Dandy Eschricht

Numerical Simulation of Jet Mixing Noise Associated with Engine Exhausts

In this article the work conducted at ISTA and LMFA during the French/German research project Noise Generation in Turbulent Flows is presented. The hybrid approach applied for the noise prediction at ISTA consists of a flow simulation which is coupled to the acoustic analogy of Ffowcs-Williams& Hawkings including predictions based on volume sources. The turbulence modelling approach used in the flow simulations is the Detached-Eddy Simulation as well as the Large- Eddy Simulation. Along with details on the numerical method and the turbulence modelling approach, results from compressible, three-dimensional flow simulations and their acoustic predictions are presented for the rod-airfoil configuration which acts as a simple broadband noise emitter. Additionally, results for a single-stream jet and for a coplanar, coaxial, unheated jet flow are presented. The results of the Detached-Eddy Simulation of the latter case are compared with those of a Large- Eddy Simulation of the same configuration.

Prediction of Jet Noise from a Coplanar Nozzle

The unsteady flow field of a compressible, unheated, coplanar jet has been simulated using Detached-Eddy Simulation (DES) at a Reynolds number of about 1 million and using a grid of about 8 million cells. To make a direct comparison of turbulence modelling aspects, the same configuration has been simulated using Large-Eddy Simulation, applying the same numerical conditions (i.e flow solver, grid, local grid length scale definition...) as used for the DES. This will allow a direct comparison of the two modelling approaches. The results of both flow simulations have been extended to the acoustic far-field by means of the acoustic analogy of Ffowcs Williams&Hawkings. The numerical results obtained for the flow and the acoustic field will be compared to experimental data that have been obtained by LEA Poitiers for the same configuration that is numerically studied here. These measurement included flow field PIV and LDA data, near field pressure and acoustic far-field measurements and will be used for a detailed comparison with the numerical results.

Jet Noise Prediction of a Serrated Nozzle

The flow fields of two configurations of high Reynolds number, heated, co-axial jets have been simulated using Detached-Eddy Simulation (DES). In this work two longcowl configurations are considered that dier in the design of the outer nozzle lip. The serrated configuration has shown to reduce the far-field noise by about 1-2 dB in all directions in experiments when compared to the smooth lipped configuration. Additional to the computations of the two configurations that follow the approach of Yan et al, a DES with a modified length scale definition has been carried out for the baseline configuration to study the eect of the turbulence modelling approach. The modification showed to reduce the modelled turbulent viscosity in the simulation. To examine the quality of the simulations, comparisons are made with PIV results for the baseline configuration. The flow-field results have been extended to the far field by means of the acoustic analogy of Ffowcs-Williams&Hawkings. The far-field predictions of the combined DES/FWH approach are compared to measurements. Even as limitations in the flow-field prediction are present, the predicted overall sound pressure levels in the far field are within a 3 dB range to the measurements.

Turbulence Modelling Effects on Tandem Cylinder Interaction Flow and Analysis of Installation Effects on Broadband Noise Using Chimera Technique

The NASA tandem cylinder benchmark case (L/D = 3.7) is studied numerically using novel variants of DES the Delayed Detached Eddy Simulation (DDES) and the Improved Delayed Deteched Eddy Simulation (IDDES). The ow Mach number is 0.1285 and the Reynolds number is set to 166.000 to match th corresponding experiments at the NASA Langley Research Center (Jenkins et al., Lockard et al.). Incompressible simulations are carried out on a mandatory grid from the EC ATAAC project with approx. 9.5 million cells with a spanwise extent of 3.0D and ow normal domain extent of approx. 17.8D with symmetry boundary condition to mimic the closed experimental test section. In a second step a compressible IDDES simulation is done to evaluate the broadband noise of the con guration. Measurements are available from the QFF open jet facility. To capture the installation e ects a combined grid of above mentioned cylinder core grid and a single stream jet grid is used. The communication between the grids is done by an overset chimera technique. The noise in the far eld is calculated by a standard FW-H method. The IDDES approach is designed to extend the LES region of the original DES approach (hybrid RANS/LES) from Spalart et. al. (1997) to the turbulent boundary layer, as proposed rst by Travin et. al. in 2006. The non-zonal blending occures therefore inside the boundary layer the RANS model acts as a wall model for the LES. The comparison of the simulations show the improvement of the results by using a IDDES.

