Professional Documents
Culture Documents
By:
Sandeep Sovani, Ph.D.
Senior Consulting Engineer
Fluent Inc., Ann Arbor, MI
March 18th, 2005
Presented at Fluent Lunch and Learn seminar series
St. Johns Conference Center, Plymouth, MI
Welcome!
Fluent Inc.s LunchNLearn Seminar Series
z Topical seminars on leading edge CFD applications
z Held frequently
z Purpose
Inform the FLUENT community about the subject
f Discuss basics, physics, theory, modeling techniques,
f Tools available in FLUENT to model the subject
f Examples
Outline
Aeroacoustics Basics
Simulation Methods
z Computational Aeroacoustics (CAA)
z Segregated Source-Propagation Methods (SSPM)
Fundamentals
Variational Methods
Boundary Element Methods
Integral Methods
Summary
Bibliography
Simulation Guide
3
Basics: Acoustics
Definitions
z Acoustics = The scientific study of sound[1]
z Sound = Pressure waves radiating in any material medium
History
z Sound was recognized to be a wave phenomenon over 2000
years ago![2]
Chrysippus (Greek philosoper, 240 BC)
Vetruvius (Roman architect and engineer, 25 BC)
z It propagates in a medium
Transports energy without transporting matter
Basics: Acoustics
Source, Medium, Propagation, Receiver
Source
Medium
Wave Propagation
Receiver
Source
Medium
Wave Propagation
Basics: Acoustics
Acoustics is sub-classified based on[2]
z Source
Aeroacoustics
Vibroacoustics
Etc.
z Medium
Hydroacoustics
Seismology
Etc.
z Etc
Basics: Aeroacoustics
Aeroacoustics
z Sub-area of acoustics where the source of sound is fluid flow
Characteristics
z No moving boundaries
such as electric speakers, vibrating strings, or vocal chords
Examples
z Whistles
z HVAC vent noise
z Automotive wind noise
Basics: Aeroacoustics
Receiver
Acoustic Medium
Flow
Sound
Source
Source Transient pressure variation caused by the flow
Sound Pressure waves propagating in the acoustic medium
1atm = 1E+5 Pa
SPL (dB)
100
80
60
40
20
0
1.E-04
1.E-03
1.E-02
1.E-01
1.E+00
Pressure (Pascal)
p' rms
SPL = 20log10
p
ref
1.E+01
1.E+02
p ref = 2 10 5 N / m 2
Outline
Aeroacoustics Basics
Simulation Methods
z Computational Aeroacoustics (CAA)
z Segregated Source-Propagation Methods (SSPM)
Fundamentals
Variational Methods
Boundary Element Methods
Integral Methods
Summary
Bibliography
Simulation Guide
10
Simulation: Objectives
Outputs typically desired from a aeroacoustics study
z Source Strengths
Source Ranking
z Frequency Spectrum
At observer
z Directivity
11
Simulation: Aspects
To obtain the desired outputs two aspects need to
be simulated
z Sound source
Provides source characteristics and rankings
z Sound propagation
Propagation of sound from the source to the receiver
f Requires input of source characteristics
f Provides
Sound spectrum and receiver
Sound directivity
12
Simulation: Approaches
There are 3 primary simulation approaches
z Computational Aeroacoustics (CAA)
Sometimes referred to as Direct Noise Computation (DNC)
Sound sources and propagation solved in a single
comprehensive model
Theory
Applicability
Advantages/Disadvantages
Examples
13
Outline
Aeroacoustics Basics
Simulation Methods
z Computational Aeroacoustics (CAA)
z Segregated Source-Propagation Methods (SSPM)
Fundamentals
Variational Methods
Boundary Element Methods
Integral Methods
Summary
Bibliography
Simulation Guide
14
CAA: Theory
CAA = Computational AeroAcoustics
CAA :: Aeroacoustics
DNS :: Turbulence
Premise
z Fluid flow at sound source and sound propagation, both
are fluid phenomena
Therefore both are governed by Navier-Stokes equations
15
CAA: Theory
Receiver
Propagation
p(t)
Sound Source
Computational Domain
16
CAA: Applicability
Practical problems in using CAA
1] Frequency range (20 Hz ~ 20,000 Hz)
z Acoustic timescales are often orders of magnitude
greater than turbulence timescales
z Simulation needs to be run for long real time with a
small timesteps, i.e. for large no. of timesteps
17
CAA: Applicability
140
120
SPL (dB)
100
80
60
40
patm ~ 1E+5 Pa
20
0
1.E-04
