Christoph Richter

Comparison of Time-Domain Impedance Boundary Conditions for Lined Duct Flows

impedance of liners. This part of the model, however, is limited to single frequency sound source in its present form. The EHR is based on the z‐transformation of the corresponding impedance function for the Extended Helmholtz Resonator in the frequency‐domain. The implementations of the two models in the numerical simulation tool with the DRP scheme are discussed in details. The results from the both models in general agree notably well with each other. However, in one case, the results from the two models dier significantly due to the presence of strongly attenuated modes, which renders the approximations of the EFI invalid. The validation and verification of both models are carried out for the latest NASA impedance tube experiment and the generic aero‐engine frequency domain results of Rienstra and Eversman. In addition, an impedance eduction technique based on the EHR and the optimization process has been developed. This further extends applications of the time‐domain models. The simultaneous optimization of all frequencies considered in the NASA’s impedance tube experiments results in a physically reasonable set of parameters for the EHR. The fitted terminal impedances and the measured mean flow profiles are used for the impedance eduction. The numerical results of both models for the generic aero‐engine inlet test case with higher azimuthal modes also compare well if the same flow assumptions and eigenvalues are used as for the frequency domain methods. Overall both models give a similar physical behavior, but the EHR requires smaller time steps in the case of small face sheet reactances.

The Stability of Time Explicit Impedance Models

from the ones presented ten years ago. This is surprising in the background of the fact, that with the development of the first time domain impedance boundary conditions the authors were facing the problem of a hydrodynamic instability. Such an instability usually would disable the resulting methods for an application in an industrial environment. Current studies do report the problem, too. It seems, that the instability usually shows a higher wave number than the acoustic solution. Therefore, it could be suppressed by an adequate mesh size. However, the obtained solution would be strongly mesh dependent. Therefore, all approaches presented recently address the hydrodynamic instability by damping, filtering or smoothing of the terms resulting from the Myers boundary condition. Additional dissipation is common for all reported stable approaches. The present paper investigates the impact of additional dissipation on the predicted attenuation by the liner by using the acoustic intensity. The Myers boundary condition, which models a grazing flow with infinitely thin shear layer at the lined surface, is identified as the cause of the hydrodynamic instability. Additional dissipation is only required and added for the convective terms of the Myers boundary condition, therefore. Consequently, the results with additional dissipation do not significantly dier from the ones without, expect for the lack of the instability in most of the cases. The current paper discusses the possibility to use a poorly resolved shear layer to remove the instability, as second stabilization approach. The method leads to a strong dependency of the observed attenuation for a generic intake test configuration to the shear layer thickness. The results obtained for a partially lined duct with the above stabilization methods are encouraging. However, for unusual impedances like a purely reacting liner, it is dicult to suppress the instability without changing the acoustic solution. Additional low order filtering always removes the instability, but the leading and trailing edge eects dominate the acoustic solution then. A grazing flow with resolved shear layer always adds dissipation, when applied at a hard wall. With liner it excites an instability for downstream propagating acoustic waves. This is in agreement to experimental observations of an hydrodynamic instability given in the literature for an approximately purely reacting liner. Altogether the results indicate, that the resolution of a grazing flow with physical shear layer is recommended to account for the diraction eects and prove the possible presence of an instability. The alternative is to apply low order filtering to stabilize the Myers boundary condition, which does not require small mesh spacings. The first approach would increase the accuracy of the numerical simulation while the second approach is recommended for fast simulation results.

A review of time-domain impedance modelling and applications☆

Time-domain impedance boundary conditions are reviewed and similarities between the models are shown. The extended Helmholtz resonator model is used with a time-domain impedance boundary condition for impedance eduction and simulation of the noise radiation from a lined aeroengine nozzle. Model parameters obtained by impedance eduction from flow duct measurements and then used to predict the noise radiation for realistic aeroengine nozzle flow and geometry. The prediction of sound radiation from a coaxial jet nozzle at approach conditions with the educed model parameters is performed in 2D and 3D. Various features of the numerical method that enable it to be used for realistic applications are presented. This includes the suppression of flow instabilities in the linear inviscid solution and analysis of the solution to prove the global conservation of acoustic energy in each individual result. It is found that the current approach is too time consuming for optimisation, even for a 2D simulation using parallel processing. Thus, finally the potential of porting a CAA method to the graphics processing unit (GPU) is shown in one example. The GPU based computation is about 100 times faster. The results are encouraging, even though the GPU variant needs further validation.

Indirect Combustion Noise Generation in Gas Turbines

The paper investigates the indirect combustion noise, which is generated during the acceleration of the convected entropy nonuniformities of the combustion products in the outlet nozzle of the combustion chamber. The generation mechanism of the indirect noise is proven experimentally and through numerical simulation. Probe microphones and fast thermocouple probes were used to measure pressure and temperature uctuations. The generation of indirect noise is veried via the phase relationship between thermocouple and microphone signals. The o w eld in the combustion chamber is simulated by means of an unsteady RANS computation. Self excited oscillations are used for the computation of the direct and indirect noise generation of the combustion chamber. Since the related frequencies are low and the corresponding scales much larger than the turbulent scales, a CAA-method is employed for both the propagation of sound waves as well as entropy perturbations. It is shown that the CAA method is capable to describe the acoustical properties of the combustion system found in the experiments when the URANS simulation is used as input. The experimental results also show that indirect combustion noise may contain high frequency noise contributions, which are generally attributed to turbine noise.

