Unified Optimization for Offshore Wind Farm Design

Offshore wind energy is key in the transition to sustainable and renewable energy sources. Improving the design of offshore wind farms contributes to ensuring affordable energy supply, reducing greenhouse gas emissions, and fighting climate changes.

There are two main challenges in the design of offshore wind farms, which are typically solved in sequence: first, placing the turbines to maximize their power production, and second, connecting the turbines with electrical cables. In the first problem, turbines are preferably placed apart as much as possible to reduce production losses due to the wake effect. In the second problem, instead, turbines are preferably positioned close together, to reduce total cable lengths and hence costs. We can exploit the opposing tendencies of the problems by unifying them in a combined optimization. Thus, we can balance these contradicting tendencies and design wind farms with higher power production and lower costs.

In this thesis, we apply Operations Research techniques to optimize the design of wind farms. First, we develop, for the separate problems, fast and scalable heuristics that match and improve results from the literature. Then, we address the unified problem, developing novel methods to combine the separate heuristics. Compared to the traditional sequential approach, the combined optimization significantly improves the business case of a wind farm, by considerably lowering the investment costs, while only slightly affecting the lifetime revenues from power production.

We also consider additional aspects of wind farm design, integrating them into the optimization. We propose a multilevel framework to select the best region for a wind park and then, for the first time, optimize the shape and area of the wind farm. Finally, we explore alternative technologies for wind farms by studying Vertical Axis Wind Turbines. These types of turbines generate asymmetric wakes, depending on the direction in which they rotate. We present the first formulation to choose the spinning of Vertical Axis turbines, clockwise or counterclockwise. This new model can thus exploit the wake asymmetry in the wind farm layout.

Thanks to the collaboration with Vattenfall, offshore developer and partner of this project, the optimization tools that we developed have been designed to handle the constraints of reallife cases. The novel heuristics are already being applied to design future wind farms and come with a proven track record of significantly improved wind farm business cases.

We hope that this thesis work will promote innovation in the industry and contribute to developing the full potential of wind energy.