Securing Power System Data in Motion by Timestamped Digital Text Watermarking
Siddhartha Deb Roy,Ankush Sharma,Saikat Chakrabarti,Sanjoy Debbarma
DOI: https://doi.org/10.1109/tsg.2024.3370892
IF: 10.275
2024-01-01
IEEE Transactions on Smart Grid
Abstract:Modern power systems rely on Wide Area Network (WAN) for real-time data transmission and control. However, integrating WAN introduces vulnerabilities that adversaries can exploit, leading to inaccurate information, compromised decisions, and potential grid instability. The Automatic Generation Control (AGC) system, responsible for maintaining power balance, is an attractive target for attackers due to its reliance on WANs for critical signal transmission. Traditional security mechanisms like encryption, multi-factor authentication, antivirus, etc., may not be suitable for power systems due to system priorities, processing limitations, and compatibility issues. To address this, we propose a novel digital text watermarking-based approach that ensures data integrity and detects replay attacks. The watermarking technique dynamically embeds a unique timestamped watermark into sensor data, enabling detection of unauthorised modifications. The algorithm confirms data integrity by comparing the extracted and expected watermark utilizing Damerau-Levenshtein index. Timestamp analysis is employed to assist in identifying replay attacks. Notably, our technique is independent of system scale, making it practical for real-world implementation. Furthermore, the framework can be extended to secure other power system applications, enhancing overall cyber resilience. Experimental results demonstrate the efficacy of our algorithm in AGC systems while ensuring almost zero performance loss to system operations.
engineering, electrical & electronic