IoT-driven anomaly detection in smart grids using multimodal deep learning models

Abstract

The increasing complexity of modern power systems, driven by the integration of Internet of Things (IoT) devices and distributed energy resources, has amplified the need for robust anomaly detection mechanisms in smart grids. This study proposes an IoT-driven multimodal deep learning framework that integrates time series data from SCADA/PMU systems, environmental sensor readings, and thermal/visual images to enhance anomaly detection performance. The proposed architecture combines Long Short Term Memory (LSTM) networks for temporal modeling, Convolutional Neural Networks (CNNs) for spatial feature extraction, and a hybrid feature fusion strategy to exploit complementary information across modalities. Experiments conducted on benchmark datasets demonstrate that the framework outperforms traditional machine learning and single modality deep learning methods, achieving an F1 score of 0.88 and a ROC AUC of 0.94. These results confirm the potential of multimodal deep learning to improve the reliability, resilience, and situational awareness of smart grids.

Keywords

• Smart grids
• Anomaly detection
• Internet of Things (IoT)
• Multimodal deep learning
• Hybrid feature fusion

References

1. [1] Banik S, Saha SK, Banik TB, Hossain SMM. Anomaly detection techniques in smart grid systems: a review. IEEE World AI IoT Congress 2023.
2. [2] Guato Burgos J, González J, Rivas D. A review of smart grid anomaly detection approaches pertaining to artificial intelligence. Applied Sciences 2023; 14(3): 1194.
3. [3] Pang G, Shen C, Cao L, van den Hengel A. Deep learning for anomaly detection: a review. ACM Computing Surveys 2021; 54(2): 1-38.
4. [4] Zhang JE, Wu D, Boulet B. Time-series anomaly detection for smart grids: a survey. IEEE Access 2022; 10: 129481-129497.
5. [5] Zhou F, Wen G, Ma Y, Geng H, Huang R, Pei L, Yu W, Chu L, Qiu R. A comprehensive survey for deep learning based abnormality detection in smart grids with multimodal image data. Applied Sciences 2022; 12(11): 5336.
6. [6] Cui J, Li F, Jiang J. Anomaly detection for cybersecurity in smart grids: a survey. IEEE Transactions on Smart Grid 2019; 10(6): 5724-5735.
7. [7] Zangrando N, Fraternali P, Petri M. Anomaly detection in quasi periodic energy consumption data series. Energy Informatics 2022; 5: 62.
8. [8] Chatterjee A, Ahmed BS. IoT anomaly detection methods and applications: a comprehensive review. Journal of Network and Computer Applications 2023; 213: 103600.

9. [9] Zheng S, Ma X, Wang R. Multi-model fusion for energy metering anomaly detection. Measurement 2025; 216: 112893.
10. [10] Jabbari Zideh A, et al. Physics-informed machine learning for anomaly detection in smart grids. arXiv 2023; arXiv:2309.10788.
11. [11] Giraldo J, Cardenas A, Quijano N. Integrity attacks and anomaly detection in smart grids. IEEE Transactions on Smart Grid 2017; 8(5): 2414-2422.
12. [12] Matthews B, O’Neill M, Clark A. Real time synchrophasor data anomaly detection using machine learning. IEEE Transactions on Smart Grid 2019; 10(3): 3046-3055.
13. [13] Sharma P, Bedi P. ML based anomaly detection in SCADA IoT systems. International Journal of Advanced Computer Science 2021; 12(6): 768-777.
14. [14] Li W, Luo Y, Sun Y. SCADA data driven anomaly detection for smart grids using LSTM networks. Energies 2020; 13(9): 2290.
15. [15] Ahmed M, Mahmood AN, Hu J. A survey of network anomaly detection techniques. Journal of Network and Computer Applications 2016; 60: 19-31.
16. [16] Mao Z, Zhou B, Huang J, Liu D, Yang Q. Anomaly detection model for power consumption data based on time-series reconstruction. Energies 2024; 17(19): 4810.
17. [17] Wen Y, Wang Q, Gao Y. Hybrid CNN LSTM for anomaly detection in smart grid time series. IEEE Access 2023; 11: 45678-45689.
18. [18] Kang M, Kim H, Choi D. Image based anomaly detection for electrical equipment using deep CNNs. IEEE Sensors Journal 2021; 21(13): 15264-15273.
19. [19] Ma X, Liu Z, Guo J. Transformer GAN for multivariate time series anomaly detection in smart grids. Frontiers in Energy Research 2024; 12: 1364456.
20. [20] Wang J, et al. Attention enhanced federated learning for privacy preserving anomaly detection in smart grids. Computers & Electrical Engineering 2025; 116: 109083.
21. [21] Silva R, Almeida M, Gomes L. Hybrid feature fusion for multimodal anomaly detection in power systems. Electric Power Systems Research 2022; 212: 108479.
22. [22] Li X, Wang Z, Huang K. Late fusion strategies for multimodal deep learning in smart grid monitoring. Neurocomputing 2023; 518: 162-173.
23. [23] Xu W, Chen Z. Federated deep learning for collaborative anomaly detection in power networks. IEEE Internet of Things Journal 2022; 9(14): 12837-12849.
24. [24] Guo W, Zhang Y. Multimodal representation learning: methods and applications. Information Fusion 2019; 57: 1-14.
25. [25] Duan J, Li Y, Wang S, Zhao X. Transformer-GAN for anomaly detection in power systems. Frontiers in Energy Research 2024; 12: 1364456.
26. [26] Patrizi G, Rossi L, Bianchi F. Anomaly detection for power quality using smart meters. Sensors 2024; 24(3): 1124.
27. [27] Zhao H, Liu Y, Chen Z. Decentralized federated anomaly detection in smart grids. arXiv 2024; arXiv:2403.01891.
28. [28] Wang J, Li X, Sun Y. Federated learning for smart grids: a comprehensive survey. IEEE Communications Surveys & Tutorials 2024; Early Access.
29. [29] Guato Burgos MF, et al. A review of smart grid anomaly detection approaches. Applied Sciences 2024.