Abstract EN:

The model problem of an airfoil embedded in the fluctuating wake of a rod is investigated in the first part. Such a flow configuration is an effective benchmark for developing and validating numerical methodologies of aeroacoustic prediction. Furthermore, the rod-airfoil configuration is of great academic concern, for it allows to investigate some underlying mechanisms involved in the generation of vortex-airfoil interaction noise. The numerical and theoretical relevance of the rod-airfoil configuration reflects the structure of the first part. An analytical model based on the circulation theory is initially developed. This is used to investigate the influence of the airfoil geometry and the vortex size on the far pressure field, for a given distribution of vortices convected past an airfoil. Particular emphasis is given to the nonlinear effects related to the airfoil thickness and camber, the interaction of the airfoil wake with the incident vortices, and the vortex distortion near the airfoil leading edge. The limits of applicability of the frozen convection hypothesis and other linear approximations are discussed in great detail. Wall pressure and acoustic measurements are carried out with a twofold aim in mind: (i) to obtain data for comparisons with numerical results, (ii) to investigate the three-dimensional character of a nominal two-dimensional flow. Numerical results are obtained by means of a hybrid RANS/Ffowcs-Williams & Hawkings approach. The RANS solver is a finite volume code developed at LMFA, and the FW-H solver is the rotor-noise code Advantia. The latter has been developed in the context of the present study and is based on the so-called advanced time approach, firstly proposed by the author. The intrinsic three-dimensional behaviour of the flow past a bluff body is described for the first time by means of a spanwise statistical model. This allows to perform acoustic analogy predictions by using a two-dimensional aerodynamic field, but accounting, to some extent, for the three-dimensional character of the flow. The hybrid RANS/FW-H approach and the spanwise statistical model are applied to the rod-airfoil system. It is shown that, despite the tonal character of the RANS solution, the spectral broadening around the tonal frequency, as observed in the experiments, is partially recovered. The second part of the work illustrates theories and models in fluid-body aeroacoustics. It is only concerned with sound generation in the absence of acoustic feed-back. Therefore, only external flows are considered, the propagation problem in complex geometries is not addressed, and the aeroacoustic feed-back in resonant configurations is not examined. Unsteady aerodynamic theories are presented in the second part as an ineluctable theoretical basis for devel¬oping aeroacoustic models. Nevertheless, the author's feeling in writing this part was that "aeroacoustic sound is an aerodynamic by-product, bid aeroacoustics are not a by-product of aerodynamics".