Large Eddy Simulation of compressible flows: the “Charles” solver and application to jet aeroacoustics
Guillaume A. Brès – Cascade

With advancement in high-performance computing, the large-eddy simulation (LES) approach is emerging as an accurate yet cost-effective computational method for the prediction of turbulent flows and their acoustic fields. The lecture will present an overview of the ongoing efforts to improve understanding and develop predictive capabilities of propulsive jet noise through large-eddy simulations with the unstructured compressible flow solver “Charles” developed at Cascade Technologies. The emphasis will be on two important aspect of the jet noise problem. First, in the context of idealized single-stream nozzles typically used for fundamental studies, we will review recent progress on the modeling of the nozzle interior turbulent flow and its effects on the nozzle-exit boundary layer, the jet plume and ultimately the acoustic field. The large LES databases generated as part of this work are being used by different research group (including the group of Prof. Cavalieri at ITA), and flow-physics insights coming from the analysis of the databases will be presented. Second, in the context of more industrial applications and new designs of high performance turbine engine exhaust systems, we will discuss an innovative meshing approach based on Voronoi diagram under development at Cascade, to enable simulations of more complex 3-D configurations, such as non-axisymmetric and airframe-integrated nozzles.
 

Aula (Arquivo PDF)

Referências (arquivo pdf)


An overview of research conducted at the EESC-USP Fan Rig
Paulo Greco – USP

The EESC-USP Fan Rig is a long-duct low-speed experimental setup recently built at the Department of Aeronautical Engineering of the University of São Paulo. The objective is to study fan aeroacoustics with a flexible configuration that allows changes in operational conditions of the rig. Several parametric campaigns were carried out exploring the effects of fan rotational speed, fan loading and rotor-stator spacing. The fan has 16 blades and 14 vanes, with speeds up to 4250 rpm and maximum blade tip Mach number of 0.34. Acoustic measurements were taken using an array of 77 wall-mounted microphones and the data was processed to obtain the modal decomposition and power spectrum for each configuration. An in-duct rotating beamforming technique was developed for locating broadband noise sources froman aeroengine fan. A parameterization methodology for broadband and tonal noise was proposed to represent the modal noise spectra and compared with the method presented by Heidmann.

Aula (Arquivo PDF)

Time-domain representation of acoustic impedance surfaces for aero-acoustic applications
Gwénaël Gabard – Univ. Maine

The acoustic impedance of a surface is a concept that is primarily formulated in the frequency domain. However an increasing number of acoustic numerical simulations are performed in the time domain. It is then necessary to construct a numerical representation of this impedance in the time domain. A first part in this talk will present a number of best practice to construct an accurate and stable numerical implementation of time-domain impedance conditions. A second part will focus on situations when mean flow is present and a boundary layer is developing above the impedance surface. Special boundary conditions have been developed for this purpose but their implementations in the time domain require particular care.
 

Aula (Arquivo PDF)

Towards high-order FEM schemes for application to problems involving aeroacoustics
Luc Mongeau – McGill Univ.

Efforts were made to develop a finite element code for aeroacoustics problems involving complex internal/external flows such as jet flows through lobed mixers.  The initial goal was to validate prior results obtained using the Lattice Boltzmann method for such problems.  A high resolution, low-pass differential filter for Large Eddy Simulations  (LES) on unstructured grids in an Approximate Deconvolution Model (ADM) framework was first developed. The finite element discretization of the filter equation on a structured grid results in a discrete compact filter. The proposed filter was found to be significantly less dissipative than the known explicit filters, e.g. Germano's differential filter, on unstructured grids. The new filter is also capable of completely suppressing fluctuations at the grid cut-off frequency, a property lacking in alternative approaches. Manufactured solutions were used to verify the performance of the proposed filter. The results suggest that the proposed filter will be very effective for explicit filtering and ADM in LES. To obtain minimal numerical dissipation while stabilizing the finite element code, a multi-stage approach was adopted. This approach was also used to investigate similarities and differences between the explicit Taylor-Galerkin and the explicit Runge-Kutta time integration schemes. It was found that the substitution of some, but not all, of second-order temporal derivatives in a Taylor-Galerkin scheme by additional stages makes it analogous to a Runge-Kutta scheme while preserving its original dissipative property  for node-to-node oscillations. The substitution of all second-order temporal derivatives transforms Taylor-Galerkin schemes into Runge-Kutta schemes with zero attenuation at the grid cut-off. The application of this approach to an existing two-stage Taylor-Galerkin scheme yields a low-dissipation low-dispersion Taylor-Galerkin formulation. Two one-dimensional benchmarks were simulated to study the performance of this new scheme. The reverse process yields a general approach for transforming m-stage Runge-Kutta schemes into (m - 1)-stage Taylor-Galerkin schemes while preserving the same order of accuracy. The dissipation and dispersion properties for several new Taylor-Galerkin schemes were compared to those of their corresponding Runge-Kutta form.  
Using the new Taylor-Galerkin scheme with and without explicit filtering has suggested that a more fundamental approach for stabilization is required if excessive numerical dissipation at low and moderate wavenumbers is to be avoided. It is considered that the stabilization mechanism should overcome the near to grid cut-off aliasing arising from non-linear inviscid fluxes at locally convective dominant areas while the explicit filtering in ADM emulates the subgrid scale (SGS) effects. The challenge is to minimize the dissipation and dispersion at low and moderate wavenumbers so that high fidelity aeroacoustic LES can be conducted. This work will be described along with ongoing work towards high speed jet noise simulations.



Experimental methods for determination of nonlinear acoustic properties of liners (Resumo não foi fornecido)
Hans Boden - KTH
 

Aula (Arquivo PDF)

Revisiting the Cremer Impedance
Mats Abom - KTH

In a classical paper (Acustica 3, 1953), Cremer demonstrated that in a rectangular duct, with locally reacting walls, there exits an impedance (“the Cremer impedance”) that maximizes the propagational damping for the lowest mode. Later (JSV 28, 1973), Tester extended the analysis to include a plug flow and ducts of both circular and rectangular cross-section. One limitation in the work of Tester is that it simplified the analysis of the effect of flow only considering high frequencies or well cut-on modes. This approximation is reasonable for large duct applications, e.g., aeroengines, but not for many other cases of interest. Kabral et al. (Acta Acustica united with Acustica 102, 2016) removed this limitation and investigated the exact Cremer impedance including flow effects. As demonstrated in that paper the exact solution exhibits some special properties at low frequencies, e.g., a negative real part of the wall impedance. In this lecture, the exact Cremer impedance is further analyzed and discussed.

Aula (Arquivo PDF)