Resilient Self-Healing of Active Distribution Networks Using Physics-Informed Graph Reinforcement Learning
DOI:
https://doi.org/10.68050/JAMS.2026.410Keywords:
Active distribution networks, physics-informed graph reinforcement learning, resilient self-healing, renewable distributed generation, battery energy storage systems, feeder reconfiguration, voltage regulation; network resilience.Abstract
The rapid integration of renewable distributed generation (DG) into active distribution networks (ADNs) introduces substantial operational challenges arising from renewable intermittency, load variability, changing network topology, and feeder contingencies. These conditions require control strategies that can simultaneously maintain secure operation and rapidly restore service following disturbances. This paper proposes a Physics-Informed Graph Reinforcement Learning (PGRL) framework for resilient operation and self-healing of renewable-integrated ADNs. The proposed framework represents the distribution network as an electrical graph to capture bus connectivity and topology changes, while physics-informed decision guidance links the learning process to network operating constraints. Coordinated DG reactive-power control, battery energy storage system (BESS) operation, and feeder reconfiguration are integrated into a unified control strategy for both normal and post-fault operation. The framework is evaluated on the IEEE 33-bus ADN under variable photovoltaic generation, load variations, and feeder contingencies. Compared with the benchmark approaches, PGRL achieves a minimum voltage of 0.980 p.u., a maximum voltage deviation of 0.020, and reduces network power loss to 115.7 kW, corresponding to a 46.6% reduction. Under a feeder outage between buses 9 and 10, PGRL restores the affected load within 11 min, limits energy not supplied (ENS) to 65 kWh, and achieves a resilience index of 0.97. Furthermore, stable learning convergence is reached at approximately 540 episodes. The results demonstrate that integrating topology-aware graph learning with physics-guided control and coordinated DG-BESS operation can substantially improve the operational efficiency, adaptability, and recovery capability of renewable-rich ADNs.
References
Behzadi, S., Bagheri, A., & Rabiee, A. (2024). Optimal operation of reconfigurable active distribution networks aiming at resiliency improvement. Electric Power Systems Research, 230, 110169. https://doi.org/10.1016/j.epsr.2024.110169.
2. Byeon, G., & Kim, K. (2024). Distributionally robust decentralized Volt-Var control with network reconfiguration. IEEE Transactions on Smart Grid, 15(5), 4705–4718,. https://doi.org/10.1109/TSG.2024.3377910.
3. Dominguez-Garcia, A. D., Zholbaryssov, M., Amuda, T., & Ajala, O. (2024). An online feedback optimization approach to voltage regulation in inverter-based power distribution networks. IEEE Transactions on Power Systems, 39(1), 1145–1156,. https://doi.org/10.1109/TPWRS.2023.3304112.
4. Glavic, M. (2019). Deep) reinforcement learning for electric power system control and related problems: A short review and perspectives. Annual Reviews in Control, 48, 22–35,. https://doi.org/10.1016/j.arcontrol.2019.09.008.
5. Glavic, M., Moreno, R., Bi, T., & Shahidehpour, M. (2021). Reinforcement learning for electric power system decision and control: Past considerations and perspectives. IEEE Transactions on Smart Grid, 12(6), 5128–5142,. https://doi.org/10.1109/TSG.2021.3070934.
6. Huang, Y., Yang, Q., Tan, J., & Ukil, A. (2022). Deep reinforcement learning for real-time energy management in smart grids. Energy Reports, 8, 1245–1257,. https://doi.org/10.1016/j.egyr.2022.08.190.
7. Li, B., & Xu, Q. A. (2024). A machine learning-assisted distributed optimization method for inverter-based Volt-VAR control in active distribution networks. IEEE Transactions on Power Systems, 39(2), 2668–2681,. https://doi.org/10.1109/TPWRS.2023.3279303.
8. Qiu, W., Yadav, A., You, S., Dong, J., Kuruganti, T., Liu, Y., & Yin, H. (2024). Neural networks-based inverter control: Modeling and adaptive optimization for smart distribution networks. IEEE Transactions on Sustainable Energy, 15(2), 1039–1049,. https://doi.org/10.1109/TSTE.2023.3324219.
9. Wang, S., Hou, Y., Guan, X., Liu, S., & Huo, Z. (2025). Resiliency-informed optimal scheduling of smart distribution network with urban distributed photovoltaic: A stochastic P-robust optimization. Energy, 313, 133449. https://doi.org/10.1016/j.energy.2024.133449.
10. Xu, Y., Dong, Z. Y., Zhang, R., & Hill, D. J. (2017). Multi-timescale coordinated voltage/VAR control of high renewable-penetrated distribution systems. IEEE Transactions on Power Systems, 32(6), 4398–4408,. https://doi.org/10.1109/TPWRS.2017.2659779.
11. Yin, C., Dong, J., & Zhang, Y. (2025). Distributionally robust bilevel optimization model for distribution network with demand response under uncertain renewables using Wasserstein metrics. IEEE Transactions on Sustainable Energy, 16(1), 118–130,. https://doi.org/10.1109/TSTE.2024.3509314.
12. Zhang, B., Cao, D., Hu, W., Ghias, A. M. Y. M., & Chen, Z. (2024). Physics-informed multi-agent deep reinforcement learning enabled distributed voltage control for active distribution network using PV inverters. International Journal of Electrical Power & Energy Systems, 155, 109641. https://doi.org/10.1016/j.ijepes.2023.109641.
13. Zhang, C., Xu, Y., Dong, Z. Y., & Wong, K. P. (2018). Robust coordination of distributed generation and energy storage in distribution networks. IEEE Transactions on Smart Grid, 9(2), 1488–1499,. https://doi.org/10.1109/TSG.2016.2583461.
14. Zhang, Z., Wang, J., & Chen, B. (2021). Deep reinforcement learning for energy management in microgrids. Applied Energy, 288, 116618. https://doi.org/10.1016/j.apenergy.2021.116618.
15. Zhao, J., Zheng, T., & Litvinov, E. (2016). A unified framework for defining and measuring flexibility in power system. IEEE Transactions on Power Systems, 31(1), 339–347,. https://doi.org/10.1109/TPWRS.2015.2390039.
16. Zhou, K., Yang, S., & Shao, Z. (2021). Digital twin-based optimization for energy systems. Applied Energy, 285, 116419. https://doi.org/10.1016/j.apenergy.2020.116419.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Author(s)
This work is licensed under a Creative Commons Attribution 4.0 International License.
Articles published in the Journal of Advanced Multidisciplinary Studies (JAMS) are licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0), unless otherwise stated. Authors retain copyright of their work and grant JAMS the right of first publication.
