Search

A detailed aeroacoustic analysis of the flow induced by the clearance between the fan tip and the shroud is performed in a scale-model fan stage of an ultrahigh bypass ratio turbofan engine, which was designed to operate at transonic regimes. A wall-modeled large eddy simulation is performed at approach condition, which corresponds to a fully subsonic operating point. The contributions of the tip-gap noise to the total fan noise are investigated using the Ffowcs Williams and Hawkings analogy. The surface is split into two parts: the tip region and the rest of the blade in order to analyze the acoustic contributions of these two regions separately. It is shown that the tip-gap region generates a significant noise component above 2 kHz, which corresponds to approximately 1.2 times the blade passing frequency. Two separate tip-leakage vortices are identified in the vicinity of the fan tip. The dominant noise sources in the tip-gap region are observed at the trailing edge of the fan blade. The wall pressure spectra in the tip-gap region and the coherence of pressure fluctuations between monitor points at different positions show an acoustic contribution of the tip-leakage flow at two different frequency ranges. The first range corresponds to medium frequencies between 2 and 9 kHz, and the second range corresponds to high frequencies between 10 and 25 kHz. The analysis of dynamic mode tracking, fluctuating pressure and velocity spectra, and instantaneous flow fields relates specific vortices in the tip-gap flow to their spectral signature and paves the way for further analytical modeling of tip-gap noise sources.

A scale model of a fan stage operating at approach condition is investigated using wall-resolved Large Eddy Simulations. The configuration is a state-of-the-art fan stage (rotor + stator) of an Ultra High Bypass Ratio turbofan engine, and the computational cost has been reduced by simulating a radial slice of a periodic sector. At low fan speeds and Reynolds number, a recirculation bubble can be observed on the suction side of the fan blades, near the leading edge. This recirculation bubble causes the boundary layer transition to turbulence and is associated with high levels of wall pressure fluctuations. The noise spectra upstream of the fan blades show high frequency tones, which are related to the noise generation mechanism of the recirculation bubble. It has been found that the size, position and pressure disturbances in the bubble depend on the mass flow rate. Furthermore, the frequencies and amplitudes of the high frequency tones in the noise spectra also depend on the mass flow rate. As the mass flow rate decreases, the size of the bubble increases, the bubble shifts towards donwstream locations, and high levels of wall pressure fluctuations can be found along the suction side of the fan blade. The acoustic signature of the bubble shifts towards lower frequencies with the mass flow rate, but its amplitude increases. A coherence analysis and a Dynamic Mode Tracking technique are used to confirm that high frequency tones can be generated by the recirculation bubble in the separation region. The flow features at these frequencies suggest that a vortex shedding mechanism is related to the noise radiated from the recirculation bubble.

International audience; In this paper the scattering of acoustic waves by a moving vortex is studied in two dimensions to bring further insight in the physical mechanisms responsible for the spectral broadening caused by a region of turbulence. When propagating through turbulence, a monochromatic sound wave will be scattered over a range of frequencies, resulting in typical spectra with broadband sidelobes on either side of the tone. This spectral broadening, also called 'haystacking', is of importance for noise radiation from jet exhausts and for acoustic measurements in open-jet wind tunnels. A semi-analytical model is formulated for a plane wave scattered by a vortex, including the influence of the convection of the vortex. This allows to perform a detailed parametric study of the properties and evolution of the scattered field. A time-domain numerical model for the linearised Euler equations is also used to consider more general sound fields, such as that radiated by a point source in a uniform flow. The spectral broadening stems from the combination of the spatial scattering of sound due to the refraction of waves propagating through the vortex, and two Doppler shifts induced by the motion of the vortex relative to the source, and of the observer relative to the vortex. The fact that the spectrum exhibits sidebands is directly explained by the directivity of the scattered field which is composed of several beams radiating from the vortex. The evolution of the acoustic spectra with the parameters considered in this paper are compared with the trends observed in previous experimental work on acoustic scattering by a jet shear layer.

