ANALYSIS OF CASCADE EFFECT SCENARIOS IN CRITICAL INFRASTRUCTURE BASED ON GRAPH NEURAL NETWORKS
DOI:
https://doi.org/10.32689/maup.it.2025.4.28Keywords:
critical infrastructure, machine learning, neural network, similarity coefficient, graphs, cascading effect, cluster analysisAbstract
The complexity of critical infrastructure is increasing, which increases the potential negative consequences of cascading effects, so studies of cascading effects scenarios in critical infrastructure are being conducted to determine actions to reduce the negative impact of failures on the system. Since data on the operation of power grids are limited, modeling and simulation of power grid operation scenarios with different values of component parameters are being conducted for research. When processing power grid operation scenarios in cascading effects scenarios, there is a need to search for similar scenarios or identify scenario patterns to generate recommendations based on existing data and make decisions. Machine learning methods (graph neural networks) are a promising direction for analyzing power grid operation in various scenarios. Modern technologies provide the opportunity to create software for real-time data processing, which increases the speed of response to potentially dangerous events. The aim of the article is to develop a method and software architecture for forming a representation of the power grid in cascading effects scenarios for comparison, identification of similar and alternative paths of scenario development. Methodology. The article describes a method for identifying similar alternative scenarios of cascading effects in the power grid, which uses a graph-layered autoencoder model to form a representation of the vertices and edges of the graph in a step or sequence of steps of the scenario. The trained model is used to form representations of the power grid operation in the scenarios and similarities based on the cosine of similarity value. The DBSCAN clustering algorithm is used to identify scenario patterns. The described software architecture provides a tool for interacting with cascading effects scenarios in the power grid operation and a neural network. The scientific novelty of the work lies in the development of a method that improves the process of searching for similar scenarios of cascading effects, patterns and generating recommendations for decision-making based on existing scenarios. The software architecture provides the ability to process data in real time and quickly respond to anomalies in the data. Conclusions. A method and software architecture have been developed to form a representation of the power grid in cascade effects scenarios for comparison, identification of similar and alternative paths of scenario development. The developed method uses an autoencoder model based on a graph neural network to form a representation of the power grid in cascade effects scenarios, cosine similarity to determine the similarity of scenarios, and the DBSCAN clustering algorithm to search for patterns in scenarios. The developed method improves the process of comparing a step or sequence of steps in cascade effects scenarios of the power grid, identification of similar power grid states and patterns for decision-making based on existing scenarios. The developed software architecture provides the ability to process data in real time for rapid response to events in the system.
References
Хоменко О. М., Сенченко В. Р., Коваль О. В. Мережевий підхід при дослідженні каскадних ефектів критичних інфраструктур. Реєстрація, зберігання і обробка даних, 2024. Том 26. № 2. С. 44–72. DOI 10.35681/1560-9189.2024.26.2.316908
Change data capture. URL: https://en.wikipedia.org/wiki/Change_data_capture
Cosine similarity. URL: https://en.wikipedia.org/wiki/Cosine_similarity
Cuadra L., Salcedo-Sanz S., Del Ser J., Jiménez-Fernández S., Geem ZW. A Critical Review of Robustness in Power Grids Using Complex Networks Concepts. Energies. 2015. 8(9), 9211–9265. https://doi.org/10.3390/en8099211
DBSCAN. URL: https://uk.wikipedia.org/wiki/DBSCAN
Di Nardo A., Giudicianni C., Greco R., Herrera M., Santonastaso, G.F. Applications of Graph Spectral Techniques to Water Distribution Network Management. Water. 2018, 10, 45. https://doi.org/10.3390/w10010045
Freitas S., Yang D., Kumar S., Tong H., Chau D. H. Graph Vulnerability and Robustness: A Survey. https://doi.org/10.48550/arXiv.2105.00419
Gilmer J., Schoenholz S. S., Riley P. F., Vinyals O., Dahl G. E. Neural Message Passing for Quantum Chemistry. https://doi.org/10.48550/arXiv.1704.01212
Gjorgiev B., David A.E., Sansavini G. Cascade-risk-informed transmission expansion planning of AC electric power systems. Electric Power Systems Research, Volume 204, 2022, 107685, ISSN 0378-7796, https://doi.org/10.1016/j.epsr.2021.107685
Hernandez J. M., Van Mieghem P. Classification of graph metrics. November 2011. URL: https://www.nas.ewi.tudelft.nl/people/Piet/papers/TUDreport20111111_MetricList.pdf
Hines P., Balasubramaniam K., Sanchez E. C. "Cascading failures in power grids," in IEEE Potentials, vol. 28, no. 5, pp. 24–30, September-October 2009, doi: 10.1109/MPOT.2009.933498
Kadri F., Birregah B., Châtelet E. The Impact of Natural Disasters on Critical Infrastructures: A Domino Effect-based Study. Homeland Security & Emergency Management. 2014. 11(2), 217–241. https://doi.org/10.1515/JHSEM-2012-0077
Koç Y., Warnier M., Kooij R., Brazier F. Structural vulnerability assessment of electric power grids. https://doi.org/10.48550/arXiv.1312.6606
Mattsson L.-G., Jenelius E., Vulnerability and resilience of transport systems a discussion of recent research, Transportation Research Part A: Policy and Practice, vol. 81, pp. 16–34, 2015. https://doi.org/10.1016/j.tra.2015.06.002
Md Sami N., Naeini M. Machine Learning Applications in Cascading Failure Analysis in Power Systems: A Review. https://doi.org/10.48550/arXiv.2305.19390
Nair A. S., Abhyankar S., Peles S., Ranganathan P. Computational and numerical analysis of AC optimal power flow formulations on large-scale power grids. URL: https://www.osti.gov/servlets/purl/1846582
Noebels M., Preece R., Panteli M., "AC Cascading Failure Model for Resilience Analysis in Power Networks," in IEEE Systems Journal, vol. 16, no. 1, pp. 374–385, March 2022, doi: 10.1109/JSYST.2020.3037400
Principal component analysis. URL: https://en.wikipedia.org/wiki/Principal_component_analysis
Varbella A., Gjorgiev B., Sansavini G. Geometric deep learning for online prediction of cascading failures in power grids. Reliability Engineering & System Safety, Volume 237, 2023. https://doi.org/10.1016/j.ress.2023.109341
Vaswani A., Shazeer N., Parmar N., Uszkoreit J., Jones L., Gomez A. N., Kaiser L., Polosukhin I. Attention Is All You Need. https://doi.org/10.48550/arXiv.1706.03762
Yazdani A., Jeffrey P. Complex network analysis of water distribution systems. https://doi.org/10.48550/arXiv.1104.0121
