Yogeshwari Yawalkar, Sakshi Vinod Borkar, Sohel Sharif Shaikh, Vijay Babasaheb Amrute

DOI: https://doi.org/10.5281/zenodo.23837428

Deep learning models now support load forecasting, fault detection, state estimation and renewable energy management in smart grids. Small, deliberately crafted changes to sensor data, known as adversarial attacks, can push these models to wrong answers and put grid reliability at risk. We reviewed twenty papers on adversarial attacks and defences in smart-grid and related energy systems, seventeen of them published from 2021 onwards, along with foundational work on FGSM, PGD and ensemble adversarial training. Attack success rates above 90% were reported against undefended detectors and adversarial training improved robustness considerably. Yet a defence trained against one attack often fails against another, most studies defend a single model and ensemble defences are rarely tested for sustainable grids. From these findings we draw four research gaps and outline a CNN–LSTM ensemble framework with adversarial training.

Adversarial Attacks, Deep Learning, Smart Grid, Ensemble Learning, Adversarial Training, Cybersecurity

This review looked at twenty papers on adversarial attacks and defences relevant to deep learning in smart grids. Deep learning is widely used for forecasting, fault detection, state estimation and theft detection and it helps integrate renewable energy. The studies show that these models can be fooled by small manipulations of their input, with attack success rates above 90% against undefended detectors [4]. Adversarial training is the most common defence and improves robustness considerably, but it is costly and a defence trained against one attack may not hold against another [4]. Ensemble and hybrid defences look promising but are under-evaluated [3] and most smart-grid work relies on a single model. The gaps identified here support developing a combined CNN–LSTM ensemble with adversarial training, tested against several attacks while keeping clean-data accuracy high. Such a framework could contribute to more secure and reliable AI in sustainable smart-grid systems.

1. Abdulaal, M. J., Ibrahem, M. I., Mahmoud, M. M. E. A., Khalid, J., Aljohani, A. J., Milyani, A. H., & Abusorrah, A. M. (2022). Real-time detection of false readings in smart grid AMI using deep and ensemble learning. IEEE Access, 10, 47541–47556.
2. Acharya, R., Al Sardy, L., Muhammad, M., & German, R. (2026). ADVIS-G: An adversarially defended intrusion detection system for smart grids using deep learning. Künstliche Intelligenz. https://doi.org/10.1007/s13218-026-00905-3
3. Askhatuly, A., Berdysheva, D., Berdyshev, A., Adamova, A., & Yedilkhan, D. (2025). Adversarial attacks and defense mechanisms in machine learning: A structured review of methods, domains and open challenges. IEEE Access, 13, 185145–185168. https://doi.org/10.1109/ACCESS.2025.3624409
4. El-Toukhy, A. T., Mahmoud, M. M. E. A., Bondok, A. H., Fouda, M. M., & Alsabaan, M. (2023). Countering evasion attacks for smart grid reinforcement learning-based detectors. IEEE Access, 11, 97373–97390. https://doi.org/10.1109/ACCESS.2023.3312376 7
5. Goodfellow, I. J., Shlens, J., & Szegedy, C. (2014). Explaining and harnessing adversarial examples. arXiv. https://arxiv.org/abs/1412.6572
6. Hamedani, K., Liu, L., Jagannath, J., & Yi, Y. (2021). Adversarial classification of the attacks on smart grids using game theory and deep learning. arXiv. https://arxiv.org/abs/2106.03209
7. Hao, J., & Tao, Y. (2022). Adversarial attacks on deep learning models in smart grids. Energy Reports, 8, 123–129. https://doi.org/10.1016/j.egyr.2021.11.026
8. Ilahi, I., Usama, M., Qadir, J., Janjua, M. U., Al-Fuqaha, A., Hoang, D. T., & Niyato, D. (2022). Challenges and countermeasures for adversarial attacks on deep reinforcement learning. IEEE Transactions on Artificial Intelligence, 3(2), 90–109. https://doi.org/10.1109/TAI.2021.3111139
9. Karim, F., Majumdar, S., & Darabi, H. (2021). Adversarial attacks on time series. IEEE Transactions on Pattern Analysis and Machine Intelligence, 43(10), 3309–3320.
10. Madry, A., Makelov, A., Schmidt, L., Tsipras, D., & Vladu, A. (2017). Towards deep learning models resistant to adversarial attacks. arXiv. https://arxiv.org/abs/1706.06083
11. Mahmoudnezhad, et al. (2024). Electric load forecasting under false data injection attacks via denoising deep learning and generative adversarial networks. IET Generation, Transmission & Distribution. https://doi.org/10.1049/gtd2.13273
12. Saber, A. M., Maheshwari, A., Youssef, A., & Kundur, D. (2025). Adversarial attacks on deep learning-based false data injection detection in differential relays. arXiv. https://arxiv.org/abs/2506.19302
13. Sampedro, G. A., Ojo, S., Krichen, M., Alamro, M. A., Mihoub, A., & Karovic, V. (2024). Defending AI models against adversarial attacks in smart grids using deep learning. IEEE Access, 12, 157408–157417. https://doi.org/10.1109/ACCESS.2024.3473531
14. Sánchez, G., Elbez, G., & Hagenmeyer, V. (2024). Attacking learning-based models in smart grids: Current challenges and new frontiers. In Proceedings of the 15th ACM International Conference on Future and Sustainable Energy Systems (pp. 589–595). https://doi.org/10.1145/3632775.3661984
15. Takiddin, A., Ismail, M., & Serpedin, E. (2023). Robust data-driven detection of electricity theft adversarial evasion attacks in smart grids. IEEE Transactions on Smart Grid, 14(1), 663–676.
16. Tian, J., et al. (2022). Adversarial attack and defense methods for neural network based state estimation in smart grid. IET Renewable Power Generation, 16(16), 3507–3518. https://doi.org/10.1049/rpg2.12334
17. Tian, J., Wang, B., Li, J., & Wang, Z. (2022). Adversarial attacks and defense for CNN based power quality recognition in smart grid. IEEE Transactions on Network Science and Engineering, 9(2), 807–819.
18. Tramèr, F., Kurakin, A., Papernot, N., Goodfellow, I., Boneh, D., & McDaniel, P. (2017). Ensemble adversarial training: Attacks and defenses. arXiv. https://arxiv.org/abs/1705.07204
19. Tusher, A. S., Rahman, M. A., Islam, M. R., Hossain, M. A., Anwar, A., & Hossain, M. J. (2025). Cyber-attack resilient data-driven approaches for early fault prediction system for wind turbines. Sustainable Energy Technologies and Assessments, 84.
20. Xu, W., & Teng, F. (2023). Availability adversarial attack and countermeasures for deep learning based load forecasting. arXiv. https://arxiv.org/abs/2301.01832

Yogeshwari Yawalkar, Sakshi Vinod Borkar, Sohel Sharif Shaikh, Vijay Babasaheb Amrute (2026). A Comprehensive Review of Robust AI Techniques for Securing Deep Learning Models in Smart Grid Systems. International Journal of Computer Techniques, 13(5), 618–625. ISSN: 2394-2231. DOI: https://doi.org/10.5281/zenodo.23837428