CHP 21

Chapter 21.
Predicting Aerodynamically Generated Noise

The discipline of acoustics, which is a study of pressure waves in uids, is intimately
related to uid dynamics. Many sounds that are technologically important in industrial
applications are generated by and propagated in uid ows. The phenomena associ-
ated with sounds can therefore be understood and analyzed in the general framework of
uid dynamics. The governing equations for acoustics are indeed the same as the ones
governing uid ows.
The main challenge in numerically predicting sound waves stems from the well-recognized
fact that sounds have much lower energy than uid ows, typically by several orders of
magnitude. This poses a great challenge to the computation of sounds in terms of di-
culty of numerically resolving sound waves, especially when one is interested in predicting
sound propagation to the far eld. Another challenge comes from the diculty of pre-
dicting the very ow phenomena (e.g., turbulence) in the near eld that are responsible
for generating sounds.
This chapter provides an overview of FLUENTs approaches to computing aerodynami-
cally generated sound, the underlying theories, the model setup, and the procedure for
computing sound.
Section 21.1: Overview
Section 21.2: Acoustics Model Theory
Section 21.3: Using the Ffowcs Williams and Hawkings Acoustics Model
Section 21.4: Using the Broadband Noise Source Models
21.1 Overview
Considering the breadth of the discipline and the challenges encountered in comput-
ing sounds, it is not surprising that a number of computational approaches have been
proposed over the years whose sophistication, applicability, and cost widely vary.
FLUENT oers three approaches to computing sounds, a direct method, an integral
method based on acoustic analogy, and a method that utilizes broadband noise source
models.
c Fluent Inc. September 29, 2006 21-1
21.1.1 Direct Method
In this method, both generation and propagation of sound waves are directly computed
by solving the appropriate uid dynamics equations. Prediction of sound waves always
requires time-accurate solutions to the governing equations. Furthermore, in most prac-
tical applications of the direct method, one has to employ governing equations that are
capable of modeling viscous and turbulence eects, such as unsteady Navier-Stokes equa-
tions (i.e., DNS), RANS equations, and ltered equations used in DES and LES.
The direct method is thus computationally dicult and expensive inasmuch as it requires
highly accurate numerics, very ne computational meshes all the way to receivers, and
acoustically nonreecting boundary conditions. The computational cost becomes pro-
hibitive when sound is to be predicted in the far eld (e.g., hundreds of chord-lengths in
the case of an airfoil). The direct method becomes feasible when receivers are in the near
eld (e.g., cabin noise). In many such situations involving near-eld sound, sounds (or
pseudo-sounds for that matter) are predominantly due to local hydrodynamic pressure
which can be predicted with a reasonable cost and accuracy.
Since sound propagation is directly resolved in this method, one normally needs to solve
the compressible form of the governing equations (e.g., compressible RANS equations,
compressible form of ltered equations for LES). Only in situations where the ow is low
subsonic and the receivers in the near eld sense primarily local hydrodynamic pressure
uctuations (i.e., pseudo sound) can incompressible ow formulations be used. But
this incompressible treatment will also not allow to simulate resonance and feedback
phenomena.
21.1.2 Integral Method Based on Acoustic Analogy
For predictions of mid- to far-eld noise, the methods based on Lighthills acoustic anal-
ogy [208] oer viable alternatives to the direct method. In this approach, the near-eld
ow obtained from appropriate governing equations such as unsteady RANS equations,
DES, or LES are used to predict the sound with the aid of analytically derived integral
solutions to wave equations. The acoustic analogy essentially decouples the propagation
of sound from its generation, allowing one to separate the ow solution process from the
acoustics analysis.
FLUENT oers a method based on the Ffowcs Williams and Hawkings (FW-H) equation
and its integral solutions [103]. The FW-H formulation adopts the most general form
of Lighthills acoustic analogy, and is capable of predicting sound generated by equiv-
alent acoustic sources such as monopoles, dipoles, and quadrupoles. FLUENT adopts a
time-domain integral formulation wherein time histories of sound pressure, or acoustic
signals, at prescribed receiver locations are directly computed by evaluating a few surface
integrals.
Time-accurate solutions of the ow-eld variables, such as pressure, velocity components,
and density on source (emission) surfaces, are required to evaluate the surface integrals.
21-2 c Fluent Inc. September 29, 2006
21.1 Overview
Time-accurate solutions can be obtained from unsteady Reynolds-averaged Navier-Stokes
(URANS) equations, large eddy simulation (LES), or detached eddy simulation (DES) as
appropriate for the ow at hand and the features that you want to capture (e.g., vortex
shedding). The source surfaces can be placed not only on impermeable walls, but also on
interior (permeable) surfaces, which enables you to account for the contributions from
the quadrupoles enclosed by the source surfaces. Both broadband and tonal noise can be
predicted depending on the nature of the ow (noise source) being considered, turbulence
model employed, and the time scale of the ow resolved in the ow calculation.
The FW-H acoustics model in FLUENT allows you to select multiple source surfaces and
receivers. It also permits you either to save the source data for a future use, or to carry
out an on the y acoustic calculation simultaneously as the transient ow calculation
proceeds, or both. Sound pressure signals thus obtained can be processed using the fast
Fourier transform (FFT) and associated postprocessing capabilities to compute and plot
such acoustic quantities as the overall sound pressure level (SPL) and power spectra.
One important limitation of FLUENTs FW-H model is that it is applicable only to
predicting the propagation of sound toward free space. Thus, while the model can be
legitimately used to predict far-eld noise due to external aerodynamic ows, such as
the ows around ground vehicles and aircrafts, it cannot be used for predicting the noise
propagation inside ducts or wall-enclosed space.
21.1.3 Broadband Noise Source Models
In many practical applications involving turbulent ows, noise does not have any distinct
tones, and the sound energy is continuously distributed over a broad range of frequencies.
In those situations involving broadband noise, statistical turbulence quantities readily
computable from RANS equations can be utilized, in conjunction with semi-empirical
correlations and Lighthills acoustic analogy, to shed some light on the source of broad-
band noise.
FLUENT oers several such source models that enable you to quantify the local contribu-
tion (per unit surface area or volume) to the total acoustic power generated by the ow.
They include the following:
Proudmans formula
jet noise source model
boundary layer noise source model
source terms in the linearized Euler equations
source terms in Lilleys equation
Considering that one would ultimately want to come up with some measures to mitigate
the noise generated by the ow in question, the source models can be employed to extract
useful diagnostics on the noise source to determine which portion of the ow is primarily
responsible for the noise generation. Note, however, that these source models do not
predict the sound at receivers.
Unlike the direct method and the FW-H integral method, the broadband noise source
models do not require transient solutions to any governing uid dynamics equations. All
the source models need is what typical RANS models would provide, such as the mean
velocity eld, turbulent kinetic energy (k) and the dissipation rate (). Therefore, the
use of broadband noise source models requires the least computational resources.
21.2 Acoustics Model Theory
This section describes the theoretical background for the Ffowcs Williams and Hawkings
model and the broadband noise source models.
21.2.1 The Ffowcs Williams and Hawkings Model
The Ffowcs Williams and Hawkings (FW-H) equation is essentially an inhomogeneous
wave equation that can be derived by manipulating the continuity equation and the
Navier-Stokes equations. The FW-H [42, 103] equation can be written as:
1
a
2
0
2
p
t
2

