PowerGNN: Using Spatial Intelligence for Autonomous Grid Resilience

Aaryyan Pradhan; Soumyajeet Biswal; Devpriya Panda; Kavita Sharma; Neelamadhab Padhy

doi:10.65890/race.v1i2.157

Authors

Aaryyan Pradhan Department of CSE, IIT Madras, India Author
Soumyajeet Biswal Department of CSE, ITER, S'O'A Deemed to be University, Odisha, India Author
Devpriya Panda ITER, S‘O’A Deemed to be University, Odisha Author
Kavita Sharma Department of CSE, Galgotias College of Engineering and Technology, Greater Noida, India Author
Neelamadhab Padhy Department of CSE, GIET University, Odisha, India Author

DOI:

https://doi.org/10.65890/race.v1i2.157

Keywords:

Multi-Agent Reinforcement Learning, Graph Neural Networks, Power System Restoration, Decentralised Control, N-k Contingency, Cyber-Physical Security

Abstract

The increasing amount of extreme weather conditions and the increasing complexity of cyber-physical threats pose a significant threat to the resilience of modern power distribution systems. The conventional approaches to restoration rely on centralised control and information flow across the globe and therefore, are susceptible to failure at one point and delays in the event of a major disaster. This paper presents a decentralised Multi-Agent Reinforcement Learning (MARL) algorithm with Graph Neural Networks (GNNs) to restore the power grid in real time and independently. The inductive bias of GNNs allows every substation agent to learn to jointly plan switching actions and power dispatch depending on local topological characteristics and neighbour messages, eliminating the need to have a central supervisor. The suggested framework is assessed using some IEEE standard test systems under various load stressors, including N-k contingencies, blackouts in communication, and adversarial False Data Injection (FDI) attacks. The methodology is aimed at creating a policy that is agnostic to grids and puts forward the priority of restoring critical loads, but assuring voltage stability by the use of localised spatial intelligence. This work shows the theoretical and practical benefits of moving from complex centralised optimisation to a scalable, O(N) decentralised graph-inference model. By demonstrating that learned policies can transfer across different topologies, this research offers a strong foundation for the next generation of self-healing, “dark-start” resilient smart grids that can effectively handle the challenging environment of post-disaster recovery.

References

[1] P. Kundur, J. Paserba, V. Ajjarapu, et al., “Definition and classification of power system stability,” IEEE Transactions on Power Systems, vol. 19, no. 3, pp. 1387–1401, Aug. 2004. DOI: https://doi.org/10.1109/TPWRS.2004.825981

[2] M. Panteli and P. Mancarella, “The grid: Stronger, bigger, smarter? Presenting a resilience framework for sustainable energy systems,” IEEE Power and Energy Magazine, vol. 13, no. 3, pp. 58–66, May–Jun. 2015. DOI: https://doi.org/10.1109/MPE.2015.2397334

[3] N. H. El-Amary and Y. J. Wang, “Power system restoration: A survey,” International Journal of Electrical Power and Energy Systems, vol. 32, no. 6, pp. 545–555, 2010.

[4] Y. Liu, R. Fan, and V. Terzija, “Power system restoration: A literature review from 2006 to 2016,” Journal of Modern Power Systems and Clean Energy, vol. 4, no. 3, pp. 332–341, 2016. DOI: https://doi.org/10.1007/s40565-016-0219-2

[5] C. Chen, J. Wang, and D. Ton, “Modernizing the grid: A review of resilience-related metrics,” Applied Energy, vol. 192, pp. 132–145, Apr. 2017.

[6] V. Mnih et al., “Human-level control through deep reinforcement learning,” Nature, vol. 518, no. 7540, pp. 529–533, 2015. DOI: https://doi.org/10.1038/nature14236

[7] T. P. Lillicrap et al., “Continuous control with deep reinforcement learning,” arXiv:1509.02971, 2015.

[8] J. Schulman, F. Wolski, P. Dhariwal, A. Radford, and O. Klimov, “Proximal policy optimization algorithms,” arXiv:1707.06347, 2017.

[9] R. Lowe et al., “Multi-agent actor-critic for mixed cooperative-competitive environments,” in Proc. Advances in Neural Information Processing Systems (NeurIPS), 2017.

[10] Y. Zhang, Q. Yang, W. Itani, and S. Cui, “Deep reinforcement learning for power system applications: A review,” IEEE Access, vol. 8, pp. 152857–152869, 2020. DOI: https://doi.org/10.1109/ACCESS.2019.2961914

[11] T. N. Kipf and M. Welling, “Semi-supervised classification with graph convolutional networks,” in Proc. Int. Conf. on Learning Representations (ICLR), 2017.

[12] P. Veličković, G. Cucurull, A. Casanova, A. Romero, P. Liò, and Y. Bengio, “Graph attention networks,” in Proc. Int. Conf. on Learning Representations (ICLR), 2018.

