TUTORIALS

FINE™/Acoustics 8.1

[Link]
 CONTENTS
TUTORIALS

TUTORIAL 1. NACELLE (2D)

1.1 Introduction 5
1.1.1 Prerequisites 5
1.1.2 Problem description 5
1.2 Aerodynamic Analysis 7
1.3 Acoustic Analysis 9
1.4 Post-processing 36
1.5 Acknowledgment 43
1.6 References 43

TUTORIAL 2. NACELLE (3D) - WIZARD APPROACH - TONAL NOISE ANALYSIS

2.1 Introduction 45
2.1.1 Prerequisites 45
2.1.2 Problem description 45
2.2 Aerodynamic Analysis 46
2.3 Acoustic Analysis 50
2.4 Acknowledgment 69
2.5 References 69

TUTORIAL 3. CONTRA ROTATING OPEN ROTOR

3.1 Introduction 71
3.1.1 Prerequisites 71
3.1.2 Problem description 72
3.2 Aerodynamic analysis 74
3.3 Acoustic analysis 78
3.3.1 Solid / Rotating FW-H surface 80
3.3.2 Permeable/ Fixed FW-H surface 96
3.4 Comparison of acoustic results 112

TUTORIAL 4. SIROCCO-TYPE FAN - BROADBAND NOISE ANALYSIS

TUTORIAL 5. NOZZLE
5.1 Introduction 142
5.1.1 Prerequisites 142
5.1.2 Problem description 142
5.2 CFD - RANS simulation 144
5.3 Flow-Noise analysis 147
5.4 Post-processing 155
5.5 Acknowledgment 160
5.6 References 160

2 FINE™/Acoustics 8.1 Tutorials

TUTORIAL 6. AIRFOIL WITH SLAT
6.1 Introduction 162
6.1.1 Prerequisites 162
6.1.2 Problem description 162
6.2 CFD - RANS simulation 163
6.3 Acoustics analysis 164
6.3.1 Flow-Noise analysis 164
6.3.2 BEM approach (uniform mean flow) 180
A. BEM setup 180
B. BEM Post-processing 185
6.3.3 FEM approach (non-uniform mean flow) 190
A. FEM setup 190
B. FEM Post-processing 200
6.4 Acknowledgment 202
6.5 References 203

TUTORIAL 7. MUFFLER
7.1 Introduction 204
7.1.1 Prerequisites 204
7.1.2 Problem description 204
7.1.3 Preparation 206
7.2 Simulate sound absorbing material 207
7.3 Predict transmission loss using a Wizard 207
7.4 Manually setup Muffler FEM analysis 214
7.5 References 234

3 FINE™/Acoustics 8.1 Tutorials

TUTORIALS
The tutorials for FINE™/Acoustics as listed below provide detailed step-by-step instructions for the
entire simulation process. Most of the tutorials are designed so as to limit use of computational
resources but some of them may require a significant amount of computational time assuming they are
executed on the finest mesh level.

Where to Find the Files Used in the Tutorials?

Each of the tutorials starts from mesh generation and uses an existing geometry. The appropriate
files (and any other relevant files used in the tutorial) can be downloaded from the Products page
of the Customer Area ([Link]).

Conventions Used in the Tutorials

Several conventions are used in the tutorials to facilitate your learning process.
Following a short introduction, each tutorial is divided into 3 sections respectively related to mesh
generation, settings and calculation, and post-treatment.
Inputs required to execute the tutorials are most often restricted to the geometry, either in a
".geomTurbo" or CAD related format, and profiles of boundary conditions.
The sequence of actions to be executed are described through a step-by-step approach, in the form
of Arabic numbers.
Additional insight about some specific actions and/ or features is frequently added to illustrate the
tutorial further. This information is proposed for the purpose of clarity and completeness, and
should not be executed. It appears as a note with a light blue background.
The calculations executed to illustrate the behavior of the flow solver include a "reasonable"
number of grid points, so as to limit the computational resources required to complete the tutorials.
As a consequence, most of the underlying flow solutions must be considered as qualitative.
Additional mesh refinement may be required in some circumstances so as to obtain a more
quantitative picture.

Contact NUMECA local sales or support office for any question or information you may require. To
allow NUMECA local sales or support office to help you out within the shortest delays, please
provide a detailed description of the observed behavior and performed analysis.

4 FINE™/Acoustics 8.1 Tutorials

 TUTORIAL 1.

NACELLE (2D)

This tutorial provides an introduction to axi-symmetric aero-engine nacelle modeling, in order to show
how it is possible to predict the in-duct sound propagation of a component (i.e. a duct mode) of the
tonal noise generated by the fan.

1.1 INTRODUCTION

1.1.1 Prerequisites

The CFD simulation performed with the NLH method is not part of this tutorial; specific tutorials
on the CFD-NLH analysis of turbomachinery configurations are available in the FINE™/Turbo
documentation.
The goal of this tutorial is to provide an introduction to the application of the Finite Element
Method (FEM) module of FINE™/Acoustics for the simulation of the fan duct radiated noise.
The user is supposed to be already familiar with the menu structure of FINE™/Acoustics.
The acoustic mesh (*.vnx) and the NLH harmonic results (*.cgns) for the noise source are
provided in specific files enclosed with this tutorial.

1.1.2 Problem description

The Advanced Nacelle Control Fan (ANCF) (FIGURE 1.1) test rig has been developed in the
nineties to provide a test environment and experimental data for noise propagation in nacelles.
The test measurements have been wildly used to validate aerodynamic and acoustic codes.

5 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.1
Aero-Acoustic Propulsion Laboratory (left) and Advanced Noise Control Fan (right)

The ANCF is composed of a 4-foot diameter ducted fan; the nacelle is composed of 14 stator
vanes and 16 rotor blades, depicted respectively in red and in blue in FIGURE 1.2. The blades
are rotating at a speed of 1800 RPM, leading to a the fundamental blade passing frequency (BPF)
of 480 Hz.

FIGURE 1.2
Geometrical configuration

6 FINE™/Acoustics 8.1 Tutorials

1.2 AERODYNAMIC ANALYSIS

The domain taken into account for the computation includes the whole nacelle duct, the lip and
the near field. FIGURE 1.3 shows the computation domain and the extension to the near field.
Compared to full unsteady 360° sliding grid computations, the number of meshed points is greatly
reduced as only one blade passage is meshed instead of all of them. External, solid wall and outlet
boundary conditions are applied on the domain boundary.

FIGURE 1.3
ANCF nacelle CFD computational domain

The CFD mesh has been generated with the AutoGrid5™, the NUMECA International automatic
hexahedral structured grid generator. As shown in FIGURE 1.4, a fictive shroud line is defined,
separating the main channel and the radial extension. The radial extension is meshed as a
meridional effect and is matching with the main channel. The first cell size is fixed to 10 μm on
the solid surfaces, ensuring values of y+ not to exceed 10 above solid surfaces.

7 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.4
Meridional mesh

FIGURE 1.5 summarizes the boundary conditions that have been used for the CFD computation.
The turbulence is modeled by the one- equation Spalart- Allmaras model, with 3 harmonics
considered for the perturbation in the framework of the NHL computation. To increase the
convergence speed, a full-multigrid strategy is used with 3 grid levels. The simulation took about
8 hours on 48 processors.

FIGURE 1.5
Boundary conditions

A snapshot of the blade wake development is shown in FIGURE 1.6 for blade-to-blade planes at
50% span and 95 % span. The efforts in terms of mesh generation and the simulation approach
have allowed for a good resolution of the important flow phenomena. The transition of the
contour lines is barely visible across the rotors interface. The continuity cannot be rigorously
reproduced since a limited number of harmonics have been used. The interaction between the
wakes of the rotor and the stator, in particular the slicing of the rotor wake by stator can be clearly
observed.

8 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.6
Entropy wake at 50 % span (left) and at 95 % span (right)

Experimental aerodynamic results are available for this configuration. FIGURE 1.7 shows the
comparison between the experiment and the CFD computation for the mean static pressure in
front of the fan. One can see that the CFD computation is matching almost perfectly the
experiment.

FIGURE 1.7
Mean static pressure

1.3 ACOUSTIC ANALYSIS

Here below is reported the list of the computational steps required to assemble the acoustic model.

9 FINE™/Acoustics 8.1 Tutorials

 1. Import the acoustic mesh:
l The mesh is imported with the command Discretization > Import New > File (FIGURE
1.8).
l Select the file of type VNoise (*.vnx) and import the file "Nacelle_2D_mesh.vnx".

FIGURE 1.8
Import mesh

l the Import Option and Transformations windows appears (FIGURE 1.9).

l Select FEM for Mesh type (if not already selected).
l Unselect all the quantities for Global data to be imported.
l Unselect Extract Boundary surfaces from volumes (if not already unselected).

10 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.9
Import of CFD mesh and NLH harmonic data

Set the model dimensions with command Mesh > Set Dimension:
l Select Axial symmetric in Dimension and set Azimuthal order equal to <16>, which is
the number of the fan blades and the azimuthal order of the acoustic duct mode that will be
excited in the source region.
l Set the Z axis as the Axial symmetric axis and the XZ plane as the Model plane location.

11 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.10
Axi-symmetric settings

Rename the new discretization created as <FEM model>. For this purpose apply the command
Discretization > Rename.
With the command Analysis > Mesh info it is possible to visualize the information on the
domain and the mesh size (FIGURE 1.11). The maximum frequency that can be computed is
calculated by default considering a spatial resolution of 6 points per wavelength (this parameter
can be changed in Options > Parameters for > Mesh Quality).

12 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.11
Acoustic mesh information

2. With the button < F3 > pressed select all the nodes and the volume elements. Select the
operation Extract Skin in the dialog box to extract the skin (FIGURE 1.12). The group
"Skin2D" is automatically created.

13 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.12
Group Skin2D

3. Import the mean flow field:

l Use the command Boundary_C. > Import_Adv. .
l Select the file type FINE/Turbo (*.cgns) and import the file "ANCF-Stator-5_25in-ext_
[Link]".
l Import on group "Global".
l Keep the default setting in the import mask (FIGURE 1.13).

14 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.13
Import mask

l Set the Min to <1700> and Max to <1900>, click on the button Select RPM to activate the
blocks characterized by a rotational speed of 1800 RPM in the Enable Zones to be
imported dialogbox, click on the OK button (FIGURE 1.14).

15 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.14
Selection of mesh blocks for importing

l Set <16> for Number of Blades for Row1 and click OK button (FIGURE 1.15 top)
l In the NLH Data mask set the Name Prefix as Mean Flow (FIGURE 1.15 bottom).
l Select Pressure, Density, Velocity for data to be extracted.
l First/last harmonic to extract: indicate the interval 0-100 in order to consider all the
harmonics used in the NLH simulation.
l Select Mean Flow and pick Convert Rotating To Fixed in Reference Frame drop-down
menu for processing the NLH data during import.
l Generate symmetric data: None. In this nacelle model the possibility to define symmetry
planes is not exploited.
During the transformation from the rotating to the fixed reference system, FINE™/Acoustics
takes into account the rotor kinematic effects on the mean-flow reconstructed in the fixed
frame. This transformation process requires the generation of a periodic time signal at each
node of the acoustic mesh. The time signal is thereafter FFT analyzed to yield the final
frequency content in the fixed frame (the mean flow corresponds to the frequency of 0 Hz).
The user shall then set the following parameters used for the time signal generation:
l Number of Time steps: 32 (default value). This parameter indicates the time steps present
in the time signal (32 points are considered a good value to obtain an accurate signal
discretization).

16 FINE™/Acoustics 8.1 Tutorials

 Click OK to start the CFD Mean Flow data import. Click OK to remove the boundary layer
automatically on the group "Global".

17 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.15
Mean flow import

The mean flow field is finally imported (Boundary C > Show).

FIGURE 1.16
Mean flow static pressure

4. Create boundary groups:

l Select the command Options > Parameters for > Coincident nodes and set the parameter
related Do not eliminate if angle between the normals greater than to < 30 > deg
(FIGURE 1.17). This parameter controls the sensitivity of the command (applied in the
next step) to extract free faces of volume elements on the basis of the angle between their
normals.

18 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.17
Extract group from sharp angles

l To create groups containing free faces of volume elements (skin), based on the angle
between their normals, use the command Groups > Create From > Sharp Angle.
l Select the group prefix name as <grp> and click OK button to create the groups. Inspect
the groups created (FIGURE 1.18).

19 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.18
Extracted groups from sharp angles

l Use the command Groups > Transform / Delete to join the groups "grp:2" and "grp:9"
(FIGURE 1.19). Keep the Operation as Join groups, set New Group name as <duct>,
keep the key <Ctrl> pressed and select the groups "grp:2" and "grp:9", select the Select as
Yes, click on the button OK.

FIGURE 1.19
Boundary groups modification

l Use the command Groups > Transform / Delete to join the groups "grp:4" and "grp:10"
and name it as <inlet>. Join the groups "grp:1", "grp:5", "grp:6", "grp:7" and "grp:8" and
name it as <PML>.

20 FINE™/Acoustics 8.1 Tutorials

 l Use the command Groups > Transform / Delete and delete the "grp:1", "grp:2", "grp:4",
"grp:5", "grp:6", "grp:7", "grp:8" ,"grp:9" and "grp:10".
l Use the command Groups > Change Order/Rename to rename the "grp:3" as <hub>.
(FIGURE 1.20).

FIGURE 1.20
Editing group name

l After boundary name editing the boundary group names should be as shown in the
FIGURE 1.21.

21 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.21
Boundary groups

5. Boundary conditions:
Define the PML boundary condition for the absorbing buffer layers:
l Use the command Boundary C.> PML > New to create the PML boundary condition for
the group PML (FIGURE 1.22):
l Set Name as <PML BC> and select Group as PML.
l Select Type as Standard and Thickness as <0.5> (this means 0.5 times the wavelength
corresponding to the analyzed frequency. This value is automatically computed by
FINE™/Acoustics).
l Keep Damping as <1.5> and Exponent as <2> (these default values are sufficient to
provide a general good behavior of the PML non-reflecting BC).
l Select External (Automatic) for PML Buffer.
l Select the option Far Field Radiation, unselect the option Use BEM at low Freq.
l Click on OK to create the PML boundary condition.

22 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.22
"PML front" boundary condition

l Apply the following commands to create the "PML Inlet" boundary condition (FIGURE
1.23):
l Set Name as <Inlet> and select Group as inlet.
l Select Type as Annular with duct modes and Thickness as 1.
l Keep Damping as 1.5 and Exponent as 2.
l Click on the button Get from Mesh to get the inner radius and outer radius.
l Select the PML Buffer as Internal.
l Select the source mode type as Duct Mode and click on Define Duct Modes.

23 FINE™/Acoustics 8.1 Tutorials

 l Set the Order Azimuthal to <16>, the Order radial to <1> and the Amplitude to
<100> dB.
l Click on OK to create the PML boundary condition.

24 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.23
PML boundary condition for Inlet

The "PML Inlet" boundary condition is a Perfectly Matched Layer (PML) boundary condition
that imposes an incident sound field (the noise source) expressed in terms of acoustic duct
modes [8]. These latter provide a mathematical description of the acoustic pressure distribution
along the duct. Duct modes are interesting mathematically because they form, in general, a
complete basis by which any acoustic pressure solution can be represented. Physically, modes
are interesting because they are solutions in their own right, not just mathematical building
blocks, and by their simple structure the usually complicated behavior of the total field is more
easily understood.
The acoustic source decomposition in duct modes is used in nacelle calculations, where the
common practice is to define an input source in terms of duct modes characterized by an
azimuthal order (m), a radial order (n), a mode amplitude and a phase (FIGURE 1.24).

FIGURE 1.24
Duct modes formulation

25 FINE™/Acoustics 8.1 Tutorials

 The source definition in terms of duct modes is typically applied when the real noise source
(coming for example from the fan) is unknown. The common practice is to build an estimated
acoustic source defined on the basis of semi-empirical models and/or measurements.
In the frame of this tutorial one single mode is excited, however it is possible to compose a
source with several modes. With reference to FIGURE 1.23 this can be done by adding a row
for each new duct mode introduced. In FINE™/Acoustics duct modes can be specified for
circular and annular ducts with constant section, in presence of a mean-flow and also with
impedance distribution over the inner and the outer duct walls.
6. Now the non-uniform mean flow imported from the CFD simulation has to be applied to the
computational domain:
l Define a new domain using the command Boundary_C .> Domains > New.
l Select Name as <Air> (FIGURE 1.25).
l Activate Non-uniform button and click on the button Flow Definition.
l In the Flow Definition menu select the appropriate flow quantities to define the non-
uniform flow properties Pressure, Rho, Vx, Vy and Vz.
l Click on OK to create the domain.