2
p
=

2
x
i
x
j
{T
ij
H(f)}

x
i
{[P
ij
n
j
+ u
i
(u
n
v
n
)] (f)}
+

t
{[
0
v
n
+ (u
n
v
n
)] (f)} (21.2-1)
where
u
i
= uid velocity component in the x
i
direction
u
n
= uid velocity component normal to the surface f = 0
v
i
= surface velocity components in the x
i
direction
v
n
= surface velocity component normal to the surface
(f) = Dirac delta function
H(f) = Heaviside function
p
is the sound pressure at the far eld (p
= p p
0
). f = 0 denotes a mathematical
surface introduced to embed the exterior ow problem (f > 0) in an unbounded space,
which facilitates the use of generalized function theory and the free-space Green function
to obtain the solution. The surface (f = 0) corresponds to the source (emission) surface,
and can be made coincident with a body (impermeable) surface or a permeable surface
o the body surface. n
i
is the unit normal vector pointing toward the exterior region
(f > 0), a
0
is the far-eld sound speed, and T
ij
is the Lighthill stress tensor, dened as
T
ij
= u
i
u
j
+ P
ij
a
2
0
(
0
)
ij
(21.2-2)
P
ij
is the compressive stress tensor. For a Stokesian uid, this is given by
P
ij
= p
ij
_
u
i
x
j
+
u
j
x
i
2
3
u
k
x
k
ij
_
(21.2-3)
The free-stream quantities are denoted by the subscript 0.
The solution to Equation 21.2-1 is obtained using the free-space Green function ((g)/4r).
The complete solution consists of surface integrals and volume integrals. The surface inte-
grals represent the contributions from monopole and dipole acoustic sources and partially
from quadrupole sources, whereas the volume integrals represent quadrupole (volume)
sources in the region outside the source surface. The contribution of the volume integrals
becomes small when the ow is low subsonic and the source surface encloses the source
region. In FLUENT, the volume integrals are dropped. Thus, we have
p
(x, t) = p
T
(x, t) + p
L
(x, t) (21.2-4)
where
4p
T
(x, t) =
_
f=0
_
_
0
_

U
n
+ U
n
_
r (1 M
r
)
2
_
_
dS
+
_
f=0
_
_
0
U
n
_
r

M
r
+ a
0
(M
r
M
2
)
_
r
2
(1 M
r
)
3
_
_
dS (21.2-5)
4p
L
(x, t) =
1
a
0
_
f=0
_

L
r
r (1 M
r
)
2
_
dS
+
_
f=0
_
L
r
L
M
r
2
(1 M
r
)
2
_
dS
+
1
a
0
_
f=0
_
_
L
r
_
r

M
r
+ a
0
(M
r
M
2
)
_
r
2
(1 M
r
)
3
_
_
dS (21.2-6)
where
U
i
= v
i
+

0
(u
i
v
i
) (21.2-7)
L
i
= P
ij
n
j
+ u
i
(u
n
v
n
) (21.2-8)
When the integration surface coincides with an impenetrable wall, the two terms on the
right in Equation 21.2-4, p
T
(x, t) and p
L
(x, t), are often referred to as thickness and
loading terms, respectively, in light of their physical meanings. The square brackets in
Equations 21.2-5 and 21.2-6 denote that the kernels of the integrals are computed at the
corresponding retarded times, , dened as follows, given the observer time, t, and the
distance to the observer, r,
= t
r
a
0
(21.2-9)
The various subscripted quantities appearing in Equations 21.2-5 and 21.2-6 are the
inner products of a vector and a unit vector implied by the subscript. For instance,
L
r
=