The interaction of the rotor with inflow turbulence is a source of broadband noise, dominant at high wind speeds. In the vicinity of the leading edge of realistic blades, the mean flow distorts the turbulence, resulting in an attenuation of the high-frequency part of the radiated noise compared to zero-thickness blades. In this paper, to study this turbulence distortion by the mean flow, two different numerical approaches are considered. First, the linearized Euler equations are solved in the time-domain using a finite difference code to model the response of an isolated blade interacting with synthetic turbulence. Second, a vorticity approach is applied, using the Biot-Savart law combined with a vortex panel method. Using these approaches on multiple configurations, the up-wash velocity fluctuations along a streamline which goes to the stagnation point, show a decrease of the turbulent levels, from a threshold distance independent of the wavenumber. This decay is then inverted for low frequencies after a...

International audience; Turbulent-flow interactions with the outlet guide vanes are known to mainly contribute to broadband-noise emission of aeroengines at approach conditions. This paper presents a three-dimensional computational aeroacoustics hybrid method aiming at simulating the aeroacoustic response of an annular cascade impacted by a prescribed homogeneous isotropic turbulent flow. It is based on a time-domain Euler solver coupled to a synthetic turbulence model implemented in the code by means of a suited inflow boundary condition. The fluctuating pressure over the airfoil surface provided by computational aeroacoustics is used as an input to a Ffowcs Williams and Hawkings integral method to calculate the radiated sound field. Euler computations are first validated against an academic computational aeroacoustics benchmark in the case of an harmonic gust interacting with an annular flat-plate cascade. Then, simulations are applied to turbulence–cascade interactions for annular configurations, in uniform and swirling mean flows, and numerical results in terms of sound power spectra in the outlet duct are compared to semi-analytical and numerical solutions, and to an available experiment.

International audience; Turbulent flow interactions with the outlet guide vanes are known to be mainly contributing to broadband noise emission of aeroengines at approach conditions. This paper presents a 3D CAA hybrid method aiming at simulating the aeroacoustic response of an annular cascade impacted by a prescribed homogeneous isotropic turbulent flow. It is based on a time-domain Euler solver coupled to a synthetic turbulence model implemented in the code by means of a suited inflow boundary condition proposed by Tam. The fluctuating pressure over the airfoil surface provided by CAA is used as an input to a FWH integral to calculate the radiated sound field. The method is first validated against an academic CAA benchmark in the case of a harmonic gust interacting with an annular flat plate cascade. Then, simulations are applied to turbulence-cascade interactions for annular configurations, in uniform and swirling mean flows, and numerical results in terms of sound power spectra in the outlet duct are compared to semi-analytical and numerical solutions, and to an available experiment. Nomenclature B = blade number L = span M x , M θ = axial, azimuthal mean flow Mach number T u = turbulence intensity U 0 , U x = uniform, axial mean flow velocity V = vane number c = chord f = frequency k = wave number p, p ∞ = static pressure, undisturbed static pressure u' = velocity disturbance (x,r,θ) = cylindrical coordinates in the annular duct ∆k, ∆f = wave number and frequency spacing χ = stagger angle φ k x , k r () = 2-wavenumbers turbulent energy spectrum Λ = integral length scale ϕ = random phase ρ, ρ ∞ = density, undisturbed density ω = angular frequency ξ,η, r () = local coordinates attached to the vane

International audience; Turbulent wakes generated by turbofan blades and interacting with the outlet guide vanes are known to be mainly contributing to broadband noise emission of aero-engines at approach conditions. Analytical approaches, such as the well-known Amiet model can be adopted to estimate the noise generated by turbulent flows impinging thin airfoils, but they are limited by the flat-plate assumptions. The development of numerical methods allowing to consider more complex geometries and realistic flows is required. The method described in the present paper, is based on a CAA code solving the nonlinear Euler equations. The upstream turbulence is synthesized from a stochastic model and injected into the computational domain through an adapted boundary condition. It is first validated in 2D and 3D against academic flat plate configurations by comparison with Amiet solutions (exact in such cases). Then, 3D computations are applied to simulate the effect of a passive treatment (leading edge serrations) aiming at reducing turbulence interaction noise of an isolated airfoil studied in the framework of European project FLOCON. First calculations on baseline conditions are shown to be able to reproduce the measured spectra and far-field directivities, and the acoustic performances of the serrations (3-4 dB PWL reduction) are fairly well assessed too.

