Posters

Abstract

In the context of wind farm design, layout optimization is a critical yet complex and time-consuming stage, as it must integrate multidisciplinary aspects and involves an evolving and iterative process. For offshore wind farms, alignment constraints are often imposed by regulatory bodies or societal considerations. The current practice typically involves manually generating a grid based on the wind rose and distributing the turbines on this grid. However, IFPEN has developed a new algorithm to numerically optimize the wind farm layout while accounting for these alignment constraints. GreenWITS is currently industrializing this solution in a new software called OptiWITS. To respond to the competitive context of wind industry, the generation of initial wind farm layouts must be accomplished within a reasonable computation time and with advanced wake models. However, the optimization problem's parameter space is vast (6 variables for grid parametrization, with numerous combinations for locating N wind turbines among X grid intersections), requiring millions of AEP (Annual Energy Production) evaluations to find an optimal layout. This is feasible with a limited number of turbines or with very simplified, and therefore faster, wake models. However, layouts obtained with simplified wake models may be suboptimal when evaluated with state-of-the-art wake models. This abstract presents a new strategy developed by GreenWITS to optimize wind farm layouts with alignment constraints that meet industrial requirements: maintaining affordable and reasonable computational time and effort while optimizing with industrial-grade wake models. The use cases considered consisting of 77 and 124 wind turbines, are inspired by the Borssele wind farm. To demonstrate how the new algorithm can be applied industrially to achieve optimally aligned wind farm layouts, several optimization studies were conducted using different wake models and optimization parameters.