[13] J. Gilmer, S. S. Schoenholz, P. F. Riley, O. Vinyals, and G. E. Dahl, “Neural message passing for quantum chemistry,” in Proc. Int. Conf. on Machine Learning (ICML), 2017.

[14] W. Hamilton, Z. Ying, and J. Leskovec, “Inductive representation learning on large graphs,” in Proc. Advances in Neural Information Processing Systems (NeurIPS), 2017.

[15] J. Zhou et al., “Graph neural networks: A review of methods and applications,” AI Open, vol. 1, pp. 57–81, 2020. DOI: https://doi.org/10.1016/j.aiopen.2021.01.001

[16] K. Zhang, Z. Yang, and T. Başar, “Multi-agent reinforcement learning: A selective overview of theories and algorithms,” in Handbook of Reinforcement Learning and Control, pp. 321–384, 2021. DOI: https://doi.org/10.1007/978-3-030-60990-0_12

[17] D. Biagioni et al., “Learning-based optimal power flow using graph neural networks,” IEEE Power and Energy Technology Systems Journal, vol. 9, no. 1, pp. 1–12, Mar. 2022.

[18] L. Yang, Y. Li, and J. Wang, “Graph-based deep reinforcement learning for decentralized service restoration in distribution systems,” IEEE Transactions on Smart Grid, vol. 12, no. 6, pp. 5207–5218, Nov. 2021.

[19] B. Chen, Z. Wang, and M. Yue, “A decentralized restoration strategy for distribution systems with high DER penetration,” IEEE Transactions on Smart Grid, vol. 11, no. 4, pp. 3211–3221, Jul. 2020.

[20] X. Jiang, J. Wang, and G. Liu, “Graph convolutional reinforcement learning for resilience-oriented microgrid formation,” IEEE Transactions on Power Systems, vol. 36, no. 4, pp. 3584–3595, Jul. 2021.

[21] Y. Liu, P. Ning, and M. K. Reiter, “False data injection attacks against state estimation in electric power grids,” ACM Transactions on Information and System Security, vol. 14, no. 1, pp. 1–33, 2011. DOI: https://doi.org/10.1145/1952982.1952995

[22] M. G. Dooms, “Cyber-physical security of the power grid: A review,” Renewable and Sustainable Energy Reviews, vol. 158, Art. no. 112090, 2022. DOI: https://doi.org/10.1016/j.rser.2022.112090

[23] O. Kosut, L. Jia, R. J. Thomas, and L. Tong, “Malicious data attacks on smart grid state estimation: Attack strategies and countermeasures,” IEEE Transactions on Smart Grid, vol. 2, no. 4, pp. 745–758, Dec. 2011. DOI: https://doi.org/10.1109/TSG.2011.2163807

[24] J. Yan, B. Tang, and H. He, “Detection of false data injection attacks in smart grid with supervised learning,” IEEE Transactions on Smart Grid, vol. 7, no. 3, pp. 1395–1403, May 2016. DOI: https://doi.org/10.1109/IJCNN.2016.7727361

[25] R. D. Zimmerman, C. E. Murillo-Sánchez, and R. J. Thomas, “MATPOWER: Steady-state operations, planning, and analysis tools for power systems research and education,” IEEE Transactions on Power Systems, vol. 26, no. 1, pp. 12–19, Feb. 2011. DOI: https://doi.org/10.1109/TPWRS.2010.2051168

[26] L. Thurner et al., “‘pandapower’—An open-source Python tool for convenient modeling, analysis, and optimization of electric power systems,” IEEE Transactions on Power Systems, vol. 33, no. 6, pp. 6510–6521, Nov. 2018. DOI: https://doi.org/10.1109/TPWRS.2018.2829021

[27] IEEE Power and Energy Society, “IEEE PES Test Systems Resources.” [Online]. Available: http://sites.ieee.org/pes-testsystems/

[28] Y. Wang and A. M. Moore, “A scalable graph neural network approach for large-scale power system state estimation,” IEEE Transactions on Power Systems, vol. 38, no. 5, pp. 4421–4432, Sept. 2023.

[29] F. Blaabjerg, Y. Yang, D. Yang, and X. Wang, “Role of power electronics in future power systems,” IEEE Journal of Emerging and Selected Topics in Power Electronics, vol. 5, no. 2, pp. 502–514, Jun. 2017.

[30] J. J. Justo, F. Mwasilu, J. Lee, and J. W. Jung, “AC-microgrids versus DC-microgrids with distributed energy resources: A review,” Renewable and Sustainable Energy Reviews, vol. 24, pp. 387–405, Aug. 2013. DOI: https://doi.org/10.1016/j.rser.2013.03.067