L
r = L
i
r
i
and U
n
=

U n = U
i
n
i
, where r and n denote the unit vectors in
the radiation and wall-normal directions, respectively. The dot over a variable denotes
source-time dierentiation of that variable.
Please note the following remarks regarding the applicability of this integral solution:
The FW-H formulation in FLUENT can handle rotating surfaces as well as station-
ary surfaces.
It is not required that the surface f = 0 coincide with body surfaces or walls. The
formulation permits source surfaces to be permeable, and therefore can be placed
in the interior of the ow.
When a permeable source surface (either interior or nonconformal sliding interface)
is placed at a certain distance o the body surface, the integral solutions given by
Equations 21.2-5 and 21.2-6 include the contributions from the quadrupole sources
within the region enclosed by the source surface. When using a permeable source
surface, the mesh resolution needs to be ne enough to resolve the transient ow
structures inside the volume enclosed by the permeable surface.
21.2.2 Broadband Noise Source Models
Proudmans Formula
Proudman [292], using Lighthills acoustic analogy, derived a formula for acoustic power
generated by isotropic turbulence without mean ow. More recently, Lilley [209] red-
erived the formula by accounting for the retarded time dierence which was neglected
in Proudmans original derivation. Both derivations yield acoustic power due to unit
volume of isotropic turbulence (in W/m
3
) as
P
A
=
0
_
u
3
_
u
5
a
5
0
(21.2-10)
where u and are the turbulence velocity and length scales, respectively, and a
0
is
the speed of sound. in Equation 21.2-10 is a model constant. In terms of k and ,
Equation 21.2-10 can be rewritten as
P
A
=
0
M
5
t
(21.2-11)
where
M
t
=
2k
a
0
(21.2-12)
The rescaled constant,
, is set to 0.1 in FLUENT based on the calibration of Sarkar and

Hussaini [315] using direct numerical simulation of isotropic turbulence.
FLUENT can also report the acoustic power in dB, which is computed from
L
P
= 10 log
_
P
A
P
ref
_
(21.2-13)
where P
ref
is the reference acoustic power (P
ref
= 10
12
W/m
3
by default).
The Proudmans formula gives an approximate measure of the local contribution to total
acoustic power per unit volume in a given turbulence eld. Proper caution, however,
should be taken when interpreting the results in view of the assumptions made in the
derivation, such as high Reynolds number, small Mach number, isotropy of turbulence,
and zero mean motion.
The Jet Noise Source Model
This source model for axisymmetric jets is based on the works of Goldstein [123] who
modied the model originally proposed by Ribner [304] to better account for anisotropy
of turbulence in axisymmetric turbulent jets.
In Goldsteins model, the total acoustic power emitted by the unit volume of a turbulent
jet is computed from
P
A
(y) =
_
2
0
_

0
I(r, ; y)r
2
sin d d
= 2r
2
_

0
I(r, ; y) sin d (21.2-14)
where r and are the radial and angular coordinates of the receiver location, and I(r, ; y)
is the directional acoustic intensity per unit volume of a jet dened by
I(r, ; y) =
12
0

4
f
L
1
L
2
2
u
2
t1
2
5 a
5
0
r
2
D
self
C
5
+
24
0

4
f
L
1
L
4
2
u
2
t1
a
5
0
r
2
_
U
r
_
2
D
shear
C
5
(21.2-15)
C in Equation 21.2-15 is the modied convection factor dened by
C = 1 M
c
cos (21.2-16)
and
D
self
= 1 + 2(
M
9
N) cos
2
sin
2
+
1
3
_
M
2
7
+ M 1.5N(3 3N +
1.5
2

2
2
)
_
sin
4
(21.2-17)
D
shear
= cos
2
_
cos
2
+
1
2
_
1
2
2N
_
sin
2
_
(21.2-18)
The remaining parameters are dened as
=
L
2
L
1
(21.2-19)
M =
_
3
2
_
__
2
(21.2-20)
N = 1
_
u
2
t2
_
_
u
2
t1
_
(21.2-21)
L
1
=
_
u
2
t1
_
3/2
(21.2-22)
L
2
=
_
u
2
t2
_
3/2
(21.2-23)
f
= 2

k
(21.2-24)
where u
2
t1
and u
2
t2
are computed dierently depending on the turbulence model chosen for
the computation. When the RSM is selected, they are computed from the corresponding
normal stresses. For all other two-equation turbulence models, they are obtained from
u
2
t1
=
8
9
k (21.2-25)
u
2
t2
=
4
9
k (21.2-26)
FLUENT reports the acoustic power both in the dimensional units (W/m
3
) and in dB
computed from
L
P
= 10 log
_
P
A
P
ref
_
(21.2-27)
where P
ref
is the reference acoustic power (P
ref
= 10
12
W/m
3
by default).
The Boundary Layer Noise Source Model
Far-eld sound generated by turbulent boundary layer ow over a solid body at low
Mach numbers is often of practical interest. The Curles integral [70] based on acoustic
analogy can be used to approximate the local contribution from the body surface to the
total acoustic power. To that end, one can start with the Curles integral
p
(x, t) =
1
4a
0
_
S
(x
i
y
i
) n
i
r
2
p
t
(y, ) dS(y) (21.2-28)
where denotes the emission time ( = t r/a
0
), and S the integration surface.
Using this, the sound intensity in the far eld can then be approximated by
p
2
1
16
2
a
2
0
_
S
cos
2
r
2
_
p
t
(y, )
_
2
A
c
(y) dS(y) (21.2-29)
where A
c
is the correlation area, r |x y|, and cos is the angle between |x y| and
the wall-normal direction n.
The total acoustic power emitted from the entire body surface can be computed from
P
A
=
1
0
a
0
_
2
0
_

