EOS

The objective of the EOS project (in French: Simulation Numérique d'Eoliennes OffShore) is to develop accurate numerical tools for the simulation of floating offshore wind turbines. ICI is working with LHEEA (Research Laboratory in Hydrodynamics, Energetics and Atmospheric Environment) on this project, funded by the West Atlantic Marine Energy Community (WEAMEC) and with industrial supported from Hydrocean. The project started in 2016 and has naturally been integrated in the strategy of WEAMEC, which intends to create a welcoming environment for the development of marine renewable energies in the Pays de la Loire Region.

Context

• The rotors installed on offshore wind turbines are massive, but their design and their rotation speed generate flows with Reynolds numbers of several million. These flows generate very small and vanishing vortices, which are particularly difficult to handle numerically. Moreover, strong aeroelasticity effects are observed on the blades. In the onshore wind community, it is common not to represent the rotor, but to use models of various accuracy to simulate its influence.

• The floating behavior of the turbines generate different problems. First of all, the simulation of statically moored bodies is a numerical challenge which is not trivial. Solving the movements of the bodies is complex, even more with waves breaking on the structure. While different techniques can allow the generation of wave fields at a moderate computational cost, their precision in extreme events is limited. On the contrary, expensive computational methods involving thin meshing are naturally more efficient for those events than they are for wave fields.
• The floating wind complexity occurs as the movements of the structure are transmitted to the rotor. The orientation of the relative wind observed at the rotor is constantly moving, which deteriorates the performances of the rotor models previously discussed.

Methodology used in EOS

• A single mesh is used in the computation, and the reconstruction of each phase is done using level-set function.
• Stabilized finite-elements are used for the resolution, with a unstructured tetrahedral mesh.
• An exact representation of all geometries is done, including the wind turbine rotor. The full reconstruction of the turbines impose the capture of the boundary layers, leading to high computational costs, but also increased precision.

• Anisotropic meshing allows the use of few points to represent the geometries, reducing the computational cost and increasing precision. The mesh is automatically generated.
• The resolution of the Navier-Stokes equations is done using the monolithic approach and a Variational-MultiScale paradigm. The coupling with the mesh adaptation procedure guarantees an appropriated concentration of points in the interest areas.
• The code is massively parallelized, allowing simulations to be deployed on supercomputers. Large computational loads can be handled through the use of a high number of processors.

Further developments can be found in the publication linked in the “More information” section.