1.E-03
1.E-02
1.E-01
1.E+00
Pressure (Pascal)
18
1.E+01
1.E+02
CAA: Applicability
CAA is practically applicable only to cases
where these 3 obstacles are relatively minor
z Frequency range
Lower the better
z Sound pressure
Larger the better
19
CAA: Applicability
Frequency
CAA: Advantages/Disadvantages
Advantages
z Simple to implement
Single simulation solves sound generation as well as
propagation
Disadvantages
z Limited applicability
As discussed on previous slides
z Computationally expensive
Large meshes
Long transient computations
21
22
MDM mesh
motion
23
/x=44
f =1500Hz
kr = 2.72
1 80.7o
160,000 grid points
/x=22
f =3000Hz
kr = 5.43
1 44.8o
f =6000Hz
Acoustic beaming
for higher frequencies
correctly predicted
Marginally resolved
propagation
/x=11
/x=22
160,000
grid points
kr = 10.86
1 20.7o
640,000 grid points
25
Example
z Aimed at predicting sound pressure spectrum at
drivers and passengers ears when side window is
open
z Parametric studies
Effect of position inside the cabin
Different window openings
Different mirror designs
26
z SUV:
An, C.-F., Alaie, S.M., Sovani, S.D., Scislowicz M.,
Singh, K., Side window buffeting characteristics of a
SUV, Vehicle Aerodynamics, Vol. SP1874, pp. 43 - 53,
SAE International Paper 2004-01-0230 (2004)[4]
27
The figure that originally appeared on this page has been removed.
See Figures 1 and 2 from SAE paper 2003-01-1316
The figure that originally appeared on this page has been removed.
See Figure 4 from SAE paper 2003-01-1316
The figure that originally appeared on this page has been removed.
See Figure 3 and Tables 2,3 from SAE paper 2003-01-1316
The figure that originally appeared on this page has been removed.
See Figure 6 from SAE paper 2003-01-1316
31
The figure that originally appeared on this page has been removed.
See Figure 7 from SAE paper 2003-01-1316
The figure that originally appeared on this page has been removed.
See Figure 10 from SAE paper 2003-01-1316
The figure that originally appeared on this page has been removed.
See Figure 13 from SAE paper 2003-01-1316
The figure that originally appeared on this page has been removed.
See Figure 14 from SAE paper 2003-01-1316
The figure that originally appeared on this page has been removed.
See Figures 18 from SAE paper 2003-01-1316
38
39
Pressure Far-Field
Walls
Velocity
Inlet
Symmetry
40
Mesh
Mirror
Inlet
41
z Discretization schemes:
Time: 2nd order implicit
Momentum: 2nd order upwind
Pressure-Velocity Coupling: SIMPLE
Transient Solution:
z
z
z
z
42
43
Side View
Pt 101
Top View
44
Experimental
CFD - CAA
CFD - AA
SPL (dB)
90
70
Reference for
Experimental Data:
Hold et al. (AIAA-991896)[5] and Seigert
et al. (AIAA-991895)[6]
50
30
10
0
500
1000
Frequency (Hz)
45
1500
2000
Inlet
Throttle
Body
Inclined
Face
Intake Outlet
Throttle Sidebranch
Manifold
Cavity
Plate
46
Airflow
47
Baseline
=43
Airflow
Modified
Inlet
(101325 Pa)
43
High Mesh
Density
Monitor
Point A
Zip Tube
Throttle
Plate
48
Sump
Outlet
92325Pa
P = 9kPa
Computational - 2D
Experimental - 3D
180
170
159dB @
2125Hz
160
153dB @
1710 Hz
SPL (dB)
150
140
130
120
110
100
90
80
0
500
1000
1500
2000
2500
3000
Frequency (Hz)
3500
4000
4500
5000
Computational - 2D
Experimental - 3D
180
170
137dB @
2125Hz
150 dB @
1710 Hz
160
SPL (dB)
150
140
130
120
110
100
90
80
0
500
1000
1500
2000
2500
3000
Frequency (Hz)
3500
4000
4500
5000
Frequency
53
Outline
Aeroacoustics Basics
Simulation Methods
z Computational Aeroacoustics (CAA)
z Segregated Source-Propagation Methods (SSPM)
Fundamentals
Variational Methods
Boundary Element Methods
Integral Methods
Summary
Bibliography
Simulation Guide
54
SSPM: Theory
Segregated Source-Propagation Methods
z Sound generation and propagation are independent phenomena
in most cases
They happen at vastly different scales
f Flow pressure ~ 1 kPa; Acoustics pressure ~ 1 mPa
f Turbulence length scales ~ 1m; Acoustic wavelengths ~1m
f Turbulence time scales ~ 1 s; Acoustic timescales ~ 1ms
z Advantages