26 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.25
Domain definition

It is finally possible to define the type of the boundary conditions.

l Define boundary condition and domain using the command Boundary C . > Define as
shown in the FIGURE 1.26.

27 FINE™/Acoustics 8.1 Tutorials

28 FINE™/Acoustics 8.1 Tutorials
 FIGURE 1.26
Set boundary condition and domains to groups

7. Import microphone array:

l Import the microphone file through the menu File > Import > File...
l Select the file "[Link]".
l Select the option Create Group and enter the name <Microphones>.
l Click OK to import the microphones.

29 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.27
Import microphone - Mask

8. Start the FEM analysis:

30 FINE™/Acoustics 8.1 Tutorials

 l Save the project with the name <Nacelle_axi-[Link]> using the command File >
Save as.
l Start the frequency analysis using the Command Analysis > Freq. Response(FIGURE
1.28).
l Specify the Name of the analysis (e.g. <Cuton>).
l Set Start Freq. to <1000> Hz, End Freq. to <2500> Hz and Delta Freq. to <500> Hz,
which will perform an analysis for 4 frequencies.
l Select Generate Mesh without run and press Start to generate a new discretization
containing the PML.

31 FINE™/Acoustics 8.1 Tutorials

 l Select the new discretization created ("Cuton") in the drop-down menu next to Geom, and
click on Analysis > Freq. Response. Press the button Start to start the analysis.

A warning message will appear because the domain Air does not ensure 6 points per wave at
2500 Hz. This tutorial just quickly shows how FINE™/Acoustics works so that a course mesh
is used. To ensure accuracy, 6 points per wave is recommended.

32 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.28
Frequency Response Analysis

l Once the computation is complete save the project.

In FINE™/Acoustics it is possible to include sound absorbing materials/structures like acoustic
liners. Their absorbing effect is modeled through an acoustic admittance distributed over the
duct surface where the acoustic treatment is located (the acoustic admittance represents the
frequency transfer function between the acoustic velocity and the acoustic pressure [9]). Apply
the following steps:
l Select the discretization "Model 0".
l Insert a liner in the duct walls by keep pressing the key <F3> and selecting the nodes
where the liner is located as shown in FIGURE 1.29.
l Select Create Group, click OK,
l Enter the name <Liner>.

33 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.29
Acoustic liner definition (white selection)

l Define the new sound absorbing material for the liner through its equivalent admittance
using the command Boundary C. > Materials > New , specify Name as <liner_mat> and
select Admittance as Constant type. Specify the real part as < 0.00115340> [m²s/kg] and
the imaginary part as <0.0002883506> [m²s/kg], as shown in FIGURE 1.30 - left. These
values are chosen to obtain a good noise attenuation at the frequency investigated (2500
Hz).
l Select the correction of Myers to allow to consider the effect of the grazing flow over the
liner admittance [10].
l Assign the "Liner" group to the respective material using the command Boundary C. >
Define > Material (FIGURE 1.30 - right).
l Repeat the Frequency Response Analysis giving the name <Cuton_liner> to the solution.

34 FINE™/Acoustics 8.1 Tutorials

35 FINE™/Acoustics 8.1 Tutorials
 FIGURE 1.30
Defining Liner Material

1.4 POST-PROCESSING

9. Once the analysis is completed it is possible to visualize the acoustic pressure [Pa], the acoustic
velocity [m/s] and the sound pressure level [dB] over the computational grid. For this purpose

36 FINE™/Acoustics 8.1 Tutorials

 visualize the "Global" group, enable the result view mode (FIGURE 1.31).
Through the toolbar on the bottom of the screen (FIGURE 1.32) it is possible to select the
quantity to be visualized: modulus, phase, real and imaginary part of the acoustic pressure and
velocity. It is also possible to create animations useful to evaluate the sound propagation.
For this purpose visualize the real part of the acoustic pressure in physical units (i.e. no
Decibels), select the number of frames desired and click on Animation. To visualize the sound
pressure level (SPL) right click on the screen then apply the command Metrics > Decibels (in
this case remember to select the visualization of the modulus).

FIGURE 1.31
Visualize the Frequency Response results

FIGURE 1.32
Selection of acoustic quantities and creation of animations

The overall results obtained are presented here below:

l FIGURE 1.33 shows the comparison between the propagation of the same acoustic mode,
i.e. mode (16,1), excited at two different frequencies, respectively 1000 Hz and 2000 Hz.
Each acoustic duct mode is characterized by a frequency called "cut-off", that is the
frequency at which the mode starts to propagate in the duct. At higher frequencies the mode
is "cut-on" and it can propagate along the duct. At lower frequencies the mode is "cut-off"
and it is damped exponentially, as show in image on the left.
l FIGURE 1.34 highlights the liner absorption and scattering effect over the propagating
sound waves. In the image on the right it is possible to appreciate how the acoustic pressure
gradient at the duct walls has a component normal to the duct surface, indicating that part of
the acoustic energy is absorbed through the admittance boundary condition.

37 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.33
Acoustic Pressure distribution (real part, [Pa]) without liner at 1000 Hz (top) and at 2000 Hz
(bottom).

38 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.34
Acoustic Pressure distribution (real part, [Pa]) without liner (top) and with liner (bottom) at
2000 Hz.

39 FINE™/Acoustics 8.1 Tutorials

 10. Post processing for microphones as follows:
l Go to the menu Post-Processing > New 2D Window,
l Right click on the window and select the contextual menu Metrics,

FIGURE 1.35
Metrics selection

l Select dB rms (SPL) and click OK for confirm,

l Right click on the window and select the contextual menu 2D Data Definition >
Solutions, keep the default selection and click OK to confirm,

40 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.36
Solution selection

l Choose the group "Microphones" and confirm.

FIGURE 1.37
Selection of the group "Microphones"

l Select one microphone and right click on it, select the contextual menu Activate All. Select
the ratio button Add, click OK to confirm.

41 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.38
Selection of the microphones

l The values on microphones will be shown.

l Right clicking on the window and select the contextual menu Curve Export > ASCII File
will export the results to a file. Then the user can post processing the results in an external
tool. The following figure shows the noise in the far field at the frequencies 2000 Hz and
2500 Hz.

42 FINE™/Acoustics 8.1 Tutorials

 FIGURE 1.39
Post-processing: Noise directivity

11. The Sound Pressure Level (SPL) is calculated by FINE™/Acoustics according to the
following formula: SPL=20 Log10 (Ppeak/Pref ) where:
Ppeak = peak amplitude of the acoustic pressure ;
Pref = reference acoustic pressure = 2e-5 Pa.

1.5 ACKNOWLEDGMENT

NUMECA International would like to acknowledge Dr. Daniel L. Sutliff and the NASA Glenn
Research Center for providing access to the ANCF database.

1.6 REFERENCES

[1] Raymond A. Loew, Joel T. Lauer, Joseph McAllister, Daniel L. Sutliff (2006), The advanced
noise control fan, 25th AIAA Aerodynamic Measurement Technology and Ground Testing
Conference, 5 - 8 June 2006, San Francisco, California

43 FINE™/Acoustics 8.1 Tutorials

 [2] Adrien Mann, Franck Pérot, Min-Suk Kim, DamianoCasalino , Ehab Fares (2012), Advanced
Noise Control Fan Direct Aeroacoustics Predictions using a Lattice-Boltzmann Method, 18th
AIAA/CEAS Aeroacoustics Conference (33rd AIAA Aeroacoustics Conference) 04 - 06 June
2012, Colorado Springs, CO
[3] Vilmin, S., Lorrain, E., Hirsch, Ch., Swoboda, M. (2006). Unsteady flow modeling across the
rotor/stator interface using the nonlinear harmonic method, ASME Turbo Expo 2006, GT2006-
90210, Barcelona, Spain.
[4] He, L., and Ning, W. (2005). Efficient Approach for Analysis of Unsteady Viscous Flows in
Turbomachines, AIAA journal, Vol. 36, No. 11, pp. 2005-2012.
[5] Purwanto, A., Deconinck, T., Vilmin, S., Lorrain, E., Hirsch, C. (2011). Efficient Prediction
of Nacelle Installation Effects at Take-Off Conditions, ETC-9, Paper No 316.
[6] di Francescantonio, P. (1997). A New Boundary Integral Formulation for the Prediction of
Sound Radiation, Journal of Sound and Vibration 202(4), 491-509.
[7] McAllister, J., Loew, R.A., Lauer, J.T., Sutliff, D.L., (2009), The Advanced Noise Control
Fan Baseline Measurements, NASA/TM-2009-215595, AIAA 2009-0624.
[8] [Link] (1999), "Sound transmission in slowly varying circular and annular lined ducts
with flow". J. Fluid mech., vol. 380, pp. 279-296.
[9] D.T. Blackstock (2000), "Fundamentals of Physical Acoustics", Wiley- Interscience
Publication.
[10] Myers, M.K., (1979),"On the acoustic boundary condition in the presence of Flow", Journal
of Sound and Vibration, vol. 65, No 1, pp. 125-136.

44 FINE™/Acoustics 8.1 Tutorials

 TUTORIAL 2.

NACELLE (3D) - WIZARD APPROACH -

TONAL NOISE ANALYSIS

This tutorial provides a description of the steps to perform in the Turbomachinery wizard in order to
compute the tonal noise generated by a 3D aero-engine nacelle configuration.

2.1 INTRODUCTION

2.1.1 Prerequisites

The CFD results (*.cgns) needed for the import of the noise sources are provided in specific files
enclosed with this tutorial. Further details about the minimum required inputs can be found in
Inputs and outputs.

2.1.2 Problem description

The Advanced Nacelle Control Fan (ANCF) test rig (see the following picture) has been
developed in the nineties to provide a test environment and experimental data for noise
propagation in nacelles. The test measurements have been wildly used to validate aerodynamic
and acoustic codes.

45 FINE™/Acoustics 8.1 Tutorials

 The ANCF is composed of a 4-foot diameter ducted fan; the fan has 14 stator vanes and 16 rotor
blades, depicted respectively in red and in blue in the following figure. The blades are rotating at a
speed of 1800 RPM, leading to a the fundamental blade passing frequency (BPF) of 480 Hz.

2.2 AERODYNAMIC ANALYSIS

The domain taken into account for the computation includes the whole nacelle duct, the lip and
the near field. the following figure shows the computation domain and the extension to the near
field. Compared to full unsteady 360° sliding grid computations, the number of meshed points is
greatly reduced as only one blade passage is meshed instead of all of them. External, solid wall
and outlet boundary conditions are applied on the domain boundary.

46 FINE™/Acoustics 8.1 Tutorials

 The CFD mesh has been generated with the AutoGrid5™, the NUMECA International automatic
hexahedral structured grid generator. As shown in the following figure, a fictive shroud line is
defined, separating the main channel and the radial extension. The radial extension is meshed as a
meridional effect and is matching with the main channel. The first cell size is fixed to 10 μm on
the solid surfaces, ensuring values of y+ not to exceed 10 above solid surfaces.

the following table summarizes the boundary conditions that have been used for the CFD
computation. The turbulence is modeled by the one-equation Spalart-Allmaras model, while the
NLH computation has been performed with 3 harmonics per perturbation. To increase the
convergence speed, a full-multigrid strategy is used with 3 grid levels. The simulation took about
8 hours on 48 processors.

47 FINE™/Acoustics 8.1 Tutorials

 A snapshot of the blade wake development is shown in the following figure for blade-to-blade
planes at 50% span and 95 % span. The efforts in terms of mesh generation and the simulation
approach have allowed for a good resolution of the important flow phenomena. The transition of
the contour lines is barely visible across the rotors interface. The continuity cannot be rigorously
reproduced since a limited number of harmonics have been used. The interaction between the
wakes of the rotor and the stator, in particular the slicing of the rotor wake by stator can be clearly
observed.

Experimental aerodynamic results are available for this configuration. the following figure shows
the comparison between the experiment and the CFD computation for the mean static pressure in
front of the fan. One can see that the CFD computation is matching almost perfectly the
experiment.

48 FINE™/Acoustics 8.1 Tutorials

 Prior to run the Turbomachinery wizard, make sure that the workstation considered for the acoustic
analysis complies with the minimum hardware requirements specified in the User Guide .
HEXPRESS™/Hybrid should also be installed on the workstation (latest version supported: v7.1).

49 FINE™/Acoustics 8.1 Tutorials

2.3 ACOUSTIC ANALYSIS

1. Launch FINE™/Acoustics and start the Turbomachinery wizard from

Wizard>Turbomachinery in the main menu bar.

The window shown in the picture below appears soon after. Press Next to start the acoustic
setup.

50 FINE™/Acoustics 8.1 Tutorials

 2. After pressing the button Browse, select the directory in which all the files generated by the
wizard will be stored. If the directory does not already exist, a new one can be created by
using the button Create New Folder. Name this directory ANCF_wizard_tonal.

Press Next to continue.

3. Import the CFD data through the button Browse located in the section "CFD Results".
Select the file "ANCF- Stator- 5_ 25in- ext- ref_ 000- nlh- 3h_ [Link]" in the
FINE™/Acoustics tutorial directory provided in the package.
Select Axial for the type of turbomachinery configuration (section "Turbomachinery
Configuration"), and both Tonal for the type of acoustic analysis to perform (section
"Analysis Type").

51 FINE™/Acoustics 8.1 Tutorials

 Additional information about this step can be found in the User Guide.
Press Next to continue.

The broadband noise analysis can also be performed, but the analysis will require a longer time
and more computational resources. An example of broadband noise analysis is described in
Tutorial 5.

4. Select the domains to consider for the acoustic analysis: keep "ROW1" and "ROW2" selected.

Additional information about this step can be found in the User Guide.
Press Next to continue.

52 FINE™/Acoustics 8.1 Tutorials

 5. Remove duct extension: keep the defaults values for each row and press Next to continue.
To switch from "ROW1" to "ROW2" and inversly, press the corresponding radio button.

Additional information about this step can be found in the User Guide.
6. Select the radiating surfaces: keep the external surface and outletselected (namely "ROW1-
External" and "ROW2-Outlet"), with the condition Far-field.

53 FINE™/Acoustics 8.1 Tutorials

 Additional information about this step can be found in the User Guide.
Press Next to continue.
7. Define the smallest geometrical not to capture for the meshes generated during the propagation
analysis: specify a tip gap of <0.000762> m.

Additional information about this step can be found in the User Guide.
Press Next to continue.
8. Define model parameters: set the maximum frequency to <1600> Hz and keep the defaults
values for the other options.

54 FINE™/Acoustics 8.1 Tutorials

 Additional information about this step can be found in the User Guide.
Press Next to continue.
9. Define model settings: tick the option Adavanced settings . Set FEM mesh refinement
[pts/wavelength] to <12.0>, and activate the option Enable Radiation surface resize. Keep
the other parameters and options to their default values/status.

55 FINE™/Acoustics 8.1 Tutorials

 Additional information about this step can be found in the User Guide.
Press Next to continue.
10. Resize the radiating surface:
l Radius of the external duct: R = 0.79017
l Forward position of the fan duct: Z0 = -1.5448
l Resizing factor: 0.3

56 FINE™/Acoustics 8.1 Tutorials

 Additional information about this step can be found in the User Guide.
Press Next to continue.
11. Split domains into subdomains:

57 FINE™/Acoustics 8.1 Tutorials

 Using the entry field the top of the window, specify the minimum position of the cutting plane
of "ROW1" at <-0.2> m, and the maximal position of the cutting place of "ROW2" at <0.5>
m.

Additional information about this step can be found in the User Guide.
Press Next to continue.
12. Select rotating domains: keep the subdomains selected by default and press Next to continue.

58 FINE™/Acoustics 8.1 Tutorials

 Additional information about this step can be found in the User Guide.
Press Next to continue.
13. Select the subdomain(s) on which the NLH will be interpolated.