0
p
2
r
2
sin dd
=
_
S
I(y) dS(y) (21.2-30)
where
I(y)
A
c
(y)
12
0
a
3
0
_
p
t
_
2
(21.2-31)
which can be interpreted as the local contribution per unit surface area of the body surface
to the total acoustic power. The mean-square time derivative of the surface pressure
and the correlation area are further approximated in terms of turbulent quantities like
turbulent kinetic energy, dissipation rate, and wall shear.
FLUENT reports the acoustic surface power dened by Equation 21.2-31 both in physical
(W/m
2
) and dB units.
Source Terms in the Linearized Euler Equations
The linearized Euler equations (LEE) can be derived from the Navier-Stokes equations
starting from decompositions of the ow variables into mean, turbulent, and acoustic
components, and by assuming that the acoustic components are much smaller than the
mean and turbulent components. The resulting linearized Euler equations for the acoustic
velocity components can be written as
u
ai
t
+ U
j
u
ai
x
j
+ u
aj
U
i
x
j
+
1
p
a
x
i
2
P
x
i
=
U
j
u
i
x
j
u
j
U
i
x
j
. .
L
sh
u
j
u
i
x
j
. .
Lse
x
i
i
t
+

x
j
u
j
u
i
(21.2-32)
where the subscript a refers to the corresponding acoustic components, and the prime
superscript refers to the turbulent components.
The right side of Equation 21.2-32 can be considered as eective source terms responsible
for sound generation. Among them, the rst three terms involving turbulence are the
main contributors. The rst two terms denoted by L
sh
are often referred to as shear-
noise source terms, since they involve the mean shear. The third term denoted by L
se
is often called the self-noise source term, for it involves turbulent velocity components
only.
The turbulent velocity eld needed to compute the LEE source terms is obtained using
the method of stochastic noise generation and radiation (SNGR) [29]. In this method,
the turbulent velocity eld and its derivatives are computed from a sum of N Fourier
modes.
u(x, t) = 2
N
n=1
u
n
cos
_
k
n
x +
n
_
n
(21.2-33)
where u
n
,
n
,
n
are the amplitude, phase, and directional (unit) vector of the n
th
Fourier
mode associated with the wave-number vector

k
n
.
Note that the source terms in the LEE are vector quantities, having two or three com-
ponents depending on the dimension of the problem at hand.
Source Terms in Lilleys Equation
Lilleys equation is a third-order wave equation that can be derived by combining the
conservation of mass and momentum of compressible uids. When the viscous terms are
omitted, it can be written in the following form:
D
Dt
_
D
2
Dt
2

x
j
_
a
2

x
j
__
+ 2
u
k
x
j
x
k
_
a
2

x
j
_
= 2
u
k
x
i
u
j
x
k
u
i
x
j
(21.2-34)
where = (1/) ln
p
po
.
Lilleys equation can be linearized about the underlying steady ow as
u
i
(x, t) = U
i
(x) + u
i
(x, t) (21.2-35)
where u
(x, t) is the turbulent velocity component.

Substituting Equation 21.2-35 into the source term of Equation 21.2-34, we have
S 2
u
k
x
i
u
j
x
k
u
i
x
j
= 2
U
k
x
i
U
j
x
k
U
i
x
j
2
u
k
x
i
u
j
x
k
u
i
x
j
. .
Self-Noise Terms
6
U
k
x
i
U
j
x
k
u
i
x
j
6
u
k
x
i
u
j
x
k
U
i
x
j
. .
Shear-Noise Terms
(21.2-36)
The resulting source terms in Equation 21.2-36 are evaluated using the mean velocity eld
and the turbulent (uctuating) velocity components synthesized by the SNGR method.
As with the LEE source terms, the source terms in Equation 21.2-36 are grouped de-
pending on whether the mean velocity gradients are involved (shear noise or self noise),
and reported separately in FLUENT.
21.3 Using the Ffowcs Williams and Hawkings Acoustics Model
The procedure for computing sound using the FW-H acoustics model in FLUENT consists
largely of two steps. In the rst step, a time-accurate ow solution is generated, from
which time histories of the relevant variables (e.g., pressure, velocity, and density) on the
selected source surfaces are obtained. In the second step, sound pressure signals at the
user-specied receiver locations are computed using the source data collected during the
rst step.
i
Note that you can also use the FW-H model for a steady-state simulation
in the case where your model has a single rotating reference frame. Here,
the loading noise due to the motion of the noise sources is computed using
the FW-H integrals (see Equations 21.2-5 and 21.2-6), except that the term
involving the time derivative of surface pressure (