Reduced computational effort
Expanded applicability to a wide variety of problems
55
SSPM: Theory
Derivation of the Wave Equation
z Linearized Continuity Equation (For fluctuations)
'
u '
+ i =0
t
xi
z Linearized Momentum Equation
(No convection, no body forces, no viscous stresses)
z Eliminate
ui ' p'
0
+
=0
t xi
2 ' 2 p'
2 =0
2
t
xi
u'i
p' = c '
2
z Wave Equation
p
c2 =
2
p'
2 p'
2
c
=0
2
2
t
xi
56
SSPM: Theory
Connection between the
two segregated parts of
the problem: source and
propagation
z Sir James Lighthill provided
the mathematical foundation
for connecting the source
and propagation parts
z Famous Lighthills Acoustic
Analogy [9]
57
SSPM: Theory
Lighthills Acoustic Analogy[9]
z Continuity Equation
+
( ui ) = 0
t xi
z Momentum Equation
(Convection included, but no viscous stresses)
u i
u i
p
=
+uj
x j
xi
t
In a conservative form
u i u i u j
p
+
=
t
x j
xi
z Eliminate
ui
2 ( u iu j )
2
2p
=
+
2
t
xix j
xixi
58
SSPM: Theory
Lighthills Acoustic Analogy (continued)
z The previous equation can be cast in the form of a wave
equation in an undisturbed medium at rest by subtracting
from both sides:
2
co2
'
xi xi
Where
2Tij
2' 2 2'
c0
=
2
t
xi xi xi x j
Tij = ui u j + ( p co2 ) ij
This is referred to as Lighthills tensor
For nearly isentropic flows:
Tij o ui u j
59
SSPM: Theory
Lighthills Acoustic Analogy (continued)
z Lighthills equation can be thought of as a wave equation with
a source term
Wave Equation
2 p' 2 2 p'
=0
c0
2
xi xi
t
Lighthills Equation
2Tij
2' 2 2'
c0
=
2
t
xi xi xi x j
Propagation
Source
60
SSPM: Theory
Computing Propagation
z Once CFD provides sound source information the problem
reduces to solving for sound propagation
z Several methods exist for this with varying level of
simplification
Rigorous Lighthill equation solution
f Finite Difference Methods
f Variational Methods (Finite Element)[10]
61
SSPM: Implementation
FLUENT provides features to compute sound propagation
using several of these methods
z Variational Methods (Finite Element Method)
ACTRAN-LA is a third party Lighthills equation solver code
developed by Free Field Technologies
f Fluent has an interface to export Lighthills tensor and other variables
to ACTRAN
z Integral Methods
FLUENT has an inbuilt sound propagation module based on the
Ffowcs-Williams and Hawkings (FWH) Method[12]
f Part of the standard FLUENT package, no add-on components
required
62
SSPM: Implementation
SSPM Implementation follows these steps for each method
z Perform well resolved transient simulation of flow in and
around the sound source regions
Save required data (such as Lighthills tensor, or time-varying
surface pressure) on source regions
Receiver
CFD Domain
Sound
Propagation
Source
63
p(t)
Outline
Aeroacoustics Basics
Simulation Methods
z Computational Aeroacoustics (CAA)
z Segregated Source-Propagation Methods (SSPM)
Fundamentals
Variational Methods
Boundary Element Methods
Integral Methods
Summary
Bibliography
Simulation Guide
64
65
Disadvantages
z Cannot account for backward effect of sound on flow
z Cannot account for reflections
z Needs straight line of sight from source to receiver
66
z Special addition
steady state fan noise (Gutin noise)
67
Targeted applications
z External aerodynamic noise
Side view mirrors, Louvers, .
Landing gear, high-lift devices,
z Fan/rotor noise
z Jet noise
68
110
Point 101
Experimental
CFD - CAA
CFD - AA
SPL (dB)
90
70
50
30
10
0
500
1000
Frequency (Hz)
69
1500
2000
Mesh:
5.3M cells
173,000 surfaces tri-elements, s=0.0135D
6 prism layers, h1=1.6x10-3D
Run time:
3min 40s per time step, 4 nodes
9950 time steps for flow to pass through
domain (18D), 25 days run time
FW-H analysis:
Source data extraction after one flow pass
Source data sampled during 0.9 flow
passes
70
71
72
Receiver
OASP
92.9 dB
98.3 dB
p(t)
93.4 dB
96.8 dB
SPL receivers 1, 3
92.3 dB
73
Microphone
D
K
Flow
74
Static pressure
contours
Static pressure
contours
Vorticity Iso-surface
colored by velocity mag.