Turbulent wakes generated by turbofan blades and interacting with the outlet guide vanes are known to bemainly contributing to broadband noise emission of aero-engines at approach conditions. Analytical approaches,such as the well-known Amiet's model can be adopted to estimate the noise generated by turbulent flowsimpacting thin airfoils, but they are limited by the flat-plate assumptions. The development of numericalmethods allowing more complex geometries and realistic flows is required. The method, described in the presentpaper, is based on a CAA code solving the nonlinear Euler equations. The upstream turbulence is synthesizedfrom a stochastic model and injected into the computational domain through an adapted boundary condition. It isfirst validated in 2D and 3D against academic flat plate configurations by comparison with Amiet solutions(exact in such cases). Then, 3D computations are applied to simulate the effect of a passive treatment (leadingedge serrations) aiming at reducing turbulence interaction noise of an isolated airfoil studied in the framework ofEuropean project FLOCON. First calculations on baseline conditions are shown to be able to reproduce themeasured spectra and far-field directivities, and the acoustic performances of the serrations (3-4 dB PWLreduction) are fairly well assessed too.

In the context of the evaluation and reduction of fan noise in modern turbofan engines, the possibility to compute the acoustic transmission through a rotor-stator system, accounting for the rotation of the rotor, is evaluated. The objective is the development of a numerical method to simulate the propagation of acoustic waves in a global fixed region which includes a partial moving zone, with equations solved in a fixed reference frame. In order to insure a high order resolution these CAA developments and computations were performed using Onera’s CAA solver sAbrinA-V0. Firstly, a full or partial 2D moving zone was employed as a reference test case. It was insured that the moving grid acts only as a resolution support and does not affect the propagation of the waves by its rotation. Monopole and planar sources were considered. In order to connect the two (fixed and rotating) zones, spline cubic interpolation functions were employed in the ghost cells rows. Secondly, a cross-shaped rigid obstacle with zero thickness walls was embedded in the moving grid and the interaction between this obstacle and a fixed monopole was computed. Note that, despite the CAA solver can handle any non-homogeneous mean flow, this case was addressed accounting for a medium at rest. The field of application of the method is restricted to non-viscous interactions between boundary layer and acoustics, so the considered mean flows have to respect this condition. Finally, a 3D test case is presented, based on a simplified rotor with four twisted blades with zero thickness in a perfect annular duct. The acoustic interaction of this rotating blades with an helicoidal modal source injected upstream the rotor within a medium at rest was evaluated. Recent developments in sAbrinA-v0 on improved inlet/outlet acoustic boundary conditions and conclusions of gust-fixed airfoil interactions will allow a proper evaluation of CAA simulations of moving surfaces.

International audience; Passive treatments aiming at reducing turbofan broadband noise have been recently studied in the framework of European Project FLOCON. A concept based on a sinusoidal variation of the leading edge of a single airfoil aiming at reducing interaction noise has been investigated by ONERA. Turbulence-airfoil interaction mechanism is achieved using a turbulence grid located upstream of a NACA airfoil tested in ISVR anechoic open wind tunnel. High noise reductions are obtained (3-4 dB) for all studied flow speeds. Moreover, aerodynamic performances are shown to be slightly increased by the treatment that tends to reduce the drag without modify the mean loading. Experimental work is supplemented by numerical simulations using Large Eddy Simulations (LES) and direct Euler approaches to predict the acoustic response of the wing. LES is chained to a FWH (Ffowcs-Williams and Hawkings) integral to assess the radiated field. Isentropic turbulence is synthetically injected by means of a suited inflow boundary condition. Present computations are focused on the reference case (without treatment). Numerical predictions are compared to the experiment, and to analytical solutions issued from Amiet theory.