L
r
in Equation 21.2-6) is
set to zero.
In computing sound pressure using the FW-H integral solution, FLUENT uses a so-called
forward-time projection to account for the time delay between the emission time (the
time at which the sound is emitted from the source) and the reception time (the time at
which the sound arrives at the receiver location). The forward-time projection approach
enables you to compute sound at the same time on the y as the transient ow solution
progresses, without having to save the source data.
In this section, the procedure for setting up and using the FW-H acoustics model is
outlined rst, followed by detailed descriptions of each of the steps involved. Remember
that only the steps that are pertinent to acoustics modeling are discussed here. For
information about the inputs related to other models that you are using in conjunction
with the FW-H acoustics model, see the appropriate sections for those models.
The general procedure for carrying out an FW-H acoustics calculation in FLUENT is as
follows:
1. Calculate a converged ow solution. For a transient case, run the transient solution
until you obtain a statistically steady-state solution as described below.
2. Enable the FW-H acoustics model and set the associated model parameters.
Dene Models Acoustics...
3. Specify the source surface(s) and choose the options associated with acquisition
and saving of the source data. For a steady-state case, specify the rotating surface
zone(s) as the source surface(s).
4. Specify the receiver location(s).
5. Continue the transient solution for a suciently long period of time and save the
source data (transient cases only).
Solve Iterate...
6. Compute and save the sound pressure signals.
Solve Acoustic Signals...
7. Postprocess the sound pressure signals.
Plot FFT...
i
Before you start the acoustics calculation for a transient case, a FLUENT
transient solution should have been run to a point where the transient
ow eld has become statistically steady. In practice, this means that
the unsteady ow eld under consideration, including all the major ow
variables, has become fully developed in such a way that its statistics do
not change with time. Monitoring the major ow variables at selected
points in the domain is helpful for determining if this condition has been
met.
As discussed earlier, URANS, DES, and LES are all legitimate candidates for transient
ow calculations. For stationary source surfaces, the frequency of the aerodynamically
generated sound heard at the receivers is largely determined by the time scale or fre-
quency of the underlying ow. Therefore, one way to determine the time-step size for
the transient computation is to make it small enough to resolve the smallest character-
istic time scale of the ow at hand that can be reproduced by the mesh and turbulence
adopted in your model.
Once you have obtained a statistically stationary ow-eld solution, you are ready to
acquire the source data.
21.3.1 Enabling the FW-H Acoustics Model
To enable the FW-H acoustics model, select Ffowcs-Williams & Hawkings in the Acoustics
Model panel (Figure 21.3.1).
When you turn on Ffowcs-Williams & Hawkings, the panel will expand to show the relevant
elds for user inputs.
Figure 21.3.1: The Acoustics Model Panel
Setting Model Constants
Under Model Constants in the Acoustics Model panel, specify the relevant acoustic pa-
rameters and constants used by the model.
Far-Field Density (for example,
0
in Equation 21.2-1) is the far-eld uid density.
Far-Field Sound Speed (for example, a
0
in Equation 21.2-1) is the sound speed in the
far eld (=
RT
0
).
Reference Acoustic Pressure (for example, p
ref
in Equation 28.10-11) is used to calcu-
late the sound pressure level in dB (see Section 28.10.4: Using the FFT Utility).
The default reference acoustic pressure is 2 10
5
Pa.
Number of Time Steps Per Revolution is available only for steady-state cases that have
a single rotating reference frame. Here you will specify the number of equivalent
time steps that it will take for the rotating zone to complete one revolution.
Number of Revolutions is available only for steady-state cases that have a single rotat-
ing reference frame. Here you will specify the number of revolutions that will be
simulated in the model.
Source Correlation Length is required when sound is to be computed using a 2D ow
result. The FW-H integrals will be evaluated over this length in the depth-wise
direction using the identical source data.
The default values are appropriate for sound propagating in air at atmospheric pressure
and temperature.
Computing Sound on the Fly
The FW-H acoustics model in FLUENT allows you to perform simultaneous calculation
of the sound pressure signals at the prescribed receivers without having to write the
source data to les, which can save a signicant amount of disk space on your machine.
To enable this on-the-y calculation of sound, turn on the Compute Acoustic Signals
Simultaneously option in the Acoustics Model panel.
i
Because the noise computation takes a negligible percentage of memory and
computational time compared to a transient ow calculation, this option
can be used by itself or along with the process of source data le export
and sound calculation. For the latter, computing signals on the y allows
you to see when the signals have become statistically steady so you can
know when to stop the simulation.
When the Compute Acoustic Signals Simultaneously option is enabled, the FLUENT console
window will print a message at the end of each time step indicating that the sound
pressure signals have been computed (e.g., Computing sound signals at x receiver
locations ..., where x is the number of receivers you specied). Enabling this option
instructs FLUENT to compute sound pressure signals at the end of each time step, which
will slightly increase the computation time.
i
Note that this option is available only when the FW-H acoustics model has
been enabled. See below for details about exporting source data without
enabling the FW-H model.
Writing Source Data Files
Although the on-the-y capability is a convenient feature, you will want to save the
source data as well, because the acquisition of source data during a transient ow-eld
calculation is the most time-consuming part of acoustics computations, and you most