Additional information about this step can be found in the User Guide.
Press Next to continue.
14. Click on Next to start the analysis.

59 FINE™/Acoustics 8.1 Tutorials

 The computation requires approximately 15 hours to complete on 12 processors (Intel Xeon
CPU E5-2620 v4 @ 2.10GHz).
15. Perform the post-processing: keep the default selection.

60 FINE™/Acoustics 8.1 Tutorials

 Additional information about this step can be found in the User Guide.
Press Next to continue.
16. Radiated Sound Power: click successively on PWL (narrow band) and on PWL (1/3 octave
bands) to generate the spectra expressed in terms of narrow band and 1/3 octave bands,
respectively. The figures displayed are automatically saved inside the directory containing the
files generated by the wizard.

61 FINE™/Acoustics 8.1 Tutorials

 Additional information about this step can be found in the User Guide.
Press Back in order to go back to the initial post-processing window. Select the option press
Sound Pressure Level at microphones and press Next to access the window enabling to
define the position of the microphones.
17. Sound Pressure Level at microphones: tick the first 6 microphones in the left part of the
window and enter the position shown in the table below. Keep the option No ground
activated in the section "Define a reflective ground surface".
Click on the buttons All Mics (narrow band) and All Mics (1/3 octave bands) to generate
the sound pressure levels expressed in terms of narrow band and 1/3 octave bands at the
different microphones.

62 FINE™/Acoustics 8.1 Tutorials

63 FINE™/Acoustics 8.1 Tutorials
 Microphone ID X Y Z
1 3.657 0 -1.545
2 3.194 0 -3.408
3 1.852 0 -4.780
4 0 0 -5.283

The picture below illustrates the position of the microphones relatively to the external radiating
surface. The microphones are located on a circular arc of radius 3.66 m (12 ft). The
microphone lying on the rotation axis has a polar angle of 0 degree.

64 FINE™/Acoustics 8.1 Tutorials

 The pictures below represent the broadband noise observed at microphone 2 and 3.

65 FINE™/Acoustics 8.1 Tutorials

 A comparison with experimental measurement is shown below for the first three Blade Passing
Frequencies (BPFs).

66 FINE™/Acoustics 8.1 Tutorials

67 FINE™/Acoustics 8.1 Tutorials
 At 30° and 60°, the results provided by the wizard are in line with the measurements (within
~5 dB).
The large under-prediction at 0° is due to the absence of low-order azimuthal modes (e.g.
planar wave) in the NLH solution. Experimentally, these modes are generated by the non-
perfect axisymmetric geometry and the flow conditions, which are not perfectly uniform.
For the microphone at 90°, the under-predication of ~10 dB which can be observed at the 3rd
BPF in the results provided by the wizard with respect to the experimental data may be
explained by numerical dissipation occurring in the CFD computation. A finer CFD mesh may
be required to better match the acoustic measurement.
Additional information about this step can be found in the User Guide.
Press Next to continue.
18. Complete the analysis: Click on Quit to exit the wizard.

68 FINE™/Acoustics 8.1 Tutorials

2.4 ACKNOWLEDGMENT

NUMECA International would like to acknowledge Dr. Daniel L. Sutliff and the NASA Glenn
Research Center for providing access to the ANCF database.

2.5 REFERENCES

[1] Raymond A. Loew, Joel T. Lauer, Joseph McAllister, Daniel L. Sutliff (2006), The advanced
noise control fan, 25th AIAA Aerodynamic Measurement Technology and Ground Testing
Conference, 5 - 8 June 2006, San Francisco, California
[2] Adrien Mann, Franck Pérot, Min-Suk Kim, DamianoCasalino , Ehab Fares (2012), Advanced
Noise Control Fan Direct Aeroacoustics Predictions using a Lattice-Boltzmann Method, 18th
AIAA/CEAS Aeroacoustics Conference (33rd AIAA Aeroacoustics Conference) 04 - 06 June
2012, Colorado Springs, CO
[3] Vilmin, S., Lorrain, E., Hirsch, Ch., Swoboda, M. (2006). Unsteady flow modeling across the
rotor/stator interface using the nonlinear harmonic method, ASME Turbo Expo 2006, GT2006-
90210, Barcelona, Spain.

69 FINE™/Acoustics 8.1 Tutorials

 [4] He, L., and Ning, W. (2005). Efficient Approach for Analysis of Unsteady Viscous Flows in
Turbomachines, AIAA journal, Vol. 36, No. 11, pp. 2005-2012.
[5] Purwanto, A., Deconinck, T., Vilmin, S., Lorrain, E., Hirsch, C. (2011). Efficient Prediction
of Nacelle Installation Effects at Take-Off Conditions, ETC-9, Paper No 316.
[6] diFrancescantonio, P. (1997). A New Boundary Integral Formulation for the Prediction of
Sound Radiation, Journal of Sound and Vibration 202(4), 491-509.
[7] McAllister, J., Loew, R.A., Lauer, J.T., Sutliff, D.L., (2009), The Advanced Noise Control
Fan Baseline Measurements, NASA/TM-2009-215595, AIAA 2009-0624.

70 FINE™/Acoustics 8.1 Tutorials

 TUTORIAL 3.

CONTRA ROTATING OPEN ROTOR

This tutorial provides an introduction to the application of the Ffowcs Williams- Hawkings (FW- H)
module of FINE™/Acoustics for the evaluation of the noise radiated by a Contra Rotating Open Rotor
(CROR) operating in representative flight environment characterized by the presence of flow with
angle of incidence.
The configuration considered is the F7/A7 Contra Rotating Open Rotor tested in the NASA Lewis 9- x
15-ft Anechoic Wind Tunnel [1].
The tonal noise source is predicted starting from the engine geometry and the CFD calculated flow
properties. The simulation approach relies on the exploitation of an efficient CFD solver based on the
Nonlinear-Harmonic (NLH) method implemented in FINE™/Turbo. The far field noise radiation of the
CROR configuration is evaluated by means of the free- field acoustic propagation tool based on the
Ffowcs Williams-Hawkings (FW-H) formulation implemented in FINE™/Acoustics.
The numerical results obtained are compared with the measurements available in Ref. [1] in order to
highlight the capability of the methodology to predict the unsteady flow field and the radiated sound
field in flight conditions with an angle of attack.

3.1 INTRODUCTION

3.1.1 Prerequisites

The CFD simulation performed with the NLH method is not part of this tutorial. More details can
be found in Ref. [2] . Specific tutorials on the CFD- NLH analysis of turbomachinery
configurations are available in the documentation of FINE™/Turbo.
The goal of this tutorial is to guide the user through the computational steps required in
FINE™/Acoustics for the simulation of the CROR radiated noise. The user is supposed to be
already familiar with the menu structure of FINE™/Acoustics.
The CFD mesh, the relative NLH harmonic results and the FW-H radiating surface meshes are
provided in specific CGNS files enclosed in this tutorial.

71 FINE™/Acoustics 8.1 Tutorials

3.1.2 Problem description

In this tutorial a working description is given to simulate the noise radiated by a Contra Rotating
Open Rotor (CROR) configuration in presence of flow with an angle of attack (FIGURE 3.1).

FIGURE 3.1
Simulated CROR configuration

The propellers have been tested in the NASA Lewis 9- x 15-ft Anechoic Wind Tunnel. The test
results are reported by Woodward (1992) [1] and include acoustic measurements performed over
a side line below the propeller (and the hypothetical aircraft) at a distance of 61 cm. The inspected
polar angles range from 45 to 135 degrees. The angle-of-attack is achieved by rotating the
propeller in the horizontal plane with respect to the incident flow (FIGURE 3.2).

72 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.2
Sketch of the propeller model and polar microphone probe (from ref. [1])

An accurate model of the F7/A7 propeller geometry is not available, therefore an arbitrary
geometry has been designed and optimized in order to be as much as possible equivalent to the
tested one. Number of blades, radius and RPM were kept the same, as well as the generated
thrust. No warranty exists that the other parameters (chord, sweep, rotors spacing) are exactly the
same as the real ones.

FIGURE 3.3
Left: F7/A7 propeller blades[1]. Right: reconstructed blade geometry

The main noise sources potentially activated by rotating propellers are represented by:
Thickness noise (monopoles) which is dependent only on the shape and motion of the blade, and can
be thought of as being caused by the displacement of the air by the rotor blades. It is primarily
directed in the plane of the rotor.
Loading noise (dipoles) which is an aerodynamic adverse effect due to the acceleration of the force
distribution on the air around the rotor blade when the blade is passing through it, and is directed
primarily below the rotor. This source tends to dominate at low blade speed.
Volume sources (quadrupoles) distributed into the flow domain surrounding the blades. These sources
are only relevant at high rotational speed while at low speed their effect on the radiated noise is
negligible and they propagate inefficiently.

73 FINE™/Acoustics 8.1 Tutorials

3.2 AERODYNAMIC ANALYSIS

A short summary is provided on the aerodynamic analysis made. The simulation is based on the
application of the NLH method [3]-[4] in which the flow is decomposed into Fourier harmonics,
typically associated with blade passing frequencies and their multiples, whose number is specified
by the user. A NLH simulation is then performed for a free stream Mach number of 0.2, which is
representative of take-off/approach operation. An angle of attack of +8 degree is studied, in order
to assess the capability of the computational chain to reproduce the effects of the flow incidence
on the blade aerodynamic loads and on the propagation of the sound generated. FIGURE 3.4
shows the CROR characteristics that were kept unchanged during the geometry reconstruction
process, and so these characteristics are identical to the ones used during the experiments (F7/A7).

FIGURE 3.4
Simulated CROR characteristics

The CFD mesh has been created with the software AutoGrid5™, the NUMECA automatic
hexahedral structured grid generator. The computational domain is radially extended to a radius of
0.75 m and is axially extended to a total length of 7.28 m as well. FIGURE 3.5 shows a blade-to-
blade view of the mesh at mid span. Only one interblade channel per row is actually meshed to
run a NLH simulation. The mesh respects standard quality criteria in terms of the maximum
expansion ratio, the minimum orthogonality, and the maximum aspect ratio of cells. The mesh
counts 7.6 million nodes.

74 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.5
Blade to blade view of CFD mesh

FIGURE 3.6 summarizes the numerical parameters that have been employed for the reference
CFD computation. A simultaneous time-marching technique is used to converge to a steady-state
solution of the time-mean and harmonic flow equations by means of an explicit Runge-Kutta
scheme. Acceleration methods to the steady state like local time stepping (with a CFL number of
3) and multigrid (with 3 grid levels) are also used. The turbulence is modeled by the eddy-
viscosity one-equation Spalart-Allmaras model, with the values of y+ not exceeding 10 on the
first layer of cells above solid surfaces. The simulation uses 3 harmonics and 2 perturbations per
blade row. The additional perturbation is due to the incidence of the flow, which is treated as a
perturbation.

FIGURE 3.6
Numerical parameters used for the CFD computation

FIGURE 3.7 shows the far- field boundary conditions used for the computation. Boundary
conditions similar to the ones observed during the experiments, which are representative to take-
off/approach operation, have been imposed. The blades and the hub are considered as smooth and
adiabatic solid wall boundaries. FIGURE 3.8 shows the boundary conditions imposed on the
solid walls.

75 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.7
Far-field boundary conditions

FIGURE 3.8
Boundary conditions on solid surfaces

Some of the results obtained are presented for reference reasons. In FIGURE 3.9 the harmonics
on the right hand side are those induced by the incidence, and the other harmonics are provoked
by the relative rotation of the adjacent rotor, being associated with the 1st BPF, i.e. passing wakes
(only in the aft rotor) and potential effect. The effect of the incidence is clearly visible near the
hub, indicating a variation of unsteady pressure over the whole rotor (360 deg) which can reach
20,000 Pa near the hub. Cumulating the contribution of these harmonics and the mean flow gives
the instantaneous static pressure, shown here on one rotor blade in FIGURE 3.10.

76 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.9
Harmonics of static pressure on suction side of rotor blades

77 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.10
Static pressure on suction side of rotor blades (instantaneous solution)

3.3 ACOUSTIC ANALYSIS

Once the unsteady loads have been evaluated with the NLH analysis, they can be directly
projected over a control surface enclosing the relevant acoustic sources and used as an input to
start the FW-H analysis with the integrated FW-H solver of FINE™/Acoustics.

78 FINE™/Acoustics 8.1 Tutorials

 The FW-H solver applies the Ffowcs Williams-Hawkings equation in several different forms
including permeable and non-permeable formulations applicable to rotating and fixed surfaces [5].
The FW-H formulation allows to compute the sound radiation in the far-field, at probes (virtual
microphones) whose location is defined by the user.
l In the case of propeller tip speed largely subsonic, the integration surface can be placed on the
propeller surface itself, and therefore the non-permeable (solid) /rotating FW-H approach can
be used. This formulation requires the knowledge of the pressure harmonics on the surface
itself.
l In the case of very high propeller tip speed (Mtip >0.9) and in transonic/supersonic conditions,
the aero-acoustic sources are not only present on the propeller surface but in principle they
could be also located in the fluid region around the blades (quadrupoles). In this case it is not
advisable to place the integration surface on the blade itself, but it is recommended to place it at
some distance in such a way to enclose all the source terms. The permeable FW-H approach
shall be selected. Different possibilities are available in FINE™/Acoustics since it is possible
to have the permeable surface rotating with the blades, or to have a fixed surface enclosing the
propellers. Typically the last approach is the most efficient since it can in principle handle all
the cases, while the rotating surface can be applied only if the surface itself is rotating with a
maximum Mach number lower than 1. FIGURE 3.11 provides guidelines for the selection of
the simulation approach.

79 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.11
Guidelines for the selection of the formulations available in FINE™/Acoustics

In order to provide a complete description of the modeling procedure both the above mentioned
FW-H approaches will be described in this tutorial, even if the test case considered involves low
rotational speed (Mtip=0.77, subsonic flow) and the quadrupole sources are negligible.

3.3.1 Solid / Rotating FW-H surface

The FW-H solver based on a non-permeable, rotating surface, lying directly on the propeller
blade surfaces, is initially applied. The aim of the FINE™/Turbo analysis is to compute the steady
and unsteady pressure distribution over the propeller blades. These are affected by two main
effects, respectively the relative interaction between the two contra rotating propellers (FIGURE
3.12 - right) and the relative interaction between the propellers and the external flow with
incidence (FIGURE 3.12 - left).

FIGURE 3.12
External flow/propellers interaction (left). Fwd/Aft propellers interaction (right).

In order to obtain a good compromise between the simulation accuracy and the computational
cost, three harmonics are used in the NLH analysis to model the first three blade passing
frequencies for each row and for each interaction effect. The harmonics and the relative
frequencies for each propeller are reported in FIGURE 3.13.

80 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.13
Harmonics considered in the propeller rotating frame

These harmonics are defined in the propeller rotating (relative) frame. The subsequent
transformation performed by the FW- H module, from the propeller "rotating" frame to the
microphones "fixed" (absolute) frame, generates the noise tones that are effectively radiated to the
far-field and perceived by the receiver. Their frequencies can be calculated as described in ref. [6].
The full integration of the CFD solver and the FW-H solver allows to reduce the mesh handling
and to simplify the setup of the computational model. For example, when using a fixed permeable
FW-H surface, the solver self-consistently reconstructs the CFD solution on the fixed surface
from the CFD solution available in the rotating reference system.
The NLH simulation results (i.e. the computational mesh and the complex pressures at each
harmonic) are imported in FINE™/Acoustics for the subsequent application of the FW- H
calculation. Only the solid surfaces of the propeller blades shall be imported in order to limit the
memory resources and speed up the computational time. The exchange file format is the standard
CGNS file containing the NLH solution on the blades surface. This file is generated by CFView,
the NUMECA CFD post-processor. The imported solution provides the pressure harmonics on
each node of the rotating mesh. This represents the input provided to the FW-H module.
Here below is reported the list of the computational steps required to assemble the
FINE™/Acoustics model and to setup and ran the FW-H solver.
1. The first step is to import in FINE™/Acoustics the file "CFView_input.cgns". This file is
generated by CFView (version 9 or higher) and contains the surface mesh of each blade and
the relative static pressure expressed in terms of mean value and harmonics (FIGURE 3.14).

