Machine Learning in Smart Grid Security: A Survey on Cyber Threat Identification and Prevention Methods

Sandeep Gupta — Thu, 30 Oct 2025 00:00:00 +0000

The revolution of old power systems into smart grids has greatly improved efficiencies, reliabilities, and sustainability of electricity distribution due to the increased ability to use advanced communications technologies, distributed energy resources and digital infrastructures. Advanced cyber-attacks, such as fake data injection, denial-of-service, malware, and ransomware, have become more common despite this technological milestone leading to a stronger defense of vital assets, including AMI, SCADA, PMUs, and DERs. This research delves into how smart grids are utilizing ML techniques for cyber threat detection, specifically looking at how these methods might offer data-driven, adaptable, and scalable solutions. Models such as KNN, SVM, RF, GNN, Transformer encoders, and federated learning are evaluated for their proficiency in detecting both existing and new threats. ML techniques, such as supervised and unsupervised learning, as well as RL, are also considered. The study's findings show that these intelligent methods may significantly improve operational risk countering, live-time anomaly detection, and detection precision. They outline the problems with data quality, model interpretability, dependability, and data privacy, and provide solutions to address them. To make next-gen smart grid systems more cybersecurity resilient, the results demonstrate the importance of developing lightweight ML frameworks that are scalable and protect users' privacy when handling high-dimensional, multidimensional data.

Cybersecurity Spyware for System Privacy and Antivirus Testing

Mahe Mubeen Akhtar D., Abhishek S., Arathi M., Ankitha S., Chandana G. — Tue, 09 Sep 2025 00:00:00 +0000

The growing sophistication of cybersecurity threats, particularly through keylogging software, poses significant risks to individual privacy and organizational security. This work proposes an integrated solution that combines advanced surveillance techniques with robust security measures to detect and mitigate keylogging threats. The system captures and analyzes user activity through keystroke logging, screen capture, audio recording, and webcam snapshots, securely transmitting data to A Mailtrap server integrated with cloud storage and HeidiSQL for efficient management. Strong encryption ensures data. security, while the system’s dual-use nature highlights both its potential as a security tool and its privacy risks. By simulating real-world spyware, this research evaluates the effectiveness of antivirus software in detecting advanced keylogging threats, emphasizing the need for enhanced detection mechanisms and ethical frameworks to balance security and privacy. The goal is to provide a proactive framework for managing keylogging threats, ensuring a more secure digital environment.

Journal of IoT Security and Smart Technologies (e-ISSN: 2583-6226)

Machine Learning in Smart Grid Security: A Survey on Cyber Threat Identification and Prevention Methods

Cybersecurity Spyware for System Privacy and Antivirus Testing