Graph Neural Network-Based Safety Evaluation and Anomaly Detection for Power Equipment Systems
Abstract
Ensuring safety and reliability in power equipment systems is critical for minimizing failures and maintaining operational efficiency. This paper introduces a novel safety evaluation and risk management framework leveraging Graph Neural Networks (GNNs). By modeling the intricate relationships among interconnected nodes in power systems, the GNN framework achieves high-precision safety score predictions, anomaly detection, and cascading failure analysis. Our model was trained and validated on a dataset comprising multi-dimensional sensor, failure, and maintenance records collected from over 1,000 equipment nodes, with more than 150,000 time-stamped entries. Experiments demonstrate that the proposed GNN framework achieves a mean accuracy of 88.9%, precision of 89.1%, recall of 87.6%, F1 score of 88.3%, and AUC-ROC of 0.93 across various hyperparameter settings. Compared to baseline methods such as traditional ML classifiers and CNN-LSTM models, the GNN exhibited superior performance in capturing spatial-temporal dependencies. The approach enables proactive identification of critical safety states and emerging risks, enhancing the resilience and reliability of complex power systems. This methodology bridges traditional safety evaluation techniques with graph-based learning, offering a scalable and intelligent solution for modern power equipment enterprises.
Full Text:
PDFReferences
Liao, W., Bak-Jensen, B., Pillai, J.R., Wang, Y. and Wang, Y., 2021. A review of graph neural networks and their applications in power systems. Journal of Modern Power Systems and Clean Energy, 10(2), pp.345-360.
Ibitoye, O.T., Onibonoje, M.O. and Dada, J.O., 2022, October. Machine learning based techniques for fault detection in power distribution grid: A review. In 2022 3rd International Conference on Electrical Engineering and Informatics (ICon EEI) (pp. 104-107). IEEE.
Samson, O.O., Gbadamosi, S.L., Onibonoje, M.O. and Ojo, E.E., 2024, April. Development of machine learning algorithms for fault detection in power systems-a review. In 2024 International Conference on Science, Engineering and Business for Driving Sustainable Development Goals (SEB4SDG) (pp. 1-6). IEEE.
Najafzadeh, M., Pouladi, J., Daghigh, A., Beiza, J. and Abedinzade, T., 2024. Fault Detection, Classification and Localization Along the Power Grid Line Using Optimized Machine Learning Algorithms. International Journal of Computational Intelligence Systems, 17(1), p.49.
Li, Q., Luo, H., Cheng, H., Deng, Y., Sun, W., Li, W. and Liu, Z., 2023. Incipient Fault Detection in Power Distribution System: A Time–Frequency Embedded Deep-Learning-Based Approach. IEEE Transactions on Instrumentation and Measurement, 72, pp.1-14.
Ghamizi, S., Bojchevski, A., Ma, A. and Cao, J., 2024. SafePowerGraph: Safety-aware Evaluation of Graph Neural Networks for Transmission Power Grids. arXiv preprint arXiv:2407.12421.
Chen, K., Hu, J., Zhang, Y., Yu, Z. and He, J., 2019. Fault location in power distribution systems via deep graph convolutional networks. IEEE Journal on Selected Areas in Communications, 38(1), pp.119-131.
Chen, Z., Xu, J., Peng, T. and Yang, C., 2021. Graph convolutional network-based method for fault diagnosis using a hybrid of measurement and prior knowledge. IEEE transactions on cybernetics, 52(9), pp.9157-9169.
Ghamizi, S., Ma, A., Cao, J. and Cortes, P.R., 2024, July. OPF-HGNN: Generalizable Heterogeneous Graph Neural Networks for AC Optimal Power Flow. In 2024 IEEE Power & Energy Society General Meeting (PESGM) (pp. 1-5). IEEE.
Varbella, A., Amara, K., Gjorgiev, B., El-Assady, M. and Sansavini, G., 2024. PowerGraph: A power grid benchmark dataset for graph neural networks. arXiv preprint arXiv:2402.02827.
Li, Y., Xue, C., Zargari, F. and Li, Y., 2023. From Graph Theory to Graph Neural Networks (GNNs): The Opportunities of GNNs in Power Electronics. IEEE Access.
Liu, S., Wu, C. and Zhu, H., 2022. Topology-aware graph neural networks for learning feasible and adaptive AC-OPF solutions. IEEE Transactions on Power Systems, 38(6), pp.5660-5670.
Wu, T., Scaglione, A. and Arnold, D., 2023. Complex-value spatio-temporal graph convolutional neural networks and its applications to electric power systems AI. IEEE Transactions on Smart Grid.
Khayambashi, K., Hasnat, M.A. and Alemazkoor, N., 2024. Hybrid chance-constrained optimal power flow under load and renewable generation uncertainty using enhanced multi-fidelity graph neural networks. Journal of Machine Learning for Modeling and Computing, 5(4).
Jiangang, L., Huijuan, T., Ruifeng, Z., Wenxin, G., Zhen, D. and Ping, L., 2024, September. Applications of Graph Computing and Graph Neural Networks in Power Systems: A Survey. In 2024 China International Conference on Electricity Distribution (CICED) (pp. 543-548). IEEE.
Huang, J., Li, Y., He, S., Hao, G., Zhou, C. and Zeng, Z., 2024. Graph Learning for Power Flow Analysis: A Global-Receptive Graph Iteration Network Method. IEEE Transactions on Network Science and Engineering.
Yang, M., Qiu, G., Liu, T., Liu, J., Liu, K. and Li, Y., 2024. Probabilistic power flow based on physics-guided graph neural networks. Electric Power Systems Research, 235, p.110864.
Ringsquandl, M., Sellami, H., Hildebrandt, M., Beyer, D., Henselmeyer, S., Weber, S. and Joblin, M., 2021, October. Power to the relational inductive bias: Graph neural networks in electrical power grids. In Proceedings of the 30th ACM International Conference on Information & Knowledge Management (pp. 1538-1547).
Chen, Y., Jiang, T., Heleno, M., Moreira, A. and Gel, Y.R., 2022, December. Evaluating distribution system reliability with hyperstructures graph convolutional nets. In 2022 IEEE International Conference on Big Data (Big Data) (pp. 1793-1800). IEEE.
DOI: https://doi.org/10.31449/inf.v49i30.8084
This work is licensed under a Creative Commons Attribution 3.0 License.