81 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.14
Data selection in CFView to be exported in CGNS format.

2. Launch FINE™/Acoustics and save a new empty project called <CROR-solid-rotating> with
the command File > Save as. It is suggested to save the project after each step as there is no
"undo" button.
3. Apply the command Discretization > Import new > File, select the filter CFView (*.cgns) in
the bottom right of the window, and select the file "CFView_input.cgns". An import mask
appears. Fill it in as indicated in FIGURE 3.15.
l BEM: only surface elements (panels) are imported.
l Global data to be imported: all data available are imported.
l Extract boundary surfaces from volumes: this checkbox enables the extraction of skin
meshes from volume meshes. Make sure that the option is unselected before pressing OK.

82 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.15
Import of CFD mesh and NLH harmonic data

In the mask shown in FIGURE 3.16, with mouse + right click, activate all the zones contained
in the CGNS file.

83 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.16
Selection of zones to be imported

Fill in the fields highlighted in red in the mask shown in FIGURE 3.17. All the other fields are
already correctly setup by default.
l Name Prefix: give name <FWH>.
l Data to be extracted: Only Pressure.
l First/last harmonic to extract: the indicated interval 0-100 considers all the harmonics
used in the NLH simulation. This filter is useful to sub-select harmonics when a single
perturbation effect (e.g. angle of flow incidence) needs to be studied.
l NLH data processing : select FWH Harmonics and pick Leave unchanged in the
Reference frame dropdown menu.
l Generate symmetric data: selection disabled in FWH simulations.
l Number of interpolation points: selection disabled in FWH simulations with rotating
surfaces.
l KMAX: selection disabled in FWH simulations with rotating surfaces.
l Sectors to be imported: <1> (only one blades set is visualized in the GUI). The user can
visualize all the blades selecting "All".
l Create Domains: when activated a domain is automatically created for each blade.

84 FINE™/Acoustics 8.1 Tutorials

 l Import CFD groups: to import groups present in the CFD mesh (option disabled in this
application).

FIGURE 3.17
Selection of data to be extracted

4. The mesh for a blade set (FIGURE 3.18) and the distribution of the pressure harmonics in
FIGURE 3.19 are finally imported. To visualize the pressure fields select the visualization of
boundary conditions with the button indicated in FIGURE 3.20.

85 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.18
Imported blades set

86 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.19
Harmonics pressure distribution [Pa] over a rotor blade

FIGURE 3.20
Visualization of the boundary conditions

5. The normals of the mesh panels have to be inspected to verify that they point outward, i.e. into
the fluid computational domain. To visualize the normals click on the buttons shown in
FIGURE 3.21. It is possible to see that there are some panels at the blade tips (highlighted in
green) whose normals are pointing inward (FIGURE 3.22). It is necessary to reverse these
normals. This is done by selecting the relative elements moving the mouse on them with the
button < F4 > pressed, until the mouse cursor becomes white. With right click + Select
connected > Sharp angles it is then possible to select the elements on the basis of the surface
discontinuities. Multiple cumulative selections may be necessary to fully select the desired

87 FINE™/Acoustics 8.1 Tutorials

 elements (FIGURE 3.23). Once done, click on the button shown in FIGURE 3.24 to reverse
the normals. The sensitivity to sharp angles can be tuned with the command Options >
Parameters for > Coincident nodes (FIGURE 3.25), modifying the value indicated in the
field Do not eliminate if angles between normal greater than. Lower values restrict the
extension of the area selected around the initial point picked with the mouse.

FIGURE 3.21
Buttons to visualize the elements normals

FIGURE 3.22
Normals to be reversed (green panels)

88 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.23
Selection of panels for which the normals shall be reversed

FIGURE 3.24
Button to reverse the orientation of the elements normals

FIGURE 3.25
Sensitivity control of the selection tool based on sharp angles

89 FINE™/Acoustics 8.1 Tutorials

 6. If required by the user it is possible to visualize the computational domains for the forward and
the aft blades. The domains are automatically created by FINE™/Acoustics on the basis of the
kinematic data contained in the CFView imported file. For this purpose apply the command
Boundary C. > Domains > Modify and select the domain of interest. The information
contained in FIGURE 3.26 are displayed. The fluid properties and the flow parameters
indicated here are not relevant as they will be overwritten by the data inserted by the user in
the FW-H solver mask at step 8. The blades kinematic setting can be inspected by clicking on
the button Kinematics.

FIGURE 3.26
Forward rotor computational domain

7. In order to compare the simulation results with the measurements available, a total number of 9
far- field polar microphones are considered, located in the same positions for which
experimental data are available. The microphone locations are defined in FIGURE 3.27. The
array is located over a side line below the propellers (and the hypothetical aircraft), at a
distance of 61 cm from the axis of rotation. The inspected polar angles (θ) ranges from 50 to
130 degrees, with a spacing of 10 deg, with the angles measured relatively to the center of the
aft propeller disk, in the aft direction (FIGURE 3.28). To create the microphones click on the

90 FINE™/Acoustics 8.1 Tutorials

 button for the microphones creation (FIGURE 3.29) and input the x-y-z coordinates. Once the
microphones are created select them with <F2> + mouse, right click on the screen and with the
command Groups > Create create a group called "Microphones". Microphone arrays (planar,
annular, spherical) can be used when it is required the visualization of the sound propagation
through field plots or animations. For a correct visualization the array shall guarantee a spatial
resolution of at least 6 points per wavelength.

FIGURE 3.27
Microphones coordinates

FIGURE 3.28
Polar microphones array located below the CROR

91 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.29
Definition of microphones coordinates

8. It is finally possible to activate the FW-H input mask (command Analysis > FWH-NLH) and
input the following parameters (FIGURE 3.24):

92 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.30
FW-H analysis input mask

Once the FW-H input mask is setup the analysis can be started. It is suggested to use a
"Cluster solver" (see in Options > Directories / solvers) to reduce the computational time.
The CPU time scales linearly with the following parameters:
l Nodes number
l Microphones number
l Calculation period
l Maximum frequency
l Number of cores (inverse proportionality)
The simulation of this model takes a total time of 1.7 hrs on an Intel Xeon @ 2.40GHz PC
equipped with 8 cores.
9. Once the results are obtained, it is possible to visualize the noise levels at the microphones with
the command Post-processing > New 2D window. An empty window appears. Make a right-
click on it and put the "Metrics" in "Decibels" (FIGURE 3.31). Then right-click and select 2D
data definition > Solutions (FIGURE 3.32). Select the solution to be visualized, the relative
quantity and click on OK. A new mask appears (FIGURE 3.33) showing the groups available
in the model. Select the "Microphones" group, activate (with right click) and add all
microphones. Click on OK and the respective sound pressure levels are finally visualized
(FIGURE 3.34). They can be exported in ASCII format through the command right-click +
Curve Export > ASCII file.

93 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.31
Metrics definition (Decibels)

94 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.32
Selection of the solution to display

FIGURE 3.33
Microphones selection

95 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.34
Sound pressure level spectra at microphones

3.3.2 Permeable/ Fixed FW-H surface

In this second model a different approach based on the FW-H permeable formulation is
exploited. The simulation procedure is similar to the previous one, i.e. the steady and unsteady
flow quantities are computed over a control surface enclosing the acoustic sources. This
information is then provided in input to the FW-H solver that computes the sound radiated to
the far-field microphones. There are however some differences:
l The radiating surface is not lying on the blades surface but it encloses also part of the
surrounding fluid (with the relative acoustic sources).
l The surface has a cylindrical shape and its extension is 360 deg (i.e. not an angular sector
enclosing one single blade).
l The harmonic quantities to be computed on the radiating surface include not only pressure
but also density and velocity (three components).
l The radiating surface is defined in the fixed (absolute) microphones reference system,
therefore FINE™/Acoustics has to apply a transformation of reference system to the
harmonic values, which are originally computed by NLH in the rotating frame.
The acoustic results obtained with a permeable surface depend on the surface location and on
its spatial discretization. In particular the optimal location of the surface mesh is driven by the
following conflicting objectives:

96 FINE™/Acoustics 8.1 Tutorials

 l The surface shall be sufficiently extended in the near-field to include the volumetric domain
containing the relevant acoustic sources (quadrupoles) and the non- uniform flow
characterized by high flow velocity (providing non negligible convective effects on the
sound propagation).
l The surface shall be located sufficiently close to the source domain in order to minimize the
numerical dissipation of the acoustic waves propagation computed with CFD.
Its optimal size is driven by the following conflicting objectives:
l The surface mesh resolution shall be sufficiently high to correctly reproduce the relevant
gradients of the static pressure, density and velocity fields.
l The surface mesh size shall be as much as possible limited to reduce the computational cost
of the radiation analysis (linearly dependent on the number of nodes).
It is then clear that, for the identification of an optimal surface mesh, it is required the execution
of a sensitivity analysis exploiting different surface sizes and mesh resolutions. This analysis is
not in the scope of this tutorial, whose main goal is to provide a detailed description of the
application procedure for the FW-H radiation analysis.
Here below is reported the list of the computational steps required to assemble the
FINE™/Acoustics model and to setup and ran the permeable FW-H solver.
10. Open a new project with File > New. Save the empty model with File > Save as giving name
<CROR-permeable-fixed>. With command Discretization > Import new > File select the file
type CGNS generic file (cgns.*) and import the file "Fwd_mesh.cgns" containing the IGG
volume mesh enclosing the forward rotor. Setup the import mask as indicated in FIGURE 3.35
- FIGURE 3.36. With the command Groups > Change order / Rename rename the group
"Base#1-" as "Fwd - volumes".

97 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.35
Import of the IGG™ volume mesh

98 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.36
Selection of the groups to be imported

Repeat the operation to import the file "Aft_mesh.cgns" containing the IGG™ volume mesh
enclosing the aft rotor. For this second import use the command File > Import > File. Rename
the group "Base#1-" as "Aft - volumes".
11. The surfaces radiating the sound to the far-field need to be created. For this purpose, select first
the group “Fwd – volumes” previously imported. With <F3> pressed, select the nodes and the
volumes of this group, and select the option Extract skin (FIGURE 3.37) in the window
popping-up.

99 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.37
Extract skin

The group “Skin” is added in the list of available groups. Select this group and press <F4>
over the lateral surface; perform then a right- click on the surface, and select Select
Connected>Sharp angles. Redo the same operation over the external surface, the one whose
normal vector is pointing outward. Finally, a new group can be created by doing a right-click,
and select Group> Create (FIGURE 3.38). Name the new group “Fwd – surface”. The
group “Skin” can then be deleted in Group> Transform/Delete.

100 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.38
Fwd - surface

The same operation should be repeated to generate the radiating surface for the aft part. Make
sure that the elements previously selected to generate the surface “Fwd - surface” are not
selected anymore. For this, click on the button Refresh .

101 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.39
Fwd and Aft surfaces

The group "Skin", as well as the groups "Fwd - volumes" and "Aft - volumes" can be deleted
after these operations.
12. Create a complementary group of the groups "Fwd - surface" and "Aft - surface" in
Group>Transform/Delete, as shown in the picture below. Name the group to delete "to_
delete". After pressing OK, select the group "Global".

102 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.40
Creation of the complementary group

FIGURE 3.41
Selection of the group on which to make the complement

Select the new group created in the menu Grp, and with <F3> pressed, select the group
displayed in the interface. In the window popping-up, select the option Delete to remove all
the entities stored inside the group, as shown in the picture below.

103 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.42
Deletion of the complementary group

To remove the group from the group list, press the button next to the menu Grp. Press OK
in the window popping-up to finalize the deletion.
Now make a check with the command Analysis > Mesh info to verify that the number of
nodes and elements is the one reported in FIGURE 3.43.

104 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.43
Mesh information

13. For a correct application of the FW-H formulation the panel normals have to point outward,
i.e. inside the computational domain. To check their orientation remove the visualization of
nodes, enable the visualization of shaded surfaces and highlight the normal orientation with the
respective buttons indicated in FIGURE 3.44. The red color means "normals outcoming from
the surface", the green color indicates "normals entering in the surface". We need then to
reverse all the normals as the computational domain considered in the FW-H analysis is
external. This can be done making a selection of the mesh with <F3> + mouse all over the
elements displayed in the working space, and select None in the menu popping-up. The user
needs then to click on Mesh > With selection > Invert normals.

105 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.44
Buttons to enable the visualization of the panels normal.

The normals have to be inverted. The button can be pressed to deactivate the selection
made previously.
14. The next step is to apply the command Boundary C. > Import advanced > File, to select the
FINE/Turbo (*.cgns) format and to import the NLH solution file "Position_22_c3f_rear_
ivelsy_ 000_ [Link]". Make the first import on group "Fwd surface". An import mask
appears (FIGURE 3.45). Keep the default values and click on OK.

106 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.45
Import of NLH harmonic data on group "Fwd surface"

With mouse + right click on the subsequent mask (FIGURE 3.46) activate the zones contained
in the CGNS file characterized by a negative rotational speed. These are the zones relative to
the forward rotor. In the subsequent mask define the number of blades relative to the forward
propeller (FIGURE 3.47).

107 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.46
Selection of zones to be imported on group "Fwd surface".

FIGURE 3.47
Definition of the number of blades for the forward propeller.

Fill in the next import mask with the fields highlighted in red in FIGURE 3.48. All the other
fields are already correctly setup by default.

108 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.48
Selection of data to be extracted from the CFD results (forward rotor).

15. The NLH harmonic data are finally interpolated over the surface mesh and transformed from
the rotating (relative) reference system to the fixed (absolute) reference system. To visualize
the harmonics fields select the desired data and frequency in the menu Properties (FIGURE
3.49).

109 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.49
Visualization of the harmonic pressure field [Pa] at the blade passing frequency (BPF) of the
forward rotor.

Reapply step 13 to import the harmonic data on group "Aft surface". Pay attention to define
the appropriate rotational speed and number of blades in the input mask shown in FIGURE
3.47 - FIGURE 3.48
16. Now it is necessary to define the computational domain and the surface kinematics (not
automatically setup when using permeable surfaces). For this purpose apply the command
Boundary C. > Domains > New . Apply the setup indicated in FIGURE 3.50. The flow
parameters indicated by default in the domain mask are not relevant, as they will be
overwritten by the flow parameters defined in the FW-H analysis mask.

110 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.50
Definition of the computational domain and of the FW-H surface kinematics

17. It is finally possible to define the microphones as shown in step 7 and to setup the FW-H
analysis as described in step 8. Pay attention to define the FW-H solver used as "Permeable
FW-H (1A)".
The post-processing of the acoustic results is made as described in step 9.
The simulation of this model takes a total time of 1.5 hrs on an Intel Xeon @ 2.40GHz PC
equipped with 8 cores.

111 FINE™/Acoustics 8.1 Tutorials

3.4 COMPARISON OF ACOUSTIC RESULTS

In the final section of this tutorial the results obtained with the solid-rotating and with the
permeable-fixed FW-H surfaces are compared with the measurements made in the NASA
facility [1]. The results shall be assessed keeping in mind the potentially poor fidelity of the
reconstructed blades geometry used in the CFD model.
FIGURE 3.51 shows the numerically predicted sound pressure level (SPL) spectra obtained at
the microphone with polar angle of 90 deg. The main tonal noise is produced at BPF1 (1399
Hz), BPF2 (1154 Hz) and at the relative interaction frequencies BPF1+BPF2 (2554 Hz) and
2*BPF1+BPF2 (3953 Hz). Index 1 refers to the BPF for of the forward propeller, index 2
refers to the BPF of the aft propeller.
The spurious peaks present aside of the main tones above mentioned are due to
approximations introduced by the reduced maximum frequency adopted with some peaks
potentially generated by aliasing effects. As explained in step 8 they can be eliminated by
increasing the "maximum frequency" parameter in the FW-H analysis input mask, at the
expenses of a longer simulation time.

112 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.51
Numerical predictions of the noise spectra at a polar microphone located along a 61 cm side
line. Flow incidence of +8 deg. "Max frequency" parameter in the FW-H analysis input mask
setup to 5269 Hz.