Noise generation due to the rotor wakes impinging the stator vanes is a dominant turbofan source at approach conditions, and the broadband noise component is significantly contributing to the overall level. A numerical method based on a CAA code solving the Euler equations is developped in this thesis in order to simulate the interaction noise between a turbulent wake and a vane row without geometry restrictions. The upstream turbulent flow is synthesized using a stochastic approach by considering an homogeneous isotropic turbulence spectrum model and a simplified spatial representation of the velocity field. These velocity gusts are injected in the CAA code by implementing a suited boundary condition. The present methodology is first validated against turbulence-flat plate interaction cases, by comparing the numerical predictions to Amiet model solutions. A chaining with an integral formulation is also performed to assess the acoustic far-field. Then, the method is used to estimate the acoustic response of an isolated airfoil with a wavy leading egde, designed and tested in ISVR wind tunnel in the framework of European FLOCON project. Finally, the computations are extended to ducted annular grid configurations. After a validation step on single-frequency cases described in a CAA benchmark, broadband noise simulations are performed, firstly on a flat-plate annular grid in a uniform axial flow tested in the LMFA wind tunnel, and then on a more complex configuration related to an unloaded grid in a swirling mean flow proposed by Atassi.

The scattering of harmonic sound waves by a single vortex is studied to provide someinsight into the more complex problem of sound scattering by a turbulent layer. Twodifferent methods are used to model the scattering by a steady vortex or by a vortexconvected in a uniform mean flow. The first method is based on the linearized Eulerequations in the frequency domain and is used to study the scattering by a steady vortex.The speed of this method allows to perform a study of the spatial scattering of a planewave over a wide range of frequencies, where analytical models previously developed arerestricted either to low or high frequencies. The second method uses a finite differencesolver, also based on the linearized Euler equations but in the time domain. This methodis used to study the scattering by a vortex convected in a uniform mean fow, where aspectral scattering occurs in addition to the spatial scattering. A parametric study isrealized, focusing on four parameters including the source frequency and the convection velocity. The spectra deduced from this study show a pattern similar to the haystackingobserved in existing studies of the scattering by a turbulent shear layer. Some of the trends deduced from the parametric study are also in agreement with these previous observations. Thus, the detailled study of the scattering by a single eddy is able to provide further understanding of the turbulence scattering.

Spectral broadening results from the unsteady scattering of acoustic waves propagating across a region of turbulence. This phenomenon has been observed previously for an harmonic source radiating through the shear layer of a cold jet at low speed. The resulting spectra displayed a reduced tone peak surrounded by lateral bands, with changes in these bands levels and width depending on the source and flow parameters. In this paper, spectral broadening is studied numerically for a simple configuration, consisting of a monopole radiation scattered by a turbulent layer with a constant thickness and convected by a uniform mean velocity. The numerical method relies on a finite difference code solving the linearized Euler equations in the time domain. The turbulent layer is synthesized with a stochastic method based on the filtering of white noise to impose prescribed statistical properties to the turbulence. The turbulent fluctuations are then added to the steady mean flow to form an unsteady base flow around which the Euler equations are linearized. This introduces terms involving products between the turbulent and acoustic fluctuations which are responsible for the scattering. Such a numerical methodology allows to vary the properties of the turbulence generated for the chosen configuration such as the turbulent kinetic energy and the integral length scale. In this paper, the effects of several parameters such as the convection velocity and source frequency on the spectral broadening are studied. The trends deduced from the results can be compared to previous models and experimental data for a jet shear layer, and they can also be related to the trends observed in a previous study of the scattering of sound by a single convected vortex.