Colloquium: Vibhor Rajoria (FPT)

Wake losses in wind farms, caused by closely spaced turbines, reduce downstream power production and overall efficiency. To address this, this thesis develops a real-time, model-based controller for wake steering, focusing on adapting to dynamic operational and environmental challenges in large-scale wind farms. The controller adjusts yaw angles to misalign turbines with incoming wind, redirecting wakes to boost overall power output. Unlike traditional controllers, it adapts to off-design conditions like changing atmospheric factors and offline turbines, enhancing adaptability. Engineering wake models are modified to account for heterogeneous wind conditions, while improvements to the Serial-Refine optimization strategy enable real-time application. Results show that adapting wake models with sensor data improves power predictions in non-uniform winds, and the distributed optimization framework for real-time yaw performs comparably to centralized methods. Including offline turbines further increases power, advancing wind farm control strategies and boosting revenue in dynamic site conditions.

Supervisor: Assistant Professor Wei Yu, Professor J.W. van Wingerden