FIGURE 3.52 shows the result obtained with a "maximum frequency" 16 times higher. The
spurious peaks are completely eliminated. It is verified that the amplitudes of the main tones
are exactly the same as those present in FIGURE 3.48. Therefore in this case it is possible to
perform a simulation with a reduced "maximum frequency", preserving the accuracy of the
results. However this possibility shall be verified on each specific application through a
convergence analysis.

FIGURE 3.52
Numerical predictions of the noise spectra at a polar microphone located along a 61 cm side
line. Flow incidence of +8 deg. "Max frequency" parameter in the FW-H analysis input mask
setup to 84304 Hz.

FIGURE 3.53 shows the SPL variation of BPF1 along the polar angle. The predictions
obtained respectively with the solid-rotating surface and with the permeable-fixed surface are
in agreement. An underprediction of the experimental data of 10 dB is observed in the angular
range around 80-90 degs.

113 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.53
BPF1 tone directivity along a 61 cm side line. Flow AOA of +8 deg. Comparison of
numerical results vs NASA measurements. "Max frequency" parameter in the FW-H analysis
input mask setup to 5269 Hz.

FIGURE 3.54 shows the SPL variation of BPF2 along the polar angle. The prediction of the
angular position of the SPL peak obtained with the permeable-fixed FW-H surface is in line
with the experimental evidence, while an underprediction of the noise levels up to 7 dB is
globally observed. A slight deviation in the results obtained with the two FW-H approaches
exploited in this tutorial is observed.

114 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.54
BPF2 tone directivity along a 61 cm side line. Flow AOA of +8 deg. Comparison of
numerical results vs NASA measurements. "Max frequency" parameter in the FW-H analysis
input mask setup to 5269 Hz.

FIGURE 3.55 shows the SPL variation of BPF1+BPF2 with respect to the polar angle. A
global agreement is observed between the numerical predictions and the experimental data,
especially for the predictions obtained with the permeable-fixed FW-H surface.

115 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.55
BPF1+BPF2 tone directivity along a 61 cm side line. Flow AOA of +8 deg. Comparison of
numerical results vs NASA measurements. "Max frequency" parameter in the FW-H analysis
input mask setup to 5269 Hz.

FIGURE 3.56 shows the SPL variation of 2*BPF1+BPF2 with respect to the polar angle. A
global agreement is observed between the numerical predictions and the experimental data.

116 FINE™/Acoustics 8.1 Tutorials

 FIGURE 3.56
2*BPF1+BPF2 tone directivity along a 61 cm side line. Flow AOA of +8 deg. Comparison of
numerical results vs NASA measurements. "Max frequency" parameter in the FW-H analysis
input mask setup to 5269 Hz.

References

[1] Woodward, R. P. (1992). "Noise of Two High-Speed Model Counter-Rotation Propellers at

Takeoff/Approach Conditions", Journal of Aircraft, Vol. 29, No. 4.
[2] Ferrante, P., Vilmin S., Hirsch, C., Bonaccorsi J.C., di Francescantonio, P., (2013).
"Integrated "CFD - Acoustic" Computational Approach to the Simulation of a Contra Rotating
Open Rotor at Angle of Attack", AIAA 2013- 2242, 19th AIAA/CEAS Aeroacoustics
Conference, Berlin, Germany.
[3] Vilmin, S., Lorrain, E., Hirsch, Ch., Swoboda, M. (2006). "Unsteady flow modeling across
the rotor/stator interface using the nonlinear harmonic method", ASME Turbo Expo 2006,
GT2006-90210, Barcelona, Spain.
[4] He, L., and Ning, W. (2005). "Efficient Approach for Analysis of Unsteady Viscous Flows in
Turbomachines", AIAA journal, Vol. 36, No. 11, pp. 2005-2012.
[5] di Francescantonio, P. (1997). "A New Boundary Integral Formulation for the Prediction of
Sound Radiation", Journal of Sound and Vibration 202(4), 491-509.
[6] Envia, E. (2012). "Open Rotor Aeroacoustic Modeling", NASA/TM-2012-217740.

117 FINE™/Acoustics 8.1 Tutorials

 [7] Deconinck, T., Capron A., Hirsch, C., Ghorbaniasl G., (2011). "Prediction of Near- and Far-
Field Noise Generated by Contra-Rotating Open Rotors", International Journal of Aeroacoustics,
issue 2, vol. 11, pp. 219 - 238.

118 FINE™/Acoustics 8.1 Tutorials

TUTORIAL 4.

SIROCCO-TYPE FAN - BROADBAND NOISE

ANALYSIS

This tutorial describes the steps to perform in the Turbomachinery wizard in order to compute the
broadband noise generated by a sirocco-type fan combined with a volute.
The CFD results used for the stochastic reconstruction of the noise sources are contained in a CGNS
file provided in specific file enclosed with this tutorial. Further details about the required CFD input
can be found in the Input and output of the User Guide.
The fan analyzed in the frame of this tutorial consists of an impeller row rotating at 3000 RPM,
encompassed within a casing. The full 360 degree geometry (3.8M control cells) has been considered
for the steady CFD simulation run in FINE™/Open with OpenLabs™. An illustration of the case is
shown below, where the magnitude of the velocity is represented on a cutting plane at mid-height.

Prior to run the Turbomachinery wizard, make sure that the workstation considered for the acoustic analysis
complies with the minimum hardware requirements (see: Limitations). HEXPRESS™/Hybrid should also be
installed on the workstation (latest version supported: v7.1).

119 FINE™/Acoustics 8.1 Tutorials

 Start the turbomachinery wizard

Launch FINE™/Acoustics and start the Turbomachinery wizard from

Wizard>Turbomachinery in the main menu bar.

The window shown in the following picture appears soon after. Press Next to start the acoustic
setup.

120 FINE™/Acoustics 8.1 Tutorials

 Acoustic setup

1. After pressing the button Browse, select the directory in which all the files generated by the
wizard will be stored. If the directory does not already exist, a new one can be created by
using the button Create New Folder. Name this directory Sirocco_fan_broadband.

Press Next to continue.

2. Import the CFD data through the button Browse located in the section CFD Results.
Select the file sr_01.cgns in the FINE™/Acoustics tutorial directory provided in the package.
Select Centrifugal for the type of turbomachinery configuration (section Turbomachinery
Configuration ), and Broadband for the type of acoustic analysis to perform (section
Analysis Type).

121 FINE™/Acoustics 8.1 Tutorials

 See also in the User Guide: Step 2: Import the CFD results, select the type of turbomachinery
configuration considered and the type of analysis to perform.
Press Next to continue.

The tonal noise analysis can also be performed, but the analysis will require a longer time and
more computational resources. An example of tonal noise analysis is described in Tutorial 2.

3. Select the domains to consider for the acoustic analysis: keep "ROW1" (fan) and "ROW2"
(volute) selected.
In order to better distinguish the fan from the volute (the fan is fully encompassed within the
volute), press the button .

122 FINE™/Acoustics 8.1 Tutorials

 See also in the User Guide: Step 3: Select domains.

The button can be pressed to retrieve the default visualization mode.

Press Next to continue.
4. Remove the duct extension: keep the default values for each row and press Next to continue.
To switch from "ROW1" to "ROW2" and inversly, press the corresponding radio button.
See also in the User Guide: Step 4: Remove duct extension.

123 FINE™/Acoustics 8.1 Tutorials

 Press Next to continue.
5. Select the radiating surfaces: keep the inlet and outlet of ROW2 ("ROW2-Inlet" and "ROW2-
Outlet", respectively) selected, with the condition Far-field.

See also in the User Guide: Step 5: Define radiation surface.

Press Next to continue.
6. Define the smallest geometrical not to capture for the meshes generated during the propagation

124 FINE™/Acoustics 8.1 Tutorials

 analysis: specify a tip gap of <0.002> m.

See also in the User Guide: Step 6: Define geometrical tolerance for meshing.
Press Next to continue.
7. Define model parameters: keep the defaults values and the default checked option.

125 FINE™/Acoustics 8.1 Tutorials

 See also in the User Guide: Step 7: Define model parameters.
Press Next to continue.
8. Define the model settings: keep the default option selected.

126 FINE™/Acoustics 8.1 Tutorials

 See also in the User Guide: Step 8: Define HEXPRESS™/Hybrid path and Flow-Noise
parameters.
Press Next to continue.
9. Split domains into subdomains: keep the default values.

See also in the User Guide: Step 9: Split large domains in subdomains.
Press Next to continue.
10. Define domains with rotating parts: unselect "ROW2_ALL".

127 FINE™/Acoustics 8.1 Tutorials

 See also in the User Guide: Step 10: Define domains with rotating parts.
Press Next to continue.
11. Click on Next to start the analysis.

The computation requires approximately 1 day to complete on 8 processors (AMD Opteron

2350 - 2 GHz).
12. Perform the post-processing: keep all the options selected by default.

128 FINE™/Acoustics 8.1 Tutorials

 See also in the User Guide: Step 13: Start post-processing.
Press Next to continue.
13. Radiated Sound Power: click successively on PWL (narrow band) and on PWL (1/3 octave
bands) to generate the spectra expressed in terms of narrow band and 1/3 octave bands,
respectively. The figures displayed are automatically saved inside the directory containing the
files generated by the wizard.

129 FINE™/Acoustics 8.1 Tutorials

130 FINE™/Acoustics 8.1 Tutorials
 See also in the User Guide: Step 14a: Radiation Sound Pressure.
Press Back in order to go back to the initial post-processing window. Select the option press
Sound Pressure Level at microphones and press Next to access the window enabling to
define the position of the microphones.
14. Sound Pressure Level at microphones: tick the first 6 microphones in the left part of the
window and enter the position shown in the table below. Keep the option No ground
activated in the section "Define a reflective ground surface".
Click on the buttons All Mics (narrow band) and All Mics (1/3 octave bands) to generate
the sound pressure levels expressed in terms of narrow band and 1/3 octave bands at the
different microphones.

131 FINE™/Acoustics 8.1 Tutorials

132 FINE™/Acoustics 8.1 Tutorials
 Microphone ID X Y Z
1 -0.707 0 0.707
2 0 0 1.0
3 0.707 0 0.707
4 -0.835 0.781 0.052
5 -1.128 0.074 0.052
6 -0.835 0.634 0.052

The following two pictures illustrate the position of the microphones relatively to the radiating
surfaces.

133 FINE™/Acoustics 8.1 Tutorials

134 FINE™/Acoustics 8.1 Tutorials
 The following pictures represent the broadband noise observed at microphone 1 and 4.

135 FINE™/Acoustics 8.1 Tutorials

136 FINE™/Acoustics 8.1 Tutorials
137 FINE™/Acoustics 8.1 Tutorials
138 FINE™/Acoustics 8.1 Tutorials
 See also in the User Guide: Step 14b: Sound Pressure Level at microphones.
Press Next to continue.
15. Change Flownoise HMesh refinement: press Next to ignore the step.

139 FINE™/Acoustics 8.1 Tutorials

 See also in the User Guide: Step 15: Restart the analysis with other level of refinement for the
Flow-Noise mesh.
16. Restart adding Tonal Noise: press Next to ignore the step.

See also in the User Guide: Step 16: Restart adding tonal noise.
Press Next to continue.
17. Complete the analysis: click on Quit to exit the wizard.

140 FINE™/Acoustics 8.1 Tutorials

141 FINE™/Acoustics 8.1 Tutorials
 TUTORIAL 5.

NOZZLE

This tutorial provides an introduction to the usage of the FlowNoise module of FINE™/Acoustics for
the evaluation of the aerodynamic noise generated by a convergent single-stream nozzle operating in
subsonic cold flow condition. The starting point is a steady CFD RANS analysis that constitutes the
input to the FlowNoise module [1] [2] which uses a stochastic approach to synthesize the frequency
content of the noise sources due to the turbulence. The reconstructed noise sources are then propagated
to the far-field microphones with the Green's function method implemented in the Boundary Element
Method (BEM) module of FINE™/Acoustics, which is used to propagate in the near-field and in the
far-field the acoustics sources reconstructed with the Lamb's formulation [3] . A uniform mean flow is
considered.

5.1 INTRODUCTION

5.1.1 Prerequisites

The tutorial assumes that the user is familiar with the menu structure of FINE™/Acoustics.
Both the CFD mesh and the RANS analysis are considered already completed and available to
the user (see the file "nozzle_ 3D_ OUT_ [Link]" realized with FINE™/Turbo, provided
together with this tutorial). The user is therefore supposed to own a background in CFD for what
concerns the mesh generation process, the setup and the analysis of the RANS computational
model.

5.1.2 Problem description

In this tutorial a working description will be given for a typical static single-stream 86 mm (3.4
inch) convergent nozzle operating with subsonic jet velocity in cold flow conditions (FIGURE
5.1). The nozzle geometry is the same tested by QinetiQ in the anechoic chamber of the Noise
Test Facility (NTF) at QinetiQ Pyestock [4].-

142 FINE™/Acoustics 8.1 Tutorials

 FIGURE 5.1
Section through test nozzle

The FINE™/Acoustics calculation described in this tutorial aims at reproducing the radiated
broadband noise through the stochastic reconstruction of the turbulent fluctuating flow field
computed with the FlowNoise module. The noise predictions are achieved at specific
microphones located in the far-field. The operative procedure can be subdivided in two main
steps:
1. Acoustic source reconstruction (Flow-Noise analysis).
A CFD steady RANS simulation is performed with FINE™/Turbo on a structured grid
generated with IGG™. The first step is to identify over the CFD mesh the regions where the
aero-acoustic sources are active. For this purpose a group is created containing the CFD
elements characterized by Turbulent Kinetic Energy (TKE) higher than a threshold defined by
the user (FIGURE 5.2). The selection of the optimal value for the threshold should be based
on a sensitivity analysis. The goal is to include all the relevant kinetic energy responsible for
the generation of the acoustic sources, limiting as much as possible the CFD domain extension
in order to minimize the CPU time. In the selected CFD sub-domain the FlowNoise module
creates an acoustic Cartesian mesh on which nodes the turbulent velocity and the associated
acoustic sources are reconstructed.

143 FINE™/Acoustics 8.1 Tutorials

 FIGURE 5.2
Turbulent Kinetic Energy selection

2. Acoustic source radiation (Flow-Noise analysis).

The reconstructed acoustic sources are then radiated by means of the Green's function
approach [3]. The radiated field is computed in free-field conditions, in order to obtain the
"incident acoustic field" at the far-field microphones. In this calculation no solid surfaces (e.g.
the nozzle) and reflective/scattering effects are considered.

5.2 CFD - RANS SIMULATION

The geometry and the mesh have been constructed with IGG™. The external domain, modeling
the far field region, is made of a pure cylinder whose dimensions have been chosen large enough
to get true far field conditions. In particular, its length L is about 4.8 m while its diameter D is
approximately 3.4 m (FIGURE 5.3).

144 FINE™/Acoustics 8.1 Tutorials

 FIGURE 5.3
CFD computational domain

The meshing has then been performed paying particular attention to the physics of the problem.
Indeed, a very fine mesh has been created near the jet exit (FIGURE 5.4), where high gradients
could be expected, and the mesh expands when moving far away from the jet. Moreover, adding
a butterfly topology inside the jet (FIGURE 5.5) has considerably increased the quality of the
mesh. Information about the quality of the final mesh are summarized here below:
l Minimum orthogonality 24.18°
l Maximum aspect ratio: 4021.99
l Maximum expansion ratio: 1.76
l Three grid levels
l 2.2 million points

FIGURE 5.4
CFD computational mesh

145 FINE™/Acoustics 8.1 Tutorials

 FIGURE 5.5
Mesh topology on the nozzle axis

The computation has been finally carried out with FINE™/Turbo, imposing the jet's inlet mass
flow rate and the outlet pressure of the external domain (given from the QinetiQ database). The
flow conditions considered are subsonic and unheated jet, with fully expanded jet velocity up to
Mach 0.5 (FIGURE 5.6). The remaining walls belonging to the external domain have been
defined as Eulerian walls. Finally, the Shear Stress Transport model (SST) has been chosen as the
turbulence model.

