Floating offshore wind turbines extend wind energy into deep-water sites that bottom-fixed foundations cannot reach. Unlike a bottom-fixed machine, a floating turbine does not hold a fixed position. Under thrust, each platform settles at a mean offset from its nominal location, a drift referred to here as passive repositioning. Wake steering adds a second source of motion. When a turbine yaws to deflect its wake, the cross-wind component of the thrust also pushes the platform sideways, shifting the rotor to a new position, an effect referred to as active repositioning. The layout the wind actually sees is therefore not the one drawn at nominal positions, and the wake interactions across the farm move with it. This motivates accounting for platform displacement inside both the layout and the control optimization. Building on the integrated layout and control co-design frameworks developed for bottom-fixed farms, this thesis extends the approach to floating wind turbines.

The resulting framework accounts for both passive and active repositioning within a single optimization loop. One XGBoost surrogate predicts the surge and sway displacement of each platform, and a second predicts the optimal yaw angle directly, which removes the nested control optimization that would otherwise run at every candidate layout. The framework is assessed on the Kriti~3 site off Crete, with 12 IEA 15~MW turbines on the VolturnUS-S semi-submersible and three mooring designs ranging from stiff to highly compliant.

Accounting for passive repositioning inside the layout optimization yields only a small median gain in annual energy production (AEP) over fixed-position optimization, though the gain increases with mooring compliance, and on the most compliant design the displacement-aware approach reaches a distinctly better layout than fixed-position optimization finds. The stronger effect is on the control side. On that same compliant design, yaw setpoints optimized under a fixed-position assumption turn a predicted gain into a net loss once applied to the moving platform, since the yaw-induced drift carries the rotors into wakes the fixed-position optimizer never evaluated. A displacement-aware optimization not only avoids this loss but unlocks further gain, using the same yaw-induced motion to steer the platforms clear of upstream wakes. The displacement response must therefore be accounted for inside the control optimization, and the requirement grows stronger the more compliant the mooring.

The two surrogates keep the layout-control co-design computationally tractable, where a nested formulation running the actual yaw optimizer and platform simulations inside the loop would be prohibitive. At this site and farm size, fixed-position, displacement-aware, and co-design optimization reach nearly the same AEP, with co-design adding little over displacement-aware layout optimization, in line with co-design results reported for bottom-fixed farms. Because many distinct layouts reach a near-equal AEP, the designer keeps the freedom to choose among them on criteria beyond energy capture, such as cabling or structural loads. ... Read more