likely will not want to discard it. By saving the source data, you can always reuse it to
compute the sound pressure signals at new or additional receiver locations.
To enable saving the source data to les, turn on the Export Acoustic Source Data option
in the Acoustics Model panel. Once this option is selected, the relevant source data at
all face elements of the selected source surfaces will be written into the les you specify.
The source data vary depending on the solver option you have chosen and whether the
source surface is a wall or not. Table 21.3.1 shows the ow variables saved as the source
data.
See Section 21.3.2: Specifying Source Surfaces for details on how to specify parameters
for exporting source data.
Table 21.3.1: Source Data Saved in Source Data Files
Solver Option Source Surface Source Data
incompressible walls p
incompressible permeable surfaces p, u, v, w
compressible walls p
compressible permeable surfaces , p, u, v, w
Exporting Source Data Without Enabling the FW-H Model
You can also export sound source data for use with SYSNOISE without having to enable
the Ffowcs Williams and Hawkings (FW-H) model. You will still need to specify source
surfaces (see Section 21.3.2: Specifying Source Surfaces), as .index and .asd les are
required by SYSNOISE. In addition, you can choose uid zones as emission sources if
you want to export quadrupole sources. To enable the selection of uid zones as sources,
use the
define models acoustics export-volumetric-sources?
text command and change the selection to yes.
SYSNOISE also requires centroid data for source zones that are being exported.
For fan noise calculations, once you have specied the source zones in the Acoustic Sources
panel and you have selected Export Acoustic Source Data from the Acoustic Model panel,
you can export geometry in cylindrical coordinates by using the
define models acoustics cylindrical-export?
text command and changing the selection to yes. By default, FLUENT exports source
zones for SYSNOISE in Cartesian coordinates.
You can then export the centroid data to a .data le using the following text command:
define models acoustics write-centroid-info
Since you will not be using the FW-H model to compute signals, you will not need to
specify any acoustic model parameters or receiver locations. Also, you will not be able to
turn on the Compute Acoustic Signals Simultaneously option in the Acoustics Model panel,
and Acoustic Signals... will not be available in the Solve menu.
21.3.2 Specifying Source Surfaces
In the Acoustics Model panel, click the Dene Sources... button to open the Acoustic
Sources panel (Figure 21.3.2). Here you will specify the source surface(s) to be used in
the acoustics calculation and the inputs associated with saving source data to les.
Figure 21.3.2: The Acoustic Sources Panel
Under Source Zones, you can select multiple emission (source) surfaces and the surface
Type that you can select is not limited to a wall. You can also choose interior surfaces
and sliding interfaces (both stationary and rotating) as source surfaces.
i
The ability to choose multiple source surfaces is useful for investigating the
contributions from individual source surfaces. The results based on the use
of multiple source surfaces are valid as long as there are negligible acoustic
interactions among the surfaces. Thus, some caution needs to be taken
when selecting multiple source surfaces.
In cases where multiple source surfaces are selected, no source surface may enclose any
of the other source surfaces. Otherwise, the sound pressure calculated based on the
source surfaces will not be accurate, as the contribution from the enclosed (inner) source
surfaces is over predicted, since the FW-H model is unable to account for the shading of
the sound from the inner source surfaces by the enclosure surface.
If you specify any interior surfaces as source surfaces, the interior surface must be gen-
erated in advance (e.g., in GAMBIT) in such a way that the two cell zones adjacent to
the surface have dierent cell zone IDs. Furthermore, you must correctly specify which
of the two zones is occupied by the quadrupole sources (interior cell zone). This will
allow FLUENT to determine the direction in which the sound will propagate. When you
rst attempt to select a legitimate interior surface (i.e., an interior surface having two
dierent cell zones on both sides) as a source surface, the Interior Cell Zone Selection
panel (Figure 21.3.3) will appear. You will then need to select the interior cell zones
from the two zones listed under the Interior Cell Zone. Figure 21.3.4 shows an example of
an interior source surface.
Figure 21.3.3: The Interior Cell Zone Selection Panel
Like general interior surfaces, if the source surfaces selected are sliding interfaces, a panel
similar to Figure 21.3.3 will appear that will show the two adjacent cell zones and you
will be asked to specify the zone which has the sound sources.
i
When a permeable surface (either interior or sliding interface) is chosen as
the source surface, other wall surfaces inside the volume enclosed by the
permeable surface that generate sound should not be chosen for the acous-
tics calculation. For example, when running an on-the-y calculation, if
both these surfaces are selected, the sound pressure will be counted twice.
inner fluid
interior source surface
outer fluid
Contours of Static Pressure (pascal) (Time=2.2210e-02)
3.36e+03
-5.47e+03
-5.03e+03
-4.59e+03
-4.14e+03
-3.70e+03
-3.26e+03
-2.82e+03
-2.38e+03
-1.94e+03
-1.50e+03
-1.05e+03
-6.13e+02
-1.71e+02
2.70e+02
7.12e+02
1.15e+03
1.59e+03
2.04e+03
2.48e+03
2.92e+03
Figure 21.3.4: An Interior Source Surface
Saving Source Data
To save the source data, you have to specify the Source Data Root Filename, Write Fre-
quency (in number of time steps), and No. of Time Steps per File in the Acoustic Sources
panel.
The Source Data Root File Name is used to give the names of the source data les (e.g.,
acoustic examplexxxx.asd, where xxxx is the global time-step index of the transient
solution) and an index le (e.g., acoustic example.index) that will store the informa-
tion associated with the source data . The Write Frequency allows you to control how
often the source data will be written. This will enable you to save disk space if the