146 FINE™/Acoustics 8.1 Tutorials

 FIGURE 5.6
Mach number distribution

5.3 FLOW-NOISE ANALYSIS

Here below is reported the list of the computational steps required to reconstruct the acoustic
sources with the Lamb's formulation and to compute the acoustic incident field over the far-field
microphones.
1. The first step is to import in FINE™/Acoustics the CFD mesh and CFD solutions.
l Select the menu Discretization > Import New > File and select the CGNS generic file
(*.cgns) file format. Import the file "nozzle_3D_OUT_000.cgns" provided in this tutorial.
l In the import mask select FEM to import volume elements ("BEM" imports only surface
elements), select all options in Global Data to be imported to import also the CFD
solution, don't extract boundary surfaces from volumes. Click the button OK to confirm.

147 FINE™/Acoustics 8.1 Tutorials

 l Through a right click on the mask activate only the domain blocks relative to the
"project3d" base.

148 FINE™/Acoustics 8.1 Tutorials

 l The CFD mesh and solution are now imported in FINE™/Acoustics. With the menu
Analysis > Mesh info it is possible to check the domain and the mesh size.

l Rename the newly imported discretization as <Flow-Noise model> applying the command
Discretization > Rename..
2. After the import save the project (File > Save). It is suggested to perform this operation
regularly as there is no "undo".

149 FINE™/Acoustics 8.1 Tutorials

 3. It is also suggested to delete the unnecessary CFD data, keeping only static pressure, density,
velocity, turbulent kinetic energy and dissipation rate. For this purpose apply the command
Boundary C. > Delete, activate the fields indicated below and press OK.

4. Visualize the CFD solution.

150 FINE™/Acoustics 8.1 Tutorials

 l With the buttons highlighted below, it is possible to visualize the CFD solution on a
clipping plane located on the symmetry axis of the nozzle.

l Select field in order to assess its distribution within the computational domain. To adjust the
displayed levels right click on the screen and apply the command Data visualization >
Global parameters. Switch off the clipped visualization

151 FINE™/Acoustics 8.1 Tutorials

 5. Create a source group "TKE>50":
l Keep Turbulent Kinetic Energy (TKE) selected and with the command Groups > Create
from > Current Data Values generate the group "TKE>50" which contains the CFD
nodes and elements in which the TKE level is bigger than the threshold imposed.

6. Include the far field microphones indicated in FIGURE 5.7. For this purpose create a new
group called "mics" (with command Groups > New), enable the visualization of this group,
add the microphones with the command highlighted in FIGURE 5.8. The microphone angles
reported in FIGURE 5.7 are referred to the nozzle axis, the distances are referred to the center
of the nozzle exit plane.

152 FINE™/Acoustics 8.1 Tutorials

 FIGURE 5.7
Microphones coordinates

FIGURE 5.8
Microphones insertion

7. Run the FlowNoise analysis by applying the command Analysis > Flow-Noise. The control
window displayed FIGURE 5.9 permits to define the input parameters to be used in the Flow-
Noise calculation. The data to be defined are the following:
l Name: name assigned to the analysis.
l Source group: group that identifies the CFD region in which the aero-acoustic sources will
be reconstructed.
l Microphones: group that contains the nodes and the panels of the acoustic mesh on which
the pressure radiated by the acoustic sources will be computed.
l Pressure: CFD solution that contains the static pressure.
l Density: CFD solution that contains the fluid density.
l Vx: CFD solution that contains the X component of the fluid velocity.
l Vy: CFD solution that contains the Y component of the fluid velocity.
l Vz: CFD solution that contains the Z component of the fluid velocity.
l Turbulent Kinetic Energy: CFD solution that contains the fluid turbulent kinetic energy
(K).
l Dissipation: CFD solution that contains the turbulence dissipation (epsilon or omega).
l Dissipation given in input: it can be epsilon or omega, dependently on the output provided
by the CFD analysis.
l Fluid properties: speed of sound and fluid density in the far-field.
l Analysis: the user shall define the minimum and the maximum frequency. The minimum
frequency corresponds to the frequency step.
Fill the menu as show below:

153 FINE™/Acoustics 8.1 Tutorials

 FIGURE 5.9
FlowNoise analysis setup

8. A new project is created with name "Flow-Noise analysis - Mesh refinement=1", containing
the HMesh.

154 FINE™/Acoustics 8.1 Tutorials

 9. Perform a second Flow-Noise analysis using the same setup of the previous run, but with a
mesh refinement equal to 2. The resulting mesh is shown below.

If an error message appears right after launching the propagation simulation (detailed in the next
section), make sure to relaunch flow noise computations with enough differences in the name
given to each of them, so that the solver can retrieve correctly the flow noise sources; the solver
uses indeed pattern recognition to retrieve the flow noise sources files. Avoid also to use any
underscore, capital or other special characters in the name given in the analysis.

10. It is now time to select the project containing the mesh with refinement=1 and to launch the
Flow-Noise analysis. The original analysis setup can be retained, except for the Analysis
Mode that now shall be put on Run on current mesh. The Flow-Noise analysis can be
started and the noise radiation is computed to the far-field microphones. Once the analysis is
completed it is possible to proceed with the second run on the HMesh with refinement=2.

5.4 POST-PROCESSING

Once the two Flow- Noise analysis are completed, it is possible to post- process inside
FINE™/Acoustics the results obtained. Here below reports the list of the computational steps
required.
11. Select the desired project (e.g. Flow-Noise analysis - Mesh refinement=1).

155 FINE™/Acoustics 8.1 Tutorials

 l In order to obtain the PSD spectra at the microphones, select Post-processing > New 2D
window,
l Right click in the displayed white window and put the Metrics on dB RMS (SPL).

l Then right-click and select 2D Data definition / Solutions. Select model, solution, type of
field to be visualized and activate or deactivate the frequencies desired in the final plot. For
our purpose we shall select "pressure" and let all the frequency fields activated.

156 FINE™/Acoustics 8.1 Tutorials

 l Select then the group "mics".
l In the following input mask activate the microphones and select Add.

l The final PSD spectra will be displayed. More than one spectrum can be represented on the
same plot (by applying iteratively the commands sequence described above). the following
figure shows the spectra computed in the FlowNoise analysis at the far-field microphones
(mesh refinement=1).

157 FINE™/Acoustics 8.1 Tutorials

 Right click on the plot obtained and select One third octave > PSD > Linear and save the
results in a file called <PSD_one_third_octave.txt>. With the previous command the spectra
are transformed into the 1/3 octave band metrics. "A-weighted" is used when it is necessary to
account for the human hear sensitivity, which changes with the frequency. The filter selected is
the "PSD" one, in order to be consistent with the physical units of the output provided by
FlowNoise. The generated output file can be used (externally to FINE™/Acoustics) to obtain
the plots contained in FIGURE 5.10 - FIGURE 5.11, showing the comparison between the
numerical results obtained (for the two mesh refinements considered) and the experimental
data.

158 FINE™/Acoustics 8.1 Tutorials

 FIGURE 5.10
PSD spectra (predictions vs measurement for microphone located at 30°)

159 FINE™/Acoustics 8.1 Tutorials

 FIGURE 5.11
PSD spectra (predictions vs measurement for microphone located at 90°)

5.5 ACKNOWLEDGMENT

QinetiQ is acknowledged for providing the test case considered in this tutorial and the relative
database.

5.6 REFERENCES

[1] Caradonna, J., Schwartz, O., di Francescantonio, P. (2010), "Innovative Computational Aero-
Acoustic Approach for Automotive Exhaust Systems", ISMA International Conference on Noise
and Vibration Engineering, Leuven, Belgium.
[2] di Francescantonio, P., Ferrante, P., Deconinck, T., Hirsch, C. (2013), "Assessment of SNGR
Method for Robust and Efficient Simulations of Flow Generated Noise", AIAA 2013-2264, 19th
AIAA/CEAS Aeroacoustics Conference, Berlin, Germany.
[3] "FineAcoustics Theoretical Manual", Release 5.2.1, August 2013.
[4] "Results from an experimental programme on static single- stream jet noise",
QinetiQ/FST/TN026455.
[5] Ewert, R., Appel, C., Dierke, J., Herr, M., (2009), "RANS/CAA based prediction of NACA
0012 broadband trailing edge noise and experimental validation", AIAA 2009- 3269, 15 th
AIAA/CEAS Aeroacoustics Conference, Miami, Florida.

160 FINE™/Acoustics 8.1 Tutorials

TUTORIAL 6.

AIRFOIL WITH SLAT

This tutorial provides an introduction to the usage of the Flow-Noise module of FINE™/Acoustics for
the evaluation of the aerodynamic noise generated by an airfoil with slat deployed.
This is a typical configuration generating airframe noise at take-off and landing conditions on aircrafts
( FIGURE 6.1 ). The starting point is a steady CFD RANS analysis that constitutes the input to the
Flow-Noise module ( [1] , [2] ) which uses a stochastic approach to synthesize the frequency content of
the noise sources due to the turbulence. The reconstructed noise sources are then propagated to the far-
field microphones with two different alternative approaches:
1. Boundary Element Method (BEM)
The BEM method can be applied when the flow non-uniformities are limited (typically for flows
with maximum Mach < 0.3) while the FEM approach provides more accuracy in presence of
relevant mean flow gradients.

FIGURE 6.1
Airframe noise generation from high-lift devices

The Green's function implemented in the BEM module of FINE™/Acoustics, which is used to
propagate in the near-field and in the far-field the acoustics sources reconstructed with the Lamb's
formulation [3].

161 FINE™/Acoustics 8.1 Tutorials

 A uniform mean flow is considered. The scattering effect due to solid surfaces present in the
computational domain is taken into account by a specific BEM analysis, which radiates the reflected
acoustic field.
2. Finite Element Method (FEM)
FEM is used to solve the Pierce- Howe acoustic wave equation ( [3] , [4] ), with embedded Flow-
Noise sources, in the airfoil near-field. This approach accounts for the convection and the refraction
effects on the noise propagation due to the mean-flow non uniformities in the vicinity of the airfoil.
The sound radiation to the far-field is afterwards computed with a BEM simulation.

6.1 INTRODUCTION

6.1.1 Prerequisites

The tutorial assumes that the user is familiar with the menu structure of FINE™/Acoustics.
Both the CFD mesh and the RANS analysis are considered already completed and available to
the user (see the project FineHexa_k-epsilon realized with FINE™/Open, provided together with
this tutorial). The user is therefore supposed to own a background in CFD for what concerns the
mesh generation process, the setup and the analysis of the RANS computational model. A demo
case on the generation of an unstructured mesh for a three elements airfoil can be found in the
HEXPRESS™ documentation.
The acoustic mesh necessary to compute the sound propagation with the FEM solver
implemented in FINE™/Acoustics has been generated with HEXPRESS™/Hybrid. This mesh is
provided in the file volume_mesh.bdf.

6.1.2 Problem description

In this tutorial a working description will be given for a typical flow noise analysis using as a test
case the "airfoil with slat" configuration studied in the VALIANT research project (FIGURE
6.2).
This airfoil configuration presents a flow in the slat cove area which is representative of the flow
in a realistic three-element configuration, minimizing significantly the global flow deviation and
the global lift force in order to allow open-jet windtunnel tests. The flow condition are Minf=0.15
and angle of attack = 18°. The airfoil chord length is 300 mm, the spanwise extension is 300 mm.

162 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.2
Left side: VALIANT "airfoil with slat" geometry (2D). Right side: experimental bench operated
into an anechoic wind-tunnel.

6.2 CFD - RANS SIMULATION

A 2D RANS analysis is initially performed with FINE™/Open to obtain the time-averaged flow
field. The k-omega SST turbulence model is considered. The simulation is performed over a
structured mesh realized with IGG™ made by 413k nodes and 205k cells (FIGURE 6.3).

FIGURE 6.3
RANS computational mesh

FIGURE 6.4 provides a sample of the computed flow field in terms of static pressure and velocity
magnitude distributions.

163 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.4
Sample of RANS results

6.3 ACOUSTICS ANALYSIS

6.3.1 Flow-Noise analysis

Here below it is reported the list of the computational steps required to reconstruct the acoustic
sources with the Lamb's formulation and compute the acoustic incident field over the airfoil
surface. This field will be applied as boundary condition in the subsequent BEM radiation
analsysis.
1. The first step is to import in FINE™/Acoustics the CFD mesh. For this purpose it is necessary
to click on Discretization > Import New > File and select the FINE/Open mesh (*.hex) file
format.(FIGURE 6.5)
Import the file "FineHexa_k-[Link]" provided in this tutorial.

164 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.5
CFD mesh selection

2. In the import mask select FEM to import volume elements ("BEM" imports only surface
elements) as shown in FIGURE 6.6:
l Unselect Global data
l Resize the mesh to 1/1000 in order to transform the coordinates from mm to m

165 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.6
CFD mesh import

l Select the refinement level 0 to import the mesh with the highest resolution (FIGURE 6.7)

166 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.7
Selection of the refinement level

3. After the import save the project (File > Save); it is suggested to perform this operation
regularly as there is no "undo".
4. Rename the discretization just created as < Flow-Noise analysis > applying the command
Discretization > Rename.
5. With the command Analysis > Mesh Info, visualize the model composition (e.g. number of
nodes, elements, …) as well as the mesh size and quality (FIGURE 6.8).

167 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.8
Mesh information

In particular it is indicated the maximum frequency at which the mesh can be used for acoustic
propagation (Figure 13). This value is by default calculated considering a minimum mesh density
of 6 points per wavelength. Check this value only for acoustic meshes used to compute the sound
propagation.

6. Define the reference system as 2D. For this purpose apply the command Mesh > Set
dimensions(FIGURE 6.9):

FIGURE 6.9
Setting of the model dimensions

7. Import the RANS solution as boundary conditions:

l Boundary C. > Import
l Choose the file format FINE/Open results (*.cfv) (*.run)(FIGURE 6.10)
l Import the file "FineHexa_k-[Link]", and keep default options selected in the mask
appearing.

168 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.10
CFD result selection

The simple Import is used when the "destination" mesh (on which the results are imported) is the
same as the "source" one, provided that the node labels are the same. The Import Adv. is used
when an interpolation is required.

8. It is suggested to delete the unnecessary CFD data (FIGURE 6.11):

l Boundary C. > Delete
l Keeping only Static pressure, Density, Vx, Vy, Vz,k and epsilon
l Activate the fields to be deleted and press OK

169 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.11
Removal of unused data

9. Create the group of acoustic source:

l Visualize the Turbulent Kinetic Energy (TKE) field.
l Go to Options > Filters > Positions, select the option node position within the specified
box, enter the following values: XMin=-0.12, XMax=0.5, YMin=-0.12, YMax=0.1. Click
on Ok to confirm.

This step is optional. It is used to select a certain region so that the following analysis can be
reproduced.

l With <F3> + mouse, select a source region that contains the relevant TKE (the selected
region should be larger than the defined position in previous step).
l Create a group called Source(FIGURE 6.12 and FIGURE 6.13)
l Go to Options > Filters > Positions , uncheck the option node position within the
specified box.

170 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.12
Creation of "Source" group based on Turbulent Kinetic Energy

FIGURE 6.13
"Source" group (Turbulent Kinetic Energy field)

171 FINE™/Acoustics 8.1 Tutorials

 10. Create the group of solid surfaces of the airfoil:
l Select with <F3> + mouse the CFD elements (with nodes included) containing the airfoil
l Operate Extract Skin in the poped-up menu (FIGURE 6.14).
l A group "Skin 2D" is created containing a mesh made by the free surfaces (skin) of the
CFD mesh selected.

FIGURE 6.14
Creation of "Skin 2D" group

l Visualize the group "Skin 2D"

l Select only the airfoil mesh with <F3> + mouse
l Create a group called <Airfoil>
To remove the panels not needed any more (FIGURE 6.15):
l Groups > Transform/Delete and select Make Complement Group for Operation,
l Make complement group (named <Free Boundaries>) of group "Airfoil",

172 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.15
Make Complement Group

l Select Skin 2D in the next dialog window,

l Visualize the group Free Boundaries, select all panels with <F3> + mouse and delete,
l Delete both groups Skin 2D and Free Boundaries.
11. Check the surface normal. For both the Flow-Noise and the BEM analysis, it is necessary that
the normals point inside the computational domain.

12. Create a new empty group called <Mic> and visualize it. A far-field microphone can be added
with the command highlighted in the following figure. The microphone coordinates are: x =
0.7556 m, y = -1.7114 m, z = 0 m.

