École Nationale Supérieure de
Techniques Avancées

Résumé

In order to mitigate global warming, green energies incentives have been enacted to increase the part of renewable energies into the energy mix. Tidal turbines are amongst the green energies which are being studied the most lately. In order to maximise the cost-efficiency of these turbines, fatigue loading and tidal farm power have to be optimised. A first approach to solve this problem is to take a look at how the wake develops, what are its characteristics and how the wakes mix with each other. The first two sides of this problem will be tackled in this study. The opensource Computational Fluid Dynamics (CFD) software OpenFOAM has been used, coupled with the Actuator Line Model (ALM) extension SOWFA created by NREL. ALM mimics the behaviour of wind turbine blades by introducing body forces in the flow which are computed through a Blade Element Momentum (BEM) model. A new solver has been created, allowing to use ALM into two-phases simulations as well as improving the way body forces are introduced in the flow. In a first part, the solver has been tested against experimental tidal turbine experiments carried out by IFREMER in a water tank. The blades have been modelled through airfoil interpolations and their polars have been obtained using XFOIL. Secondly, three Large Eddy Scale (LES) simulations have been run using the two-bladed NREL turbine. Using the transient results, downstream velocities, root and tip vortices have been identified in the wake. The Proper Orthogonal Decomposition (POD) has been applied to a series of transient snapshots which have been sampled during each of the simulations. These enabled to extract the coherent structures of the downstream flow by finding the eigenmode basis which better fits the wake dynamics.