The coupling of the legacy power infrastructure with complex computer networks and the integration of communication devices has led to the problem of cyberattacks, which in turn may induce partial/full blackouts. In this regard, an approach based on a Decision Tree for the detection and mitigation of Denial-of-Service (DoS) attacks is proposed in this paper. Based on the patterns and effects observed during DoS attacks on the Synchrophasor data flowing from the Phasor Measurement Unit (PMU) to the Phasor Data Concentrator (PDC), various supervised machine learning models are employed, which classify healthy and attack data and detect DoS attacks. Thereafter, a recurrent deep learning network employing Gated Recurrent Unit (GRU) architecture is utilized to reconstruct the lost data, which in turn is helpful in mitigating the impacts of a DoS attack. The suggested algorithms are validated using a real-world scenario of a communication dataflow between multiple PMUs placed at different buses of a power transmission network and a PDC present at a control center, representing a wide-area network, and are incorporated using Real Time Digital Simulator and openPDC. By comparing various metrics of ML classification, the results indicate that the decision tree based model gives better accuracy along with the least detection time compared to other machine learning-based models. The suggested GRU-based mitigation algorithm can reconstruct the DoS-impacted lost data from the available pre-attack phasors.
Denial-of-Service Attack Detection and Mitigation Strategy for Synchrophasor Data Flowing in Power Transmission System
2024-07-31
720266 byte
Conference paper
Electronic Resource
English
Automatic Detection and Mitigation of Denial-of-Service Attacks
European Patent Office | 2022
|E-business Denial of service attack takes down WorldPay
British Library Online Contents | 2003
Secure Data Forwarding against Denial of Service Attack Using Trust Based Evolutionary Game
British Library Conference Proceedings | 2008
|