173 FINE™/Acoustics 8.1 Tutorials

 13. With the command Groups > Tranform/Delete > Join groups create the group
"Airfoil+Mic".

14. It is now possible to start the Flow-Noise analysis; the acoustic sources are reconstructed in the
source region selected.
Run the Flow-Noise analysis by applying the command Analysis > Flow-Noise. The control
window displayed permits to define the input parameters to be used in the Flow- Noise
calculation. The data to be defined for the BEM analysis are the following (FIGURE 6.16):

174 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.16
Flow-Noise analysis settings

Once the analysis is started a new project is created with name "Flow Noise Analysis - mesh
ref =1", containing the HMesh (FIGURE 6.17).

175 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.17
HMesh (refinement=1, 21k nodes)

15. It is now possible to perform a second Flow-Noise analysis using the same setup of the
previous run, except that:
l Name is set to Flow Noise Analysis - mesh ref = 2,
l Mesh refinement is set to 2,
l Enable the option Export sources for embedded FEM,

The option Export sources for embedded FEM must be also checked to store the Flow
Noise sources on the hard disk for the subsequent propagation analysis with the FEM.

l Enable the option Enable 2D/3D correction and define the parameters: Mach: <0.15>;
Mic distance : < 1.8 > m; 3D Span : < 0.3 > m; C Constant : < 2.1 > (experimentally
determined).

For the FEM analysis, the user should use the same settings with the option Enable 2D/3D
correction checked, in order to take into account the impact of the experimental
configuration.

176 FINE™/Acoustics 8.1 Tutorials

177 FINE™/Acoustics 8.1 Tutorials
 The relative mesh obtained is shown in FIGURE 6.18

FIGURE 6.18
HMesh (refinement=2, 78k nodes)

16. To run the computations, select the desired discretization generated in the previous steps, and
change the option in Analysis Mode to Run on Current mesh. Click finally on Start to run
the calculation. In this tutorial, the computation with mesh refinement =1 will be used in the
BEM analysis and the computation with mesh refinement =2 will be used in the FEM analysis

It has to be reminded that in any BEM and FEM analysis, based on a Flow- Noise input, the
acoustic pressure is expressed at each frequency in terms of sqrt(PSD).

178 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.19
Example of Flow-Noise solution (incident sound field [dB rms] at 1000 Hz over the airfoil
surface.)

17. The solution obtained can be finally exported in VNX format for the subsequent use in the
propagation BEM model. Apply the command Solution > Export > VNoiseX and export all
quantities obtained on the group "Mic+Airfoil" (FIGURE 6.20). Name the output file Flow-
Noise_solution-Mesh_ref1.vnx.

179 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.20
Flow-Noise Solution export

The analysis of the results available in project "Flow Noise analysis - Mesh ref =2" can be
performed in the same way.

6.3.2 BEM approach (uniform mean flow)

A. BEM setup

It is now possible to start the preparation of the BEM radiation model. Here below it is
reported the list of the computational steps required.
18. The airfoil mesh and the Flow-Noise solution previously obtained can be imported in a new
discretization.
l Discretization > Import new > File,
l Select the file "Flow-Noise_solution-Mesh_ref1.vnx" previously saved,
l Set Mesh Type to BEM,
l uncheck the option Extract Boundary surfaces from volumes.

180 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.21
Import of airfoil mesh and Flow-Noise results into the BEM model

181 FINE™/Acoustics 8.1 Tutorials

 19. Definition of discretization and groups:
l Rename the new discretization as <BEM radiation analysis - Mesh ref=1>
l Set the dimension of the discretization to 2D in Mesh>Set Dimensions.
l Create a group called "Airfoil" containing nodes and elements of the airfoil and the slat.
l Select the microphone and with it create a group called "Mic".
20. With Boundary C. > Domains > New create an external domain called <Air> with the
settings shown in FIGURE 6.22.

182 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.22
Domain definition

Also the domain and the boundary conditions in Boundary C. > Define shall be defined as
shown in FIGURE 6.23 and FIGURE 6.24.

FIGURE 6.23
Domain attribution

FIGURE 6.24
Boundary conditions definition

21. Start the BEM radiation analysis with the command Analysis > Freq. response(FIGURE
6.25 ). In particular it is possible to activate/deactivate part of the boundary conditions
displayed through a right click + mouse. Click on Start and accept the warning relative to the
Aspect Ratio bigger than 10.

183 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.25
Frequency response analysis

The BEM simulation above described takes approximately 4 GB of memory. This method
requires large amount of memory. In case of very fine mesh, the FEM analysis should be
privileged.

184 FINE™/Acoustics 8.1 Tutorials

 B. BEM Post-processing

22. Here below it is reported the list of the computational steps required.
l Post-Processing > Metrics
l Select dB rms (SPL) (FIGURE 6.26)
l Check the pressure distribution on the airfoil surface (FIGURE 6.27)

FIGURE 6.26
Metrics selection

185 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.27
Power spectral density [dB] distribution over the airfoil surface at 1000 Hz

23. To obtain the PSD spectra at the microphone considered in the simulation:
l Select Post-processing > New 2D window, right click in the white window displayed and
put the Metrics on dB rms (SPL).
l Then right click and select 2D Data definition.
l Select Model, Solution , Type of field to be visualized and activate or deactivate the
frequencies desired in the final plot. (For our purpose we shall select "pressure" and let all
the frequency fields activated. )
l Select then the group "Mic". In the following input mask activate the microphone and
select Add (FIGURE 6.29). The final PSD spectra will be displayed.

186 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.28
Data definition

FIGURE 6.29
Microphone activation

187 FINE™/Acoustics 8.1 Tutorials

 More than one spectrum can be represented on the same plot (by applying iteratively the
commands sequence above described). In particular FIGURE 6.30 shows the spectra
computed in the Flow-Noise analysis and in the BEM analysis at the far-field microphone
included in the analysis.
The green spectrum is associated to the "incident" acoustic field (computed in project Flow
noise analysis – Mesh ref=1) and the red one to the "reflected" acoustic field.
The "total" PSD, sum of the PSDs due to the incident and to the reflected acoustic fields, can
be obtained by right click and applying the command 2D Composition, filling in the input
mask (FIGURE 6.31). The total PSD is shown in FIGURE 6.32.

FIGURE 6.30
Power Spectral Density spectra at the far-field microphone. Incident acoustic field (red),
reflected acoustic field (green).

188 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.31
Sum of incident and reflected acoustic fields.

FIGURE 6.32
Power Spectral Density spectra at the far-field microphone. Total PSD, sum of the PSDs due
to the incident and to the reflected acoustic fields.

Right click on the plot obtained, select Curve export > ASCII file and save the results in a
file called <Total_PSD_dB_microphone-Mesh_ref1.txt>. The generated output file can be
used to obtain plots externally to FINE™/Acoustics.

189 FINE™/Acoustics 8.1 Tutorials

6.3.3 FEM approach (non-uniform mean flow)

A. FEM setup

In the following paragraphs the procedure for the FEM simulation with embedded sources will
be outlined. The sources previously computed on the fine HMesh (refinement=2) will be
considered. The reader shall then ensure to have performed the Flow-Noise simulation "Flow
Noise analysis – Mesh ref = 2".

The acoustic sources necessary to run the FEM propagation analysis have been already calculated
and stored in the memory by selecting the option Export sources for embedded FEM in the
Flow-Noise control mask ( FIGURE 6.16).
The sources are stored in the temporary directories relative to each cluster client (C:\NUMECA_
SOFTWARE \FineAcoustics81\CLUSTER\ ClusterClient_X\FLOWNOISE_SOURCES).
This is done to avoid the handling of the huge files containing the computed sources. At the end
of the Flow- Noise analysis it is sufficient to save the project. FINE™/Acoustics can be even
closed: the sources will remain available in the temporary files. These files will be automatically
read by the FEM solver when started. Two reminders have to be considered:
l Between the storage of the Flow- Noise sources and the subsequent FEM analysis, no other
Flow- Noise analysis shall be executed in order not to overwrite and lose the information
contained in the temporary files.
l The FEM analysis shall be performed on the same PC, using the same number of cores as used
in the Flow-Noise simulation.

24. An acoustic mesh made by one layer of volume elements has to be prepared with a grid
generator external to FINE™/Acoustics. For this purpose, HEXPRESS™/Hybrid has been
used.
l With the command Discretization > Import new > File, select the file "FEM_mesh.vnx".
l Setup the import command window as indicated in FIGURE 6.33.

190 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.33
Acoustic mesh import mask

l Rename the discretization as <FEM propagation analysis>

l The mesh information can be obtained with the command Analysis > Mesh info(FIGURE
6.34). The maximum frequency achievable with this mesh is 14734.6 Hz.

191 FINE™/Acoustics 8.1 Tutorials

 l Set the model dimensions as 2D.

FIGURE 6.34
Mesh information

25. It is then necessary to select with <F3> + mouse all the FEM elements and nodes, then with
the operation Extract Skin the surface elements (panels) are extracted from the volumetric
FEM mesh and automatically added to a new group called "Skin 2D".
l Check its normals orientation which shall point inward, inside the FEM mesh. If not, select
nodes and elements of the group "Skin 2D" and apply the command Mesh > With
selection > Invert normal.
l Now select each side of the airfoil and the slat by pressing <F4> and passing the mouse of
it; when a white point appears close to head of the mouse pointer, right click and apply the
command Select connected > Sharp angles once all the sides are selected, right click over
one side and choose Groups> Create. Name this group "Airfoil". Repeat the operation
with the sides of the rectangle, and create the group "PML".

192 FINE™/Acoustics 8.1 Tutorials

 l Go back to the Flow Noise model discretization "Flow Noise analysis", and export the full
model by clicking on File > Export > VNoiseX > Full Model. Give the name "Mean_
flow" to the exported file.
l The same operation should be repeated for the microphone. On Flow Noise model
discretization, select the microphone by maintaining < F3 > pressed and capturing the
microphone with the mouse. Select None in the drop down menu. As before, the group is
exported through File>Export>VNoiseX>Selection . The name ‘Microphone" can be
given to the exported file.
26. It is now necessary to import the RANS mean flow field and interpolate it onto the acoustic
mesh.
In the discretization FEM propagation analysis import the mean flow field and interpolate it
onto the acoustic mesh with the command Boundary C. > Import Adv. > File(FIGURE
6.35 ), selecting the file "Mean_ [Link]". Import on group "Global", Set Data Prefix to
Mean_flow.
With command Boundary C. > Delete remove all the CFD fields except Static pressure,
Density, Vx, Vy, Vz. Accept the compression of the files (if proposed), useful to save space
on disk.

193 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.35
Import and projection of the RANS mean flow field over the acoustic mesh

194 FINE™/Acoustics 8.1 Tutorials

 27. Now it is necessary to remove the boundary layer, in order to comply with the acoustic wave
equation [4] solved by the FEM solver:
To remove the boundary layer, select the group "Airfoil" and apply the command Boundary
C. > CFD > Remove boundary layer; fill in the input mask as indicated in FIGURE 6.36.
The boundary layer is eliminated and the flow-field outside the boundary layer extrapolated to
the solid wall. This operation is not necessary when potential flow fields (without boundary
layer) are imported. New fields with prefix BL_removed are created. The final result is shown
in FIGURE 6.37.

FIGURE 6.36
Boundary layer removal

195 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.37
Boundary layer before (top) and after (bottom) removal

28. It is now possible to create the domain with the command Boundary C. > Domains > New
(FIGURE 6.38). The domain is made by air defined through a non-uniform mean flow field
corresponding to the one with boundary layer removed. Complete the window appearing
when clicking on Flow Definition, with the options shown in (FIGURE 6.38)
By selecting Import sources the acoustic sources previously computed by Flow-Noise, stored
in temporary directors, will be automatically imported by the FEM solver, interpolated over the
acoustic mesh and propagated inside the computational domain.

196 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.38
Non uniform flow field definition

29. The microphone exported during the step 27 needs to be imported in the discretization
considered for the propagation study. Select "[Link]" in File>Import>File .... In
the window appearing, check the option Create Group and call this group <Microphone>.
30. The last step required is the definition of the boundary conditions. A PML boundary condition
should be created following the parameters and values shown in (FIGURE 6.39) . To access
this window, click on Boundary C. > PML > New.

197 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.39
Definition of the PML boundary conditions

Before launching the calculation, make sure that the correct boundary conditions have been set
for each group; the different tabs available after clicking on Boundary C.>Define, should as
depicted in FIGURE 6.40

198 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.40
Boundary conditions and Domains settings

31. It is finally possible to run the FEM solver with the command Analysis > Freq. Response
(FIGURE 6.41). Define a frequency range covered by the previous Flow-Noise analysis (e.g.
2000-4000 Hz with frequency step of 50 Hz). This range is sufficiently extended for a
preliminary and fast comparison with the BEM radiation.
Once the analysis is started a new model is created with name "FEM analysis".

FIGURE 6.41
FEM frequency response analysis

199 FINE™/Acoustics 8.1 Tutorials

 32. To start the computation, select the new model "FEM analysis", click on Start, Accept the
warning message on the elements aspect ratio. In the Cluster server dialog box, click on the
button FlowNoise Source Selection and select the exported source in the computation "Flow
Noise analysis - Mesh ref =2".

B. FEM Post-processing

33. Once the analysis is completed it is possible to inspect the solution.

200 FINE™/Acoustics 8.1 Tutorials

 FIGURE 6.42
Acoustic field at 2000 Hz (real part of the acoustic pressure amplitude)

34. Apply the same commands described in step 22 to obtain a plot of the PSD at the far-field
microphone. The result obtained is already the total pressure, as in the FEM approach the
acoustic field is not split into an incident and into a reflected field as done in the BEM
approach.

201 FINE™/Acoustics 8.1 Tutorials

 35. The plot below shows the comparison between the BEM/FEM results and the VALIANT
experimental data.

FIGURE 6.43
Power Spectral Density spectra at the far-field microphone.

For the BEM result, an offset of -20.8dB should be manually applied in order to account for the
effective 3D span-wise extension. And an additional offset of +7 dB should be applied in order to
account for the installation effects active during the test made in the VALIANT project, due to the
presence of the airfoil end-plates.
For the FEM result, the effective 3D span-wise extension has been accounted for during the flow
noise simulation. Only an additional offset of +7 dB should be applied.

6.4 ACKNOWLEDGMENT

The provision of the CFD experimental data (Ecole Centrale de Lyon, ONERA) and the EC
support (Collaborative Project VALIANT, Grant Agreement N° 233680) are acknowledged.

202 FINE™/Acoustics 8.1 Tutorials

6.5 REFERENCES

[1] Caradonna, J., Schwartz, O., di Francescantonio, P. (2010), "Innovative Computational

Aero-Acoustic Approach for Automotive Exhaust Systems", ISMA International Conference
on Noise and Vibration Engineering, Leuven, Belgium.
[2] di Francescantonio, P., Ferrante, P., Deconinck, T., Hirsch, C. (2013), "Assessment of
SNGR Method for Robust and Efficient Simulations of Flow Generated Noise", AIAA 2013-
2264, 19th AIAA/CEAS Aeroacoustics Conference, Berlin, Germany.
[3] "FineAcoustics Theoretical Manual", Release 5.2.3, August 2014.
[4] Pierce, A., D., (1990), "Wave equation for sound in fluids with unsteady inhomogeneous
flow", J. Acoust. Soc. Am., Vol. 87, No 6, 2292-2299.
[5] De Roeck, W., (2007), "HYBRID METHODOLOGIES FOR THE
COMPUTATIONAL AEROACOUSTIC ANALYSIS OF CONFINED, SUBSONIC
FLOWS", PhD thesis, KATHOLIEKE UNIVERSITEIT LEUVEN, Leuven, Belgium.
[6] Ewert, R., Appel, C., Dierke, J., Herr, M., (2009), "RANS/CAA based prediction of
NACA 0012 broadband trailing edge noise and experimental validation", AIAA 2009-3269,
15th AIAA/CEAS Aeroacoustics Conference, Miami, Florida.

203 FINE™/Acoustics 8.1 Tutorials

 TUTORIAL 7.