Investigation of Sound Radiation from a Scarfed Intake by CAA-FWH Simulations using Overset Grids

A hybrid numerical approach for the simulation of the sound propagation and radiation from aero‐engine intakes considering non uniform mean flows is presented. The approach consists of a 3D finite dierences CAA method in time domain for the simulation of the propagation in the intake as well as a Ffowcs Williams & Hawkings integration method for the far‐field prediction. Inside the CAA solver an overset grid algorithm is implemented oering a quite simple handling of configurations with complex geometries. This hybrid CAA‐FWH method is applied for the investigation of the sound radiation from engine‐ intakes with scarfed nacelles considering mean flow eects. Therefore a parameter study investigating the dependency of the scarfing angle, the mean flow velocity and the mean flow angle of attack is carried out. Based on the obtained far‐field directivities the influence of the parameters is discussed.

Validation of a Model for Open Rotor Noise Predictions and Calculation of Shielding Effects using a Fast BEM

For application in BEM/FMM shielding calculations a simple analytical model for the loading noise of contra rotating open rotors (CRORs) with different rotational speeds is derived. Following previous work of S.L.A. Glegg, the model is formulated in frequency domain and the pressure is approximated by a set of dipoles on circles on the propeller disk. The dipole strength depends on the blade loading function and can be obtained, e.g., by CFD calculations. For arbitrary rotational speeds, the blade loading function is not a perodic function on the propeller disk anymore and must be approximated, e.g., by a least-squares Fourier approximation. The CROR model is checked against a time domain solution using rotating dipoles and validated with data from a test of Rolls-Royce’s open rotor model rig 145 in DNW. The lowest three and most dominant peaks in the spectrum of the uninstalled rotor are well predicted by the method with errors less than 3dB with a correct directivity. For higher frequencies larger errors are observed. The CROR model has been developed for shielding calculations with the DLR BEM/FMM code. Some information about the code is given and the applicability of the model is demonstrated for a CROR installed at a modified DLR F6 aircraft geometry.

Simulation of the Rearward Propagation of Fan Noise through a Long Cowl Aero-engine

Jet aero-engines are considered as the major noise source of modern aircraft today, especially during take-off. In the past, aeroacoustic research activities in this field were focused primarily on jet noise and were later extended to the aero-engine intake due to the growing size of the fan. The intake with its exposed fan is in fact the main contributor to tonal noise received in the far-field. Significant improvements have been achieved using e.g. liners and optimization of the rotor-stator geometry itself. Until recently, less attention has been paid to the downstream propagation through the bypass duct with its internal installations. To improve the overall noise characteristics of aero-engines it is necessary to further exploit the potential of noise reduction and extend the research to the less prominent but nonetheless important bypass duct. In this work a successive three-dimensional CAA-approach to simulate the noise propagation through a realistic aero-engine exhaust under consideration of the mean flow will be presented. The effect of installation parts such as the engine mount and the splitter plate that divide the bypass duct into two “c-channels” will be investigated using a geometric approximation. Finally the effect of acoustic lining of different exhaust parts will be studied. To cope with the complex geometry the overset grid technique is employed as implemented in the TUBA3D code. This extention of the classic finite differences method enables an increase in the simulation accuracy whilst maintaining efficiency.

Acoustic Investigation of a Specially Manufactured Non-Locally Reacting Liner for Aircraft Application

This work examines the influences of cell interaction on the acoustic damping performance of a recently developed liner structure (similar to Helmholtz resonators). Due to the patented production process slots remain in the cell walls and thereby the liner is non-locally reacting. For comparison another test object almost without slots is analyzed likewise. The related acoustic properties in terms of energy dissipation calculated from microphone measurements are discussed. Improvements in measurement accuracy and data acquisition lead to very precise results for this study. The focus is then put on experiments to further understand the effects of inter-cellular communication. Besides, a numerical approach is utilized to educe the model parameters of a time domain impedance model. It is shown that the techniques developed for locally reacting liners provide important information to characterize this special type of liner. Moreover, the definition of the Level of Cell Interaction (LOCI) and Effective Communication Range (ECR) is introduced to quantify effects due to cell communication of non-locally reacting liners.

Validation of a Zonal Approach Computing the Sound Radiation from Lined Ducts

A zonal approach for the prediction of the far-field radiation from flows in lined ducts is investigated. The approach combines a high-order computational aeroacoustics scheme with the recently developed acoustic intensitybased method for the calculation of far-field radiations. The advantage over the application of an acoustic analogy is the usage of an open control surface for the acoustic input. The capability of the current hybrid approach was validated on the basis of two benchmark cases. One considers the sound radiation from a two-dimensional semiinfinite duct in which a small fraction of the walls close to outlet nozzle is lined. The second problem concerns the sound radiation from a bypass enginelike annular duct with lined and hardwalled configurations of the infinite centerbody and sheared flow conditions. A broadband time-domain impedance boundary condition based on the extended Helmholtz resonator model and the Ingard–Myers boundary condition is implemented in the computational aeroacoustic code to model the acoustic linings. Although mesh refinement may improve the solutions, the good agreements of analytical solutions and numerical results, in particular for the radiation characteristics, verify this approach as an accurate and efficient prediction tool.

Mode propagation in bifurcated bypass ducts: Application oriented simulation approach

An application oriented numerical approach is presented which allows the simulation of the sound propagation in aero-engine bypass ducts with acceptable numerical e ort. Inside the structured working high order CAA solver an overset grid algorithm is implemented o ering an e cent and quite simple handling of con gurations with complex geometries. As an example the tonal noise propagation in a generic bypass duct with various installation parts is investigated. Although the studied con gurations are simpli ed, they reproduce all key features and challenges of real engine simulations. The performed simulations and analyses demonstrate on the one hand the capability of the CAA solver and the the presented analysis methods for an e cient prediction of the sound propagation of real aero-engines. On the other hand the in uence of di erent kind of installation parts inside the bypass duct on the acoustic mode propagation is analyzed and discussed.