through different deep learning neural networks by using real time data from Tabuk city power grid,” Energies, vol. 16, no. 3, p. 1026, Jan. 2023, doi: 10.3390/en16031 026.

[27] T. Li et al., “Durative monitoring of sulfur hexafluoride characteristic gases under hydrogen interference using a Time2Vec-encoded CNN–Transformer–LSTM model based on a heterogeneous gas sensor array,” ACS Sensors, vol. 10, no. 11, pp. 8809–8820, Oct. 2025, doi: 10.1021/acssensors.5c02740.

Full Text: PDF

DOI: 10.14416/j.asep.2026.08.012

Refbacks

  • There are currently no refbacks.


Copyright © 2025 by King Mongkut’s University of Technology North Bangkok. All rights reserved.