This paper establishes the effect of blade deflections on wind turbine noise directivity. Fast turn-around methods are used in a framework of integrated aeroelastic and aeroacoustic simulations: the blade element momentum theory is coupled with a RANS-informed Amiet's model for the aeroacoustic modelling of trailing-and leading-edge noise. This approach is applied to the NREL 5 MW wind turbine and the results of rigid and flexible blades are compared. The overall sound pressure level computed with flexible blades increases up to 13 dBA for listeners close to the rotor plane. This effect is attributed to the flapwise angular deflection of the wind turbine blade. Furthermore, the symmetry of the results with respect to the rotor plane is lost when the flapwise deflection is considered, indicating that the modelling of this rotation is of fundamental importance for the acoustic simulation.

@en

Accurate aerodynamic and aeroacoustic simulations of helicopters and wind turbines require the inclusion of the blade elasticity into the computational setup. If low-order aerodynamic models can be efficient for optimization purposes, they are often insufficient in predicting the complex flow phenomena responsible for structural vibrations and noise. For this purpose, high-fidelity Computational Fluid Dynamics solvers coupled with Multi-Body Dynamic tools can be exploited for more accurate simulations in the context of detailed design.

This thesis focuses on the development of a coupling methodology between the Lattice-Boltzmann flow solver PowerFLOW and the Multi-Body Dynamic tool Simpack for pitching and plunging airfoils featuring lumped structural parameters. This is achieved by verifying the fluid and multibody simulation setups separately against theoretical models. Then, the coupling is compared against analytical aeroelastic solutions for several amplitudes and reduced frequencies of motion. In addition, different approaches to model the airfoil motion in the fluid solver are assessed and compared favorably against each other. As a conclusive effort, the coupling is applied to a bi-dimensional airfoil flutter case returning a prediction of the flutter velocity within a 1% difference with respect to analytical methods.