DLR Portal
Home|Textversion|Imprint|Sitemap|Contact |Deutsch
You are here: Home:Departments:Numerical Methods:Application
Advanced Search
News
Institute
Departments
Combustor
Combustion Test
Fan and Compressor
Numerical Methods
Research fields
Application
Publications
Simulationsystem
Projects
Engine
Engine Acoustics
Engine Measurement
Turbine
IT Infrastruktur
Facilities
Publications
Offers
Service & Links
Application
Send article to a friendPrint

Aeroacoustics



 LES of a turbulent flow in a low-speed axial fan: iso-surfaces of lambda-2
zum Bild LES of a turbulent flow in a low-speed axial fan: iso-surfaces of lambda-2
Fan noise itself comprises both tonal and broadband components and arises due to a variety of mechanisms:
  • rotating shock structures
  • the interaction of inlet flow disturbances (ingested atmospheric turbulence, asymmetric inflow, boundary-layer turbulence) with the rotor blades
  • rotor/stator interaction
  • secondary flow phenomena such as that in the tip clearance gap
For subsonic blade tip speeds the radiated acoustic field is typically dominated by discrete tones at the blade passing frequency (BPF) and harmonics thereof. In such cases the interaction of the rotor-wakes with the downstream stators and the interaction of the rotor itself with inflow disturbances are the dominant noise source mechanisms.

The computation of noise generation and propagation from first principles in a realistic three-dimensional axial fan configuration is a computationally intensive endeavor. As in all computational aeroacoustic (CAA) applications particular care must be taken to ensure adequate spatial and temporal resolution of the small amplitude acoustic waves. The actual wave propagation characteristics, and mesh and time step size requirements of the TRACE code have been established in earlier studies.

 

Boundary Conditions

An equally important aspect of any CAA simulation are the boundary conditions, and in particular the radiation boundary conditions required along the external boundaries of the computational domain. Such boundary conditions should ideally be transparent to outgoing vortical, acoustic and entropy waves, while permitting incoming disturbances and external physical boundary conditions to be prescribed. Accurate non-reflecting boundary conditions are implemented in the TRACE code. In particular, an implementation was sought and developed specifically for use with the implicit dual-time stepping solver algorithm of TRACE. [Ashcroft]

 
 Validation of implicit non-reflecting boundary conditions: Upstream wave propagation (M = 0.4, Eta = 1.28, p0 = 89 Pa) Instantaneous pressure perturbation contours
zum Bild Validation of implicit non-reflecting boundary conditions: Upstream wave propagation (M = 0.4, Eta = 1.28, p0 = 89 Pa) Instantaneous pressure perturbation contours

 
 

Detached Eddy Simulation (DES)

The flow field interaction effects are shown to play a decisive role in the noise generation process. To predict the far-field noise spectra of jet engines a special numerical simulation is necessary. The DES scheme implemented in TRACE has been calibrated and applied to two different test cases that are relevant to turbomachinery flows. Either of them is the rod-airfoil testcase, which aimed at the simulation of broadband noise in the turbomachinery. In order to assess the accuracy of the employed DES model for broadband-noise prediction, simulations of the rod-airfoil test case investigated by Jacob et al. were carried out. To accurately predict the noise radiated by the highly-periodic vortex structures superimposed with more stochastic small-scale vortical fluctuations, a numerical model must be capable of resolving directly at least the larger turbulent structures in the wake of the cylinder and their interaction with the downstream airfoil.

 
   
 
 Schematic of rod-airfoil configuration superimposed with an instantaneous snapshot of the computed spanwise vorticity
zum Bild Schematic of rod-airfoil configuration superimposed with an instantaneous snapshot of the computed spanwise vorticity
 
 

 
As can be seen the far-field values are generally well predicted and, similar to the near-field data.  
 
 computed acoustic near-field
zum Bild computed acoustic near-field

 

 

 comparison of computed and measured acoustic far-fields (Dashed line - experiment, symbols - computation)
zum Bild comparison of computed and measured acoustic far-fields (Dashed line - experiment, symbols - computation)

 

 
   
 

LinearTRACE

The application of time-linearized RANS methods to the numerical simulation of aerodynamic noise generation and propagation in a modern high-bypass ratio fan is investigated. The convective and viscous fluxes of an existing URANS solver are linearized and the resulting unsteady linear equations are solved in the frequency domain, thereby transforming the problem from unsteady time-integration to a simpler complex linear system. The linear system is solved using a parallel, preconditioned restarted GMRES method. In view of the fact that nonlinear unsteady CFD computations remain costly in terms of CPU time, the results show that using the time-linearized solver for noise prediction constitutes an adequate method to efficiently obtain estimates of noise levels.

 
 nonlinear solver
zum Bild nonlinear solver
 time-linearized solver
zum Bild time-linearized solver

UHBR-Fan, Pressure contour of the first harmonic of the sixth mode: left - nonlinear solver, right - time-linearized solver

 

 

 
 

Coupling to acoustic far-field noise prediction

To evaluate the acoustic characteristics of an engine, typically a hybrid approach is used. Special optimized metheds to predict the wave propation into the far-field are available. The methods includes Boundary Element Methodens (BEM) and integral acoustics methods after Ffowcs Williams & Hawkings or Kirchhoff. TRACE provides suitable surface for all of these applications. Those are as well hard walls or in the flow field located porous surfaces. Either the definition will be made by the user befor a TRACE simulation to reduce the file output or afterwards from the complete 3D flow solution.
The definition of a porous surface by the user is made by iso-contour lines and are created with the help of VTK  (Visualization Toolkit) by Kitware. The resulting surfaces are unstructured grids, which can also provided with a band structure e.g. to make modal decomposition. The resolution is orientated on the original 3D grid and as well splitted into block to achive an efficient parallelization.

 
Counter-rotating open rotor with a enclosing Surface for acoustic far-field noise prediction
.
 MTU-mixer with various porous surfaces for acoustic far-field noise prediction
zum Bild MTU-mixer with various porous surfaces for acoustic far-field noise prediction

 

References

  • Ashcroft, G. B., "Numerical Modelling of Wake-Jet Interaction with Application to Active Noise Control in Turbomachinery", AIAA Paper, 2004

 


Contact
Dr. Graham Ashcroft
German Aerospace Center

Institute of Propulsion Technology
, Numerical Methods
Köln

Tel.: +49 2203 601-2283

Jens Wellner
German Aerospace Center

Institute of Propulsion Technology
, Numerical Methods
Köln

Tel.: +49 2203 6013245

Fax: +49 2203 64395

Application
DLR-Institute of Propulsion Technology-Engine Acoustics Department
DLR-Institut für Lufttransportsysteme
Publications (elib)
Related Topics
Fluid Mechanics and Thermodynamics
Copyright © 2013 German Aerospace Center (DLR). All rights reserved.