time-step size used in the transient ow simulation is smaller than necessary to resolve
the sound frequency you are attempting to predict. In most situations, however, you will
want to save the source data at every time step and use the default value of 1.
Since acoustics calculations usually generate thousands of time steps of source data, you
may want to split the data into several les. Specifying the No. of Time Steps per File
allows you to write the source data into separate les for dierent simulation intervals,
the duration of which (in terms of the number of transient ow time steps) is specied
by you. For example, if you specify 100 for this parameter, each le will contain source
data for an interval length of 100 time steps regardless of the write frequency.
You will nd this feature useful if you want to use a selected number of source data les to
compute the sound pressure rather than using all the data. For example, you may want
to exclude an initial portion of the source data from your acoustics calculation because
you may realize later that the ow eld has not fully attained a statistically steady state.
After you click Apply, FLUENT will create the index le (e.g., acoustics example.index),
which contains information about the source data.
i
If you choose to save source data, keep in mind that the source data can use
up a considerable amount of disk space, especially if the mesh being used
has a large number of face elements on the source surfaces you selected.
FLUENT will print out the disk space requirement per time step at the
time of source surface selection if the Export Acoustic Source Data option
is enabled in the Acoustics Model panel.
At this point, if you have chosen to perform your acoustics calculation in two steps, (i.e.,
saving the source data rst, and computing the sound at a later time), you can go ahead
and instruct FLUENT to perform a suitable number of time steps, and the source data will
be saved to the disk. If you have chosen to perform an on-the-y acoustic calculation,
then you will need to specify receiver locations (see Section 21.3.3: Specifying Acoustic
Receivers) before you run the unsteady FLUENT solution any further.
21.3.3 Specifying Acoustic Receivers
In the Acoustics Model panel, click the Dene Receivers... button to open the Acoustic
Receivers panel (Figure 21.3.5).
Figure 21.3.5: The Acoustic Receivers Panel
i
Note that you can also open the Acoustic Receivers panel by clicking the
Receivers... button in the Acoustic Sources or the Acoustic Signals panels.
Increase the No. of Receivers to the total number of receivers for which you want to
compute sound, and enter the coordinates for each receiver in the X-Coord., Y-Coord.,
and Z-Coord. elds. Note that because FLUENTs acoustics model is ideally suited for far-
eld noise prediction, the receiver locations you dene should be at a reasonable distance
from the sources of sound (i.e., the selected source surfaces). The receiver locations can
also fall outside of the computational domain.
For each receiver, you can specify a le name in the Signal File Name eld. These les will
be used to store the sound pressure signals at the corresponding receivers. By default,
the les will be named receiver-1.ard, receiver-2.ard, etc.
Once the receiver locations have been dened, the setup for your acoustic calculation is
complete. You can now proceed to instruct FLUENT to perform a transient calculation
for a suitable number of time steps. When the calculation is nished, you will have
either the source data saved on les (if you chose to save it to a le or les), or the sound
pressure signals (if you chose to perform an acoustic calculation on the y), or both
(if you chose to save the source data to les and if you chose to perform the acoustic
calculation on the y).
If you chose to save the source data to les, the FLUENT console window will print a
message at the end of each time step indicating that source data have been written (or
appended to) a le (e.g., acoustic example240.asd).
21.3.4 Postprocessing the FW-H Acoustics Model Data
At this point, you will have either the source data saved to les or the sound pressure
signals computed, or both. You can process these data to compute and plot various
acoustic quantities using FLUENTs FFT capabilities. See Section 28.10: Fast Fourier
Transform (FFT) Postprocessing for more information.
Writing Acoustic Signals
If you chose to perform the acoustic calculation on the y, you will need to write the
sound pressure data to les. To do so, select Write Acoustic Signals under Options in
the Acoustic Signals panel (Figure 21.3.6) and then click Write. The computed acoustic
pressure will be saved from internal buer memory into a separate le for each receiver
you dened in the Acoustic Receivers panel (e.g., receiver-1.ard).
Reading Unsteady Acoustic Source Data
Computing the sound pressure signals using the source data saved to les is done in the
Acoustic Signals panel (Figure 21.3.6).
Figure 21.3.6: The Acoustic Signals Panel
To compute the sound data, use the following procedure:
1. In the Acoustic Signals panel, select Read Unsteady Acoustic Source Data Files under
Options.
2. Click Load Index File... and select the index le for your computation in the Select
File dialog box. The le will have the name you entered in the Source Data Root
Filename eld in the Acoustic Sources panel, followed by the .index sux (e.g.,
acoustic example.index).
3. In the Source Data Files list, select the source data les that you want to use
to compute sound. Source data les will all contain the specied root le name
followed by the sux .asd.
i
You can use any number of source data les. However, note that you should
select only consecutive les.
4. In the Active Source Zones list, select the source zones you want to include to
compute sound. See Section 21.3.2: Specifying Source Surfaces for details about
proper source surface selection.
5. In the Receivers list, select the receivers for which you want to compute and save
sound.
Optionally, you can click the Receivers... button to open the Acoustic Receivers
panel and dene additional receivers.
6. Click the Compute/Write button to compute and save the sound pressure data. One
le will be saved for each receiver you previously specied in the Acoustic Receivers
panel (e.g., receiver-1.ard).
i