75
Rain gutter
120
Experiment (Kumarasamy and Karbon)
Fluent (LES Central Differencing)
100
80
SPL (dB)
Rain Gutter
LES
Central Differencing
Run 2
Model A
60
40
20
0
0
500
1000
1500
x (m)
77
Reference:
Kumarasamy S. and
Karbon K., Aeroacoustics
of an Automobile A-Pillar
2000Rain Gutter:
Computational and
Experimental Study, SAE
Paper 1999-01-1128,
(1999)[16]
100
80
Rain Gutter
LES
Central Differencing
Run 2
Model A
Timestep size = 8E-5 sec
Far-field microphone location
60
SPL (dB)
40
20
0
0
500
1000
1500
Frequency (Hz)
78
Reference:
Kumarasamy S. and
Karbon K., Aeroacoustics
of an Automobile A-Pillar
Rain Gutter:
2000Computational and
Experimental Study, SAE
Paper 1999-01-1128,
(1999)[16]
Outline
Aeroacoustics Basics
Simulation Methods
z Computational Aeroacoustics (CAA)
z Segregated Source-Propagation Methods (SSPM)
Fundamentals
Variational Methods
Boundary Element Methods
Integral Methods
Summary
Bibliography
Simulation Guide
79
SNGR
Rationale
z Unsteady simulations (LES and DES) are time-consuming
z Steady RANS results contain a fair amount of useful
information
SNGR
FLUENT6.2s Offering:
Broadband Noise Source Models
Source terms in the acoustic
equations
z Lilleys equation
z Linearized Euler equation (LEE)
81
SNGR
BB Model 1: Lilleys Acoustic Sources
Steady RANS results used to synthesize turbulent
velocity fields with the stochastic noise generation and
radiation (SNGR) method.
FLUENT reports the r.m.s. values of the source terms.
z Self-noise sources
z Shear-noise sources
Iso-surface of
Lilleys acoustic
source (total)
strength
82
SNGR
BB Model 2: LEE Source Terms
Source terms in LEE (Linearized Euler Equation)
uai
u
u
U i
u
+ U j ai + = U j ti utj
utj ti
x j
t
x j
x j
x j
1442443 1
424
3
Shear noise
Self noise
83
SNGR
BB Model 3: Proudman Acoustic Power
Originally derived by Proudman (1952) for noise due to
isotropic turbulence (quadrupole sources)
Recently re-derived (Lilley, 1993) and confirmed using
DNS (Sarkar and Hussaini, 1993)
Simple yet very useful to determine the local contribution
to the total acoustic power.
84
SNGR: Example
Broadband noise source models are practically
useful for
z Determine prominent noise generating regions in a flow
domain
z Determine noise rankings of different variations of a
design
85
SNGR: Example
Generic HVAC duct
z Baseline design
z Three additional design variation
z Aim: determine noise ranking with BB models and
compare with experimentally determined noise
rankings
The figure that originally appeared on this page has been removed.
See SAE paper 2005-01-2495 On Predicting the Aeroacoustic
Performance of Ducts with Broadband Noise Source Models.
86
SNGR: Example
Design Variations
Design 2
Design 4
Design 3
87
SNGR: Example
Mesh
The figure that originally appeared on this page has been removed.
See SAE paper 2005-01-2495 On Predicting the Aeroacoustic
Performance of Ducts with Broadband Noise Source Models.
88
SNGR: Example
Solver Settings and Boundary Conditions
Function
Setting
Steady state,
Segregated
Implicit
Double Precision
RNG k-e
2nd order
2nd order upwind
SIMPLEC
Air (incompressible)
Solver
Precision
Turbulence Model
Pressure discretization
Momentum
discretization
Pressure-velocity
coupling
Fluid
Boundary
Boundary
Condition
Value
Constant Velocity
Interior
No slip wall
Constant
Pressure
No slip wall
7.507 m/s
(=300 cfm)
0 Pa (gage)
Duct Inlet
Duct Outlet
Duct boundaries
Plenum Outlet
Plenum
boundaries
89
SNGR: Example
Flow Structure: Contours of velocity magnitude
The figure that originally appeared on this page has been removed.
See SAE paper 2005-01-2495 On Predicting the Aeroacoustic
Performance of Ducts with Broadband Noise Source Models.
90
SNGR: Example
Acoustic power generated inside the duct as estimated
from the BB models
z (1) Proudman Formula (Representative of quadrupole
contribution)
z (2) Turbulent Boundary Layer Noise (Representative of
dipole contribution)
Design
Design1
Design2
Design3
Design4
91
SNGR: Example
Experimental Measurements
z Sound spectrum measured at a point 1 m directly
downstream of the ducts outlet
The figure that originally appeared on this page has been removed.