MUFFLER

This tutorial explains how to predict the Transmission Loss (TL) of a typical muffler. The tutorial is
divided into two parts. The first part explains how to predict the transmission loss by using the muffler
wizard. The second part explains how to manually setup a transmission loss simulation.

7.1 INTRODUCTION

7.1.1 Prerequisites

The goal of this tutorial is to guide the user through the computational steps required in
FINE™/Acoustics for the characterization of the Transmission Loss (TL) of a typical muffler.
The tutorial assumes that the user is familiar with the menu structure of FINE™/Acoustics. Some
steps in the setup and solution procedure will thus not be shown explicitly.

7.1.2 Problem description

Mufflers are installed within the exhaust systems of most internal combustion engines. They are
engineered as acoustic soundproofing devices designed to reduce the loudness of the sound
pressure generated by the engine.
The typical noise sources in automotive engines are:
l The engine order noise generated by the pulsating gas pressure variations produced by the
opening and closing of the exhaust valves. This noise is characterized by a dominant firing
frequency and relative harmonics (tonal noise).
l The flow noise produced by the turbulence generated inside the silencer and at the outlet when
the flow expands into the atmosphere. It is characterized by a broadband spectrum.
There are two types of mufflers:

204 FINE™/Acoustics 8.1 Tutorials

 1. Reactive mufflers creates destructive interference between the sound waves or reflects the
noise back towards the source. This kind of mufflers mainly contain:
l Resonating chambers: resonant cavities at the termination of exhaust pipes.
l Expansion chambers: sudden change in cross sectional area in the pipe.

2. Dissipative mufflers dissipates the acoustic energy into heat. This kind of mufflers mainly
contain:
l Damping material: foam, fibrous material, glass fiber.
l Perforated sheets: metallic sheet punched (hole size of mm), supporting the damping
material.

A combination of the two muffler types is possible.

One of the acoustic performance parameters of a muffler is The Transmission Loss (TL). This is
the only performance parameter independent on the type of noise source and the elements located
around the muffler. It is defined as the difference between the incident power and the transmitted
power.

205 FINE™/Acoustics 8.1 Tutorials

 In this tutorial the Transmission Loss of a typical muffler will be computed. The muffler geometry
is shown below.

The gray and the yellow surfaces are rigid surfaces and the green one is a perforated surface. The
volume inside the yellow surface is filled with sound absorbing material. The red arrows indicate
the flow path.

7.1.3 Preparation

Download the tutorial archive as available from the Products page of the Customer Area
([Link]). Then extract the tutorial input files.

206 FINE™/Acoustics 8.1 Tutorials

7.2 SIMULATE SOUND ABSORBING MATERIAL

When dealing with sound absorbing materials, two different approaches are available in
FINE™/Acoustics: the surface impedance approach and the multi-domain approach.
The surface impedance approach can be used when the region to be analyzed is covered with a
relatively thin layer of sound absorbing material, in such a way that the absorbing material can be
considered locally reacting. In this case the user is simply required to define the admittance
(inverse of impedance) of the absorbing material, and so the modeling is substantially unchanged
in respect of the case without sound absorbing material. The main difficulty however is to know
the required impedance data, that typically can be obtained from experiments, analytical/semi-
empirical models or numerical simulation (e.g. Virtual Impedance Tube approach in
FINE™/Acoustics).
The multi-domain approach can be used in any cases, not only when the absorbing material is
thin. In this case the absorbing material has to be modeled as a separate domain with its own fluid
properties (that depend on the absorbing properties of the material itself). In this case the modeling
is a bit more complex since the user is required to setup a multi- domain analysis, but the
evaluation of the proper fluid parameters is somehow simpler since FINE™/Acoustics includes
some built-in models that simply require the flow resistivity as an input, and simple relations can
be found that relate the flow resistivity to the equivalent speed of sound and fluid density.
In several situations the surface impedance approach could in principle be the preferred solution
(for its modeling simplicity). In FINE™/Acoustics, a two steps procedure is proposed. In the first
step, the multi-domain approach is used to build a sort of virtual impedance tube that permits to
evaluate the impedance (admittance) of the desired sample of material. In the second step the
evaluated admittance is applied to the full muffler model.
In this tutorial, the surface impedance has been evaluated based on the virtual impedance tube
approach and provided in a file. The following part of the tutorial explains how to setup the
muffler model based on the evaluated admittance.

7.3 PREDICT TRANSMISSION LOSS USING A

WIZARD

This section explains how to predict the transmission loss by using the muffler wizard. A muffler
geometry file in STL format is used as a start.

207 FINE™/Acoustics 8.1 Tutorials

 1. Launch FINE™/Acoustics, select the menu Wizards / Muffler TL to start the wizard.
2. Select working directory and HEXPRESS™/Hybrid executable

l Select the working directory.

The muffler geometry file must be put in the working directory.

l Select the HEXPRESS™/Hybrid executable. The path to HEXPRESS™/Hybrid is

automatically detected and displayed. If necessary the user can modify it.

208 FINE™/Acoustics 8.1 Tutorials

 The HEXPRESS™/Hybrid version should be v5.2 or higher.

l Click on the button Next.

3. Define the muffler geometry.

l Select the muffler geometry file Wizard/[Link] provided in this tutorial.

The muffler geometry file must be in STL format. For more information on how to create the
geometry file, see Muffler wizard.

l Click on the button Next.

4. Define the frequency range of interest.

209 FINE™/Acoustics 8.1 Tutorials

 In this step, the user specifies the frequency range used during the transmission loss analysis
and the acoustic mesh resolution:
l Set the minimum frequency to 10 [Hz].
l Set the maximum frequency to 350 [Hz].
l Set the frequency step to 5 [Hz].
l Keep the minimum number of mesh points per wavelength to 6.
l Click on the button Next.
5. Define the surface admittance.

210 FINE™/Acoustics 8.1 Tutorials

 l Select the provided admittance file Wizard/[Link] for both Absorber 1 and
Absorber 2.
l Click on the button Plot to visualize the admittance in the provided file.
l Click on the button Next.
6. Define the fluid properties.

211 FINE™/Acoustics 8.1 Tutorials

 l Keep the default fluid properties.
l Click on the button Next.
7. The wizard starts the modeling process and finally gives the muffler’s transmission loss spectra
in narrow band and in ⅓ octave band.

212 FINE™/Acoustics 8.1 Tutorials

 After the user defines all the inputs, the wizard will automatically perform the following steps:
l Acoustic mesh generation: one or multiple acoustic meshes are automatically generated, as
needed to obtain a more efficient analysis.
l Acoustic model setup : once the meshing process is completed, the wizard automatically
setups the noise sources, the domains, the relative properties of the fluid and the absorbing
materials, the mesh interfaces to simulate the perforated sheets and all the boundary conditions.

213 FINE™/Acoustics 8.1 Tutorials

 l Transmission loss analysis : the FEM solver is then initialized considering the frequency
range provided by the user and the Transmission Loss analysis is started with the Impedance
Matrix Method (IMM) approach available in FINE™/Acoustics.
l Results post-processing: once the analysis is completed the results are assembled and post-
processed in order to obtain the muffler’s transmission loss spectra in narrow band and in ⅓
octave band. The spectra are finally displayed.

7.4 MANUALLY SETUP MUFFLER FEM

ANALYSIS

This section explains how to predict the transmission loss by manually setup a muffler model. A
mesh generated by HEXPRESS™/Hybrid and exported in CGNS format is used as a start.

214 FINE™/Acoustics 8.1 Tutorials

 1. Launch FINE™/Acoustics.
2. Import the acoustic mesh:
l The mesh is imported with the command Discretization > Import New > File .

l Select the CGNS generic file (*.cgns) and import the file GUI/muffler_ [Link]
provided in this tutorial in the file browser window.
l The Import Options and Transformations window appears as shown below.

215 FINE™/Acoustics 8.1 Tutorials

 l Deselect Panels under the frame Objects to be Imported
l Deselect all the quantities under the frame Global data to be imported.
l Select Extract Boundary surfaces from volumes under the frame Options.
l Keep all other settings and click on the button OK.
l Click on the zone muffler_normals_ok_1to activate it in the Enable Zones to be imported
window.

216 FINE™/Acoustics 8.1 Tutorials

 l Click on the button OK.
3. Save project.
l Using the command File > Save as ...to save the project.

It is suggested to save project frequently because there is no Undo option in FINE™/Acoustics.

4. Rename the discretization:

l Using the command Discretization > Rename, select Model 0 and rename it as FEM.
l Select the discretization FEM to view it.

5. Clean Groups:

217 FINE™/Acoustics 8.1 Tutorials

 l In the imported mesh, there are some duplicated groups. To delete these groups, select the
menu Groups > Transform / Delete.
l Activate the Delete Groups option, select all the groups except Global and Skin, select Yes
and then press OK to delete the groups.

The remaining groups are:

l Global: the volumes
l Skin: surface panels generated by the option Extract Boundary surfaces from volumes
during the import.

6. Create boundaries.

218 FINE™/Acoustics 8.1 Tutorials

 l Select the group Skin to view.
l Select the menu Groups > Create From > Sharp angles. Enter the name grp and click
the button OK to confirm.

l Use the menu Groups > Change Order/Rename to rename grp: 2 to Outlet, grd: 4 to
Inlet, grp: 5 to Absorber1 and grp: 6 to Absorber2. Optionally change the order.

l Use the menu Groups > Transform / Delete to create a group Complement containing the
groups grp: 1 and grp: 3.

219 FINE™/Acoustics 8.1 Tutorials

 l Use the menu Groups > Transform / Delete to delete the group Skin, grp: 1 and grp: 3.

220 FINE™/Acoustics 8.1 Tutorials

 7. Check mesh.
l With the command Analysis > Mesh info it is possible to visualize the information on
domain and mesh size. The maximum frequency that can be computed is calculated by
default considering a spatial resolution of 6 points per wavelength (this parameter can be
changed in Options > Parameters for > Mesh quality).

221 FINE™/Acoustics 8.1 Tutorials

 8. Create absorber material and assign the admittance to the perforated surfaces:

222 FINE™/Acoustics 8.1 Tutorials

 l The first step is to create a material that defines the admittance value and this can be done
with the command Boundary C. > Materials / New.

l Set the name as absorber>, select Freq. Dependant and click on the Define Admittance
button.
l Directly copy and paste the admittance values computed with the Virtual Impedance Tube
(see the file GUI/[Link]). Press now OK to insert the admittance values (delete the
extra-line in the end of the list if there is any), and OK again to save the material.
l Go to the menu Boundary C. > Define.

l Select the Material panel, and assign the absorber material to groups Absorber1 and
Absorber2.
l Press OK to confirm the material assignment.

223 FINE™/Acoustics 8.1 Tutorials

 9. Create flow domain and apply to all groups.
l Create a domain with the command: Boundary C. > Domains > New.

l Keep all default settings and click on the button OK.

224 FINE™/Acoustics 8.1 Tutorials

 When the flow speed and the velocity gradients are relevant it is possible to import a non-
uniform flow (e.g. computed with a RANS CFD simulation) by selecting Non Uniform - Flow
Definition. It is important to note that in this way it is taken into account only the effect that
mean sound velocity has on acoustic propagation while it is not considered in any way the
noise generated directly by the turbulence produced by the mean flow. This type of noise can
be computed with the FlowNoise module of FINE™/Acoustics.

l Apply the command Boundary C. > Define.

l Select the Domain tab and apply the Air domain to all groups.
10. Define boundary conditions.
l The Velocity boundary condition is assigned by default . This boundary condition
simulates perfectly sound reflective solid surfaces.

225 FINE™/Acoustics 8.1 Tutorials

 11. Start the FEM analysis:

The typical procedure for the calculation of transmission loss (TL) is based on the evaluation of
the so-called four pole parameters (A, B, C, D) that link the acoustic pressure and the acoustic
velocity at the muffler inlet to the corresponding quantities at the muffler outlet.

To calculate the four coefficients it is simply required to execute two sets of calculations that
differ only for the boundary conditions applied at the inlet and outlet.
Set Boundary condition at inlet Boundary condition at outlet
1 Imposed velocity v=1 Imposed velocity v=0
2 Imposed velocity v=1 Imposed pressure p=0

For the case of a zero mean flow,

From the set 1,

226 FINE™/Acoustics 8.1 Tutorials

 From the set 2,

It is possible to obtain an expression for the TL.

FINE™/Acoustics implements the calculation method above described, based on the Four Pole
Parameters transfer matrix. With this method the user needs to setup and execute two calculations,
required to obtain the four pole parameters, by imposing manually the relative boundary
conditions.
The two calculations can be started manually by the user, defining the proper BC at the inlet and
at the outlet. However, FINE™/Acoustics provides a specific feature that permits to simplify the
process of defining the BCs for TL analysis, and the two calculations can be started with a single
command.
This latter method, based on the Impedance Matrix approach, is applied in this tutorial.

l Use the command Analysis > TL Analysis > Impedance Matrix.

l Select the model FEM.
l Define the Inlet and the Outlet groups.

227 FINE™/Acoustics 8.1 Tutorials

 Eventually it is possible to add more than one inlet or outlet. If more than one inlet is present
the boxes |p| and <p allow to indicate the relative relation between the acoustic pressure
amplitudes and phases at the inlet sections considered.

l Click OK to accept.
l Set Name to TL.
l Set the Start Freq. as 10 Hz, the End Freq. as 350 Hz and the Delta Freq. as 5 Hz.
l Click Start to start the TL analysis.

228 FINE™/Acoustics 8.1 Tutorials

 l FINE™/Acoustics runs two linearly independent analysis, respectively initialized with BCs
at inlet and outlet. The solution provided are labeled as TL:case inlet and TL:case outlet.

229 FINE™/Acoustics 8.1 Tutorials

 12. Check the convergence of the solutions.

230 FINE™/Acoustics 8.1 Tutorials

 l Using command Solution > View Parameters.
l Then selecting the TL case:Outlet.
l Check the residuals of the analysis for all the frequencies (you get two values for each
frequency since FINE™/Acoustics performed two calculations with different BCs).

13. Post-processing: TL evaluation

l Open a new 2D window with the command Post-Processing > New 2D Window.
l Put the metric in dB with right click, Metrics > Decibel.
l Press the right mouse button inside the new window, and then select from the contextual
menu the command TL Evaluator > Impedance Matrix.

231 FINE™/Acoustics 8.1 Tutorials

 l Assign the solutions obtained to the respective groups Inlet and Outlet .

232 FINE™/Acoustics 8.1 Tutorials

 l Click on OK. The Transmission Loss curve (dB vs Frequency) is plotted.

233 FINE™/Acoustics 8.1 Tutorials

 l Once that the narrow band TL results is obtained we can transform it into 1/3 octave bands.
For this purpose press the right mouse button in the 2D window with the narrow band
results and select the command One third octave > Tonal > Linear. The TL spectrum in
one third octave bands can be saved in an external ASCII file.

7.5 REFERENCES

[1] "Analysis of concentric Tube resonators having unpartitioned cavities", [Link],

[Link], [Link]. 64, pp207-215 (1978).
[2] "Detailed Modeling of Mufflers with perforated Tubes Using Substructure Boundary Element
Method", Balasubramanian Datchanamourty, MS Degree Thesis, 2004, University of Kentucky.

234 FINE™/Acoustics 8.1 Tutorials

 [3] "ACOUSTICAL PROPERTIES OF FIBROUS ABSORBENT MATERIALS", M. E.
DELANY and E. N. BAZLEY, Applied Acoustics (3) (1970).
[4] "Noise and vibration control engineering, Principles and Applications", Leo L. Beranek,
István L. Vér.
[5] "Calculation of Acoustic Impedance of Multi-layer Absorbers", I. P. Dunn and W. A. Davern,
Applied Acoaustics 19 (1986) 321-334.
[6] "Empirical Relations between Acoustical Properties and Flow Resistivity of Porous Plastic
Open-Cell Foam", Wu Qunli, Applied Acoustics 25 (1988) 141-148.
[7] "Noise and Vibration Control", edited by: Leo L. Beranek, Istvan L. Ver Wiley Interscience,
Chapter eight.
[8] "Engineering Noise Control", David A. Bies and Colin H. Hansen Wiley Interscience,
Chapter eight - Appendix C.

235 FINE™/Acoustics 8.1 Tutorials

© NUMECA International, all rights reserved
EN201809191346