A Machine Learning Based Cyber Attack Detection Model for Wireless Sensor Networks in Microgrids

Abstract

The expanding dependence on remote sensor systems (WSNs) inside microgrids has underscored the basic require for strong security components, given their helplessness to different cyber-attacks . This study introduces a machine learning-based cyberattack detection model tailored for WSNs in microgrids, utilizing a comprehensive dataset from Kaggle. Our demonstrate coordinating a assorted set of calculations, counting Convolutional Neural Systems (CNN), Detached Forceful Classifiers, Arbitrary Timberland Classifiers, and XGBoost Classifiers, to guarantee tall exactness and productivity in recognizing peculiarities. By nourishing the framework with organize information, it can precisely classify the organize state as either typical or beneath one of three particular assault sorts: grayhole, blackhole, or flooding assault. This multifaceted approach not as it were upgrades the discovery .

Introduction

As the global energy sector shifts toward renewable and decentralized systems, microgrids become essential for efficient energy management and distribution. Wireless Sensor Networks (WSNs) within microgrids enable real-time monitoring and control but rely on open communication channels, making them vulnerable to cyber-attacks like grayhole, blackhole, and flooding. These attacks disrupt network communication, causing operational failures and financial losses. Traditional security methods struggle to detect such dynamic and evolving threats effectively.

To address this, the project proposes a machine learning-based cyberattack detection system tailored for WSNs in microgrids. It leverages advanced algorithms—including Convolutional Neural Networks (CNN), Passive Aggressive Classifiers, Random Forest, and XGBoost—to accurately identify and classify cyber threats. Each algorithm contributes uniquely: CNN excels in pattern recognition, Passive Aggressive Classifier adapts quickly to new threats, Random Forest enhances decision robustness, and XGBoost improves prediction accuracy while reducing false positives.

The system architecture involves collecting network traffic data, preprocessing it (including normalization and augmentation), splitting it into training/testing sets, training multiple models, and evaluating them through metrics like accuracy and F1-score. The best model is then deployed for real-time detection to safeguard microgrid operations against cyberattacks, ensuring reliable and continuous energy distribution.

Conclusion

The proposed system extraordinarily identifies cyberattacks in WSNs inside microgrids employing a combination of CNN, Detached Forceful Classifiers, Arbitrary Woodlands, and XG Boost. It goes past conventional strategies by identifying unpretentious deviations rather than depending exclusively on known assault marks, tending to dangers like grayhole, blackhole and flooding. CNNs empower effective highlight extraction to reveal complex assault designs and irregularities in organize activity. Detached Forceful Classifiers rapidly adjust to modern assaults by overhauling choice boundaries with approaching information. Irregular Woodlands upgrade exactness through numerous choice trees, decreasing wrong positives. XGBoost progresses decision-making and minimizes classification mistakes, optimizing by and large execution. This collaboration guarantees exact, real-time cyber danger location and fortifies microgrid security. The versatile learning demonstrate underpins progressing advancement, dealing with advancing assault procedures. By moving past rule-based frameworks, it empowers proactive, brilliantly defense instruments. Eventually, the demonstrate contributes to independent, self-sustaining cybersecurity in basic foundation.

References

[1] F. Pasqualetti, F. Dorfler, and F. Bullo, “Attack detection and identification in cyber-physical systems,” IEEE Trans. Autom. Control, vol. 58, no. 11, pp. 2715–2729, 2013. [2] C. Kwon, W. Liu, and I. Hwang, “Security analysis for cyber-physical systems against stealthy deception attacks,” in Proc. Amer. Control Conf., IEEE, 2013, pp. 3344–3349. [3] M. Pajic et al., “Design and implementation of attack-resilient cyberphysical systems,” IEEE Control Syst. Mag., vol. 37, no. 2, pp. 66–81, 2017. [4] M. Ozay et al., “Machine learning methods for attack detection in the smart grid,” IEEE Trans. Neural Netw. Learn. Syst., vol. 27, no. 8, pp. 1773–1786, 2015. [5] Y. LeCun, Y. Bengio and G. Hinton, “Deep learning,” Nature, vol. 521, no. 7553, pp. 436–444, 2015. [6] K. Crammer et al., “Online Passive-Aggressive Algorithms,” Journal of Machine Learning Research, vol. 7, pp. 551-585, 2006. [7] L. Breiman, “Random forests,” Machine Learning, vol. 45, no. 1, pp. 5-32, 2001. [8] T. Chen and C. Guestrin, “XGBoost: A scalable tree boosting system,” in Proc. 22nd ACM SIGKDD Intl. Conf. on Knowledge Discovery and Data Mining, 2016, pp. 785-794. [9] A. Jindal et al., “Machine Learning-Based Cyber Threat Detection in WSNs,” Neural Computing and Applications, Springer, 2023. [10] Ozay, Mete, Inaki Esnaola, Fatos Tunay Yarman Vural, Sanjeev R. Kulkarni, and H. Vincent Poor. “Machine learning methods for attack detection in the smart grid.” IEEE transactions on neural networks and learning systems 27, no. 8 (2015): 1773-1786. [11] Tianfield, Huaglory. “Data mining based cyber-attack detection.” System simulation technology 13, no. 2 (2017): [12] 90-104. [13] Pasqualetti, Fabio, Florian Dorfler, and Francesco Bullo. “Attack detection and ¨ identification in cyber- physical systems.” IEEE Transactions on Automatic Control 58, no. 11 (2013): 2715-2729. [14] Wireless Sensor Network Security: Real-Time Detection and Prevention of Attacks – Discusses real-time security strategies in WSNs, focusing on detection and mitigation mechanisms. [15] Intrusion Detection System for Wireless Sensor Networks: A Machine Learning-Based Approach – Explores machine learning-based intrusion detection for WSN security. [16] Gharavi, H. “Cyber-Physical Security for Distributed Smart Grid.” Proceedings of the IEEE, 2019. [17] Cui, J. et al. “Cyber-Attack Detection for Wireless Sensor Networks in Smart Grids.” Sensors, 2020. [18] Zarpelão, B. B., et al. “A survey of intrusion detection in WSNs.” Journal of Network and Computer Applications, 2017. [19] Yang, T., et al. “Anomaly Detection in WSNs using Deep Learning.” IEEE Internet of Things Journal, 2022. [20] Jindal, A. et al. “Machine Learning-Based Cyber Threat Detection in WSNs.” Springer Neural Computing and Applications, 2023. [21] Rao, R. et al. “Resilient Cybersecurity Framework for Wireless Sensor Networks in Smart Cities.” ACM Transactions on Sensor Networks, 2021. [22] Mahbub, K. et al. “Blockchain-Enabled Security Model for Wireless Sensor Networks.” Journal of Cybersecurity and Privacy, 2023. [23] Lin, C., et al. “AI-Based Intrusion Detection in Wireless Networks.” IEEE Transactions on Information Forensics and Security, 2024. [24] T. Yang et al., “Anomaly detection in WSNs using deep learning,” IEEE Internet of Things Journal, vol. 9, no. 2, pp. 1504–1514, 2022. [25] H. Tianfield, “Data mining-based cyber-attack detection,” System Simulation Technology, vol. 13, no. 2, pp. 90–104, 2017.

Copyright

Copyright © 2025 Dr. M. V. Vijaya Saradhi, Mende Puja, B Anjali, Lavudya Harshitha, Karre Bharadwaj. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.