Abstract

The increasing integration of wind energy and the reliance on digital communication systems for the remote operation of wind farms (WFs) have significantly expanded the grid’s vulnerability to cyberattacks. These risks are particularly concerning for power grids with degrading system strength (SS) and high penetration of converter-based resources, where cyber-induced instabilities threaten grid stability. This thesis investigates critical aspects of cyber risks in wind-integrated power grids, focusing on attack modeling and prevention, risk management, and mitigation strategies.
The first part introduces a cyberattack model targeting SS provision to permanent magnet synchronous generator (PMSG)-based WFs. To address this, an anomalous command verification (ACV) module is proposed to detect and prevent cyber-induced converter-driven instabilities by estimating SS buffer capacity. The second part develops a cyber-informed risk management framework using Bayesian attack graphs and distributionally robust optimization (DRO) to quantify and mitigate cyber risks. This framework incorporates converter-driven stability support (CDSS) from WFs and is validated on the IEEE 118-bus system. The third part focuses on mitigating attack-induced instabilities in offshore wind farms (OWFs) connected via High Voltage Direct Current (HVDC) systems. It introduces a physics-informed iterative control (PIC) strategy based on neural networks to reduce instability magnitudes progressively.
These contributions provide an integrated approach to addressing cyber risks in wind-integrated power grids, enhancing their security and stability in the face of evolving threats.