Advanced DES Methods and Their Application to Aeroacoustics

Detached eddy simulation (DES) is shown to be a suitable method for the simulation of the sound generation of turbulent flows, because it provides access to the resolved turbulent scales at minimal computational cost. The near-wall region is solved efficiently by RANS while LES is applied to all regions containing scales important for the noise generation. In addition to the usual LES resolution requirements, for acoustics the smallest of these scales are defined by the highest frequencies of interest. The sound radiation is generally computed by solving an integral over a data surface surrounding the source region outside the turbulent flow. The grid must be fine enough to resolve the sources and the propagation between the sources and the data surface. Examples for the simulation of noise emission problems are presented for the influence of nozzle serrations on jet mixing noise and three airframe noise sources, which are rod-airfoil interaction noise, the noise of high-lift devices, and airfoil self noise.

Analysis of noise-controlled shear-layers

Despite very similar flow-structures, the uncontrolled and controlled shear-layers of Wei & Freund 1 were found to generate radically dierent sound fields (a dierence of up to 11dB). In this work we analyse these flows—with a view to better understanding what led to this sound reduction—using two approaches: (1) we analyse the in-flow pressure fields in the x ! domain, where we find that the acoustically-matched components of the controlled flow have been reduced substantially, while the global structure remains very similar; and, (2) we use causality-correlations to study the kernel of the Lighthill solution: the spatiotemporal source-correlation fields are found to have been dramatically degenerated by virtue of very subtle changes in the controlled flow’s evolution.

Numerical Investigation of the Noise Emission from Serrated Nozzles in Coaxial Jets

Jet noise remains an important source of aircraft noise, especially at take-off. The reduction of jet noise due to the application of serrated exit nozzles in aero-engines is one of the subjects under investigation within the German research project FREQUENZ. It can be shown that the generation of additional vorticity at the nozzle exit by modifications of the exit geometry, such as serrations, influences the radiated noise spectrum giving rise to a lower acoustic emission in certain frequency ranges. However, the mechanisms behind this phenomenon are not yet well understood. Numerical simulation of the flow field allows an in-depth analysis of the flow phenomena involved and the mechanisms of noise generation. In this work, simulation results for high subsonic coaxial jet flows are presented for a typical jet-engine exhaust. The geometry has been studied with a plane nozzle and with passive flow control (serrated) on the nozzle lip. They are investigated using far-field sound characteristics obtained by the Ffowcs-Williams-Hawkings acoustic analogy and compared to experimental measurements.

Numerical Simulation of the Flow and Sound Field of a Serrated Nozzle

Nowadays, turbine engines for commercial aircraft use high-bypass-ratio, dual-stream jets to limit the acoustic radiation. Even though a large noise reduction has been achieved since the middle of last century that depends heavily on the increased bypass ratio, additional reduction is sought by modifications to the initial turbulence field behind the nozzle exit. Such modification can be achieved for instance by micro jets, allowing active control, or by geometrical changes to the nozzle lip, i.e serrations. The present paper is concerned with the latter approach. A realistic, long-cowl nozzle configuration with a modified outer nozzle lip has been studied numerically using Detached-Eddy Simulation. The dual-stream configuration is of a high Reynolds and Mach number. For this configuration a noisereducing effect of the serrated nozzle lip was proven experimentally. The work presented here aims at reproducing the noise reduction numerically, and by analysing the numerical flow and noise field, to provide some insight into the physical mechanisms that lead to this noise reduction. The flow field of the configuration with a clean nozzle lip was simulated using the same methodology and a grid of equal resolution for this purpose. This provides the data used for comparison of the two configurations.

Modeling and Simulation of Coaxial Jet Flow

This paper presents various hybrid approaches to simulate the turbulent coaxial jet flow and its fluctuating quantities, a generic configuration relevant to engineering applications. The Detached-Eddy Simulation (DES) is applied to treat the turbulence. The influence of the grid filter width used in the DES length scale has been preliminarily studied. It is found that the maximum formulation cannot resolve jet flow on the mesh used, while the cubic root formulation is able to capture the turbulent structures. Thereafter, two different DES formulations are used in combination with the cubic root formulation. In the first one, the DES length scale is substituted only in the turbulent kinetic energy equation, while in the second one additionally in the turbulence eddy viscosity expression. With respect to turbulent structure and spectral analysis, the second formulation has better resolution than the first one.