See SAE paper 2005-01-2495 On Predicting the Aeroacoustic
Performance of Ducts with Broadband Noise Source Models.
Outline
Aeroacoustics Basics
Simulation Methods
z Computational Aeroacoustics (CAA)
z Segregated Source-Propagation Methods (SSPM)
Fundamentals
Variational Methods
Boundary Element Methods
Integral Methods
Summary
Bibliography
Simulation Guide
93
Summary
94
Outline
Aeroacoustics Basics
Simulation Methods
z Computational Aeroacoustics (CAA)
z Segregated Source-Propagation Methods (SSPM)
Fundamentals
Variational Methods
Boundary Element Methods
Integral Methods
Summary
Bibliography
Simulation Guide
95
Bibliography
1.
2.
3.
4.
5.
6.
Bibliography
7.
Kannan V., Sovani S.D., Greeley, D., and Khondge A., Computational
Aeroacoustics Simulation of Whistle Noise in an Automotive Air-Intake
System, SAE International Paper No. 2005-01-2364 (2005)
8. Fluent Inc., Brochure FLUENT6 for Acoustics Modeling, Lebanon, NH
(2005)
9. Lighthill M. J., On Sound Generated Aerodynamically, I. General Theory,
Proc. Roy. Soc. London, A211, page 564, (1952).
10. Caro S., Ploumhans P. and Gallez X., Implementation of Lighthills
acoustic analogy in a finite/infinite elements framework, 10th AIAA/CEAS
Aeroacoustics Conference, AIAA Paper number 2004-2891 (2004)
11. Lyrintzis A.S., Surface integral methods in computational aeroacoustics
From the (CFD) near-field to the (Acoustic) far-field, International Journal
of Aeroacoustics, vol. 2, pp. 95-128, (2003)
12. Kim S.-E., Dai Y., Koutsavdis E., Sovani S., Kadam N., Ravuri K.M.R., A
versatile implementation of acoustic analogy based noise prediction
method in a general-purpose CFD code, AIAA paper 2003-3202 (2003)
97
Bibliography
13. Ffowcs Williams, J.E. and Hawkins, D.L., Sound generation by
turbulence and surfaces in arbitrary motion, Proceedings of the Royal
Society of London A, vol. 264, pp. 321-342 (1969)
14. Lokhande B.S., Sovani S.D., Xu J., Computational Aeroacoustic Analysis
of a Generic Side View Mirror, SAE Paper 2003-01-1698 (2003)
15. Sovani S.D. and Chen K.-H., Aeroacoustics of an "Automobile" A-Pillar
"Rain Gutter": A Numerical Study with the Ffowcs-Williams and Hawkings
Method, SAE Paper 2005-01-2492 (2005)
16. Kumarasamy S. and Karbon K., Aeroacoustics of an Automobile A-Pillar
Rain Gutter: Computational and Experimental Study, SAE Paper 199901-1128, (1999)
17. Khondge A.D., Sovani S.D., Kim S.-E., Farag A.A., and Guzy S.C., On
Predicting the Aeroacoustic Performance of Ducts with Broadband Noise
Source Models, SAE Paper 2005-01-2495 (2005)
98
Outline
Aeroacoustics Basics
Simulation Methods
z Computational Aeroacoustics (CAA)
z Segregated Source-Propagation Methods (SSPM)
Fundamentals
Variational Methods
Boundary Element Methods
Integral Methods
Summary
Bibliography
Simulation Guide
z CAA
z FWH Method
99
z In transmission region
Mesh edge length =
{shortest sound wavelength of interest}/10
Transmission
Region
Source Region
100
Receiver
p(t)
101
Outline
Aeroacoustics Basics
Simulation Methods
z Computational Aeroacoustics (CAA)
z Segregated Source-Propagation Methods (SSPM)
Fundamentals
Variational Methods
Boundary Element Methods
Integral Methods
Summary
Bibliography
Simulation Guide
z CAA
z FWH Method
102
103
Mesh
wall source
surface
interior source surface
Duct
Interior source surfaces
104
Simulation Procedure
STEP1: Processing
1(A) Setup Fluent case;
Activate FW-H acoustics model
105
Simulation Procedure
STEP2: Post-Processing
2(A) Specify receiver locations in Acoustic
Receivers panel
2(B) Read .index and .asd files
Compute/write acoustic signals
106
107
fwh3
108
109
110
111
Auto-pruned signal
112
113
Hamming
Hanning
Barlett
Blackman
114
Original Data
Pruned Data
115