If you enabled both the Export Acoustic Source Data and Compute Acoustic
Signals Simultaneously options in the Acoustics Model panel, you will need
to rst select the Write Acoustic Signals option in the Acoustic Signals panel
after the ow simulation has been completed. If you select the Read Un-
steady Acoustic Source Data Files before writing out the on-the-y data
in such a case, the data will be ushed out of the internal buer memory.
To avoid such a loss of data, you should save the FLUENT case and data
les whenever you begin to do an acoustic computation in the Acoustic
Signals panel. The sound pressure data calculated on the y will then be
saved into the .dat le. Finally, after the on-the-y data is saved, make
sure to change the le names of the receivers before doing a sound pres-
sure calculation with the Read Unsteady Acoustic Source Data Files option
enabled, to avoid overwriting the on-the-y signal les.
i
Note that you can compute and write sound pressure signals only when
the FW-H acoustics model has been enabled. See Section 21.3.1: Export-
ing Source Data Without Enabling the FW-H Model for details about
exporting source data (e.g., for SYSNOISE) without enabling the FW-H
model.
Pruning the Signal Data Automatically
Before the computed sound pressure data at each receiver is saved, it is by default
automatically pruned. Pruning of the receiver data means clipping the tails of the signal
where incomplete source information is available.
The acoustic source data is tabulated from time
0
to
n
. Without auto-pruning, the
receiver register begins receiving the earliest sound pressure signal at
t
0
=
0
+
r
min
a
0
where r
min
is the shortest distance between the source surfaces and the receiver. However,
the receiver will not receive the sound pressure signal from the farthest point on the source
surfaces (r
max
) until the receiver time becomes
t
1
=
0
+
r
max
a
0
From time t
0
to t
1
, the sound accumulated on the receiver register does not include
the contribution from the entire source surface area, and thus the sound pressure data
received during that time is not complete. The same thing occurs during the period from
t
m
=
m
+
r
min
a
0
to
t
n
=
n
+
r
max
a
0
Thus, pruning means clipping the signal on the incomplete ends, from t
0
to t
1
and t
m
to t
n
.
Auto-pruning can be disabled using the define models acoustics auto-prune
text command. Although auto-pruning can be disabled, it is expected that you will use
only the complete sound pressure data.
21.4 Using the Broadband Noise Source Models
Reporting the Static Pressure Time Derivative
The RMS value of the static pressure time derivative (p/t) is available for postprocess-
ing only on wall surfaces, which are at the same time sources of sound, when the FW-H
acoustics model is used.
You can select Surface dpdt RMS in the Acoustics... category only when you specify
at least one wall surface, which is also marked as an acoustic source, in the relevant
postprocessing panels.
Using the FFT Capabilities
Once the sound pressure signals are computed and saved in les, the sound data is
ready to be analyzed using FLUENTs FFT tools. In the Fourier Transform panel (Fig-
ure 28.10.1), click on Load Input File... and select the appropriate .ard le. If the receiver
data is still in FLUENTs memory, then it can directly be processed using the Process
Receiver option. See Section 28.10: Fast Fourier Transform (FFT) Postprocessing for
more information on FLUENTs FFT capabilities.
Plot FFT...
In this section, the procedure for setting up and using the broadband noise source models
is outlined rst, followed by descriptions of each of the steps involved.
The general procedure for carrying out a broadband noise source calculation in FLUENT
is as follows:
1. Calculate a steady or unsteady RANS solution.
2. Enable the broadband noise model and set the associated model parameters.
3. Postprocess the noise sources.
Display Contours...
21.4.1 Enabling the Broadband Noise Source Models
To enable the broadband noise sources models, select Broadband Noise Sources in the
Acoustics Model panel (Figure 21.4.1).
Figure 21.4.1: The Acoustics Model Panel for Broadband Noise
Setting Model Constants
Under Model Constants in the Acoustics Model panel, specify the relevant acoustic param-
eters and constants used by the model. See Section 21.3.1: Enabling the FW-H Acoustics
Model for the denitions of Far-Field Density and Far-Field Sound Speed.
Reference Acoustic Power (for example, P
ref
in Equation 21.2-13) is used to compute
the acoustic power outputs in decibels (dB). The default value is 10
12
. Note that
the units for the reference acoustic power will be dierent in 2D (W/m
2
) and 3D
(W/m
3
) cases.
Number of Realizations is the number of samples used in the SNGR to compute the
averaged source terms of LEE and Lilleys equations. The default value is 200.
Number of Fourier Modes (N in Equation 21.2-33) is the number of the Fourier modes
used to compute the turbulent velocity eld and its derivatives. The turbulent
velocity eld is then used to compute the LEE and Lilleys source terms. The
default value is 50.
21.4.2 Postprocessing the Broadband Noise Source Model Data
The nal step in the broadband noise source modeling process is the postprocessing
of acoustic power and noise source data. The following variables are available in the
Acoustics... postprocessing category:
Acoustic Power Level (dB)
Acoustic Power
Jet Acoustic Power Level (dB) (axisymmetric models only)
Jet Acoustic Power (axisymmetric models only)
Surface Acoustic Power Level (dB)
Surface Acoustic Power
Lilleys Self-Noise Source
Lilleys Shear-Noise Source
Lilleys Total Noise Source
LEE Self-Noise X-Source
LEE Shear-Noise X-Source
LEE Total Noise X-Source
LEE Self-Noise Y-Source
LEE Shear-Noise Y-Source
LEE Total Noise Y-Source
LEE Self-Noise Z-Source (3D models only)
LEE Shear-Noise Z-Source (3D models only)
LEE Total Noise Z-Source (3D models only)

CHP 21

Hochgeladen von

Dokumentinformationen

Originalbeschreibung:

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

CHP 21

Hochgeladen von

Copyright:

Verfügbare Formate

Chapter 21.

Predicting Aerodynamically Generated Noise

is the sound pressure at the far eld (p

, is set to 0.1 in FLUENT based on the calibration of Sarkar and

(x, t) is the turbulent velocity component.

Das könnte Ihnen auch gefallen