ArticlePDF Available

Experimentelle und numerische Untersuchung zur Strömungsakustik einer Staulippe

Authors:

Abstract and Figures

Die Staulippe befindet sich an Fahrzeugen unterschiedlicher Automobilhersteller am Unterboden jeweils vor den Rädern. Ihre akustische Wirkung im Fahrzeuginnenraum wird im Rahmen einer Untersuchung der FH Düsseldorf und BMW mit Druckschwankungen an der Fahrzeugau:sBenseite korreliert, um den Ent stehungsmechanismus von Geräuschen und strömungsinduzierter Schwingungen bereits im Quellbereich besser zu verstehen. Für die Positionie rung der Wandmikrofone und für ein tieferes Verständnis der Strömungstopografie werden stationäre CFD-Rechnungen im Bereich des rotierenden Vorderrades durchgeführt.
Content may be subject to copyright.
Experimental and
Numerical Investigation of
Air Dam Aeroacoustics
Different vehicle manufacturers mount an air dam on the vehicle undercarriage just in
front of the wheels. For better understanding of the noise generation mechanisms and
flow induced vibrations directly at their source region, the acoustical effects in the interior
of the vehicle will be correlated by the University of Applied Sciences Duesseldorf and
BMW with the pressure fluctuations of the exterior of the vehicle. Steady state CFD
calculations in the area of the rotating front wheel will be carried out for the positioning
of the surface microphones and for a better understanding of the flow topology.
RESEARCH
ATZ 07-08I2009 Volume 11154
Aerodynamics
1 Influence on the Interior Acoustics
In addition to design, safety and han-
dling characteristics of automobiles, the
acoustic perception of the car occupants
is becoming of even greater importance.
New experimental and simulation meth-
ods already play a role during the con-
ceptual phase of a vehicle design. It is for
this reason that BMW’s acoustic wind
tunnel is equipped with additional meas-
urement methods for the undercarriage,
Figure 1 [1].
Wind tunnel experiments on their
own are not sufficient in order to under-
stand the influence of engine, roll and
wind noise. Wind tunnel experiments on-
ly draw conclusions on the acoustics of
the air flow, other noise sources stay un-
considered. An additional possibility for
gathering all relevant car-related noise
would be road testing in realistic condi-
tions. In the present paper, acoustical ef-
fects in the interior of the vehicle in order
will be correlated with the pressure fluc-
tuations of the exterior of the vehicle to
understand the noise generation mecha-
nism and flow induced vibrations in their
source region. A special attention will be
drawn on the influence of the air dams,
Figure 1. They are located on the under-
carriage in front of each wheel respective
each wheel housing, Figure 2.
2 Aerodynamic Processes
at Wheel and Wheel Housing
The investigation of aeroacoustic proc-
esses at the wheels and in the wheel
housings gained importance as recently
as in the last 15 years. The high degree of
cavities, edged connection rods and
turned sheets in this area result into
highly uneven and cleft surfaces which
not only cause a remarkable f low drag,
but also represent a exceedingly source
for f low noise. According to [4] wheels,
wheel housings and undercarriage at a
passenger car with a drag coefficient of
cD = 0.3 are responsible for over half of
the aerodynamic drag.
To avoid or to reduce the direct flow
against the wheels and to achieve better
flow around the wheel, the air dam was
fitted on the part of BMW and other auto-
mobile manufacturers. The construction
at the experimental vehicle covers the
complete span of the front wheel hous-
ing. At the rear wheel housing, the air
dam is approximately only 200 mm long
which equates to the width of the wheel.
In the front part of the wheel housing
around the rotating wheel, Figure 3, the
flow sucks upwards and hits the bound-
ary layer f low close to the body on the
edge of the car body. At this location an
unrolling longitudinal vortex is expect-
ed, which probably is disturbed by the
emergent air of the wheelhouse area and
flow separation occurs.
3 Measurement Instrumentation
and Conditions
For Wagners thesis [3] Brüel & Kjær has
kindly provided the detection of dynam-
ic pressure fluctuations with surface mi-
crophones of the types 4949 and 4949B,
Figure 4. Despite the small dimensions
(Ø 20 mm, height 2.5 mm) they are,
The Authors
Prof. Dr.-Ing.
Frank Kameier
is Professor of Fluid
Mechanics and Acous-
tics at the University
of Applied Sciences
Duesseldorf (Germany).
Figure 1: Vehicle undercarriage with the position of the air dam in front of the front wheel [2]
Thomas Wagner
MScEng
was Research Assistant
at the University of
Applied Sciences
Duesseldorf and is now
Research Assistant at
the Technical University
Kaiserslautern (Germany).
Igor Horvat MScEng
is Research Assistant
at the University of
Applied Sciences
Duesseldorf (Germany).
Dipl.-Ing.
Frank Ullrich
is responsible for
Aeroacoustics and the
Acoustic Wind Tunnel
of the BMW Group in
Munich (Germany).
ATZ Peer Review
The Seal of Approval
for scientific articles
in ATZ.
Reviewed by experts from research
and industry.
Received . . . . . . . . . . . . . . . . March 02, 2009
Reviewed. . . . . . . . . . . . . . . . March 16, 2009
Accepted . . . . . . . . . . . . . . . . March 31, 2009
based on a titanium casing, very sustain-
able against wet (corrosion) and oily envi-
ronments. The acoustic sensitivity and
the dynamic range are approximately
equal to traditional ¼” condenser meas-
urement microphones. Figure 4 shows
the approximately selected positions of
the surface microphones. Parallel to the
measurements, further numerical calcu-
lations of the simplified vehicle section
were continued.
At all measurements the surface micro-
phones were provided with a little covering,
protecting the membrane from dirt parti-
cles, as well as with a mounting pad. This to
the outer diameter flattening ring causes a
better flow around the microphone. The so
called self-noise, which is measured by the
microphone itself, is generated by vortex
separation which causes pressure fluctua-
tions at the microphone body. This self-
noise can be minimized by the shaping of
the microphone. On the other hand, a struc-
ture-borne sound isolation between wall
and surface microphone can be achieved
with this indirect placement [6]. For acquisi-
tion of all relevant aeroacoustic data, a mo-
bile data acquisition unit MK II and the soft-
ware PAK of MüllerBBM VibroAkustik Sys-
teme GmbH were used.
For road testing an E92, 320d Coupé ex-
perimental vehicle was provided by BMW
Group. For reproducible measurement re-
sults the motorway A59 between Düssel-
dorf and Leverkusen was used. This route of-
fers good test conditions , due to its new and
homogenous road surface as well as its low
traffic volume.
In forefront of the measurements, a sim-
plified sketch of the wheel housing was
used in CFD calculations with Ansys 11 to
find the best measurement positions for
the surface microphones. Within the steady
state CFD calculations by using a shear
stress transport (SST) model it was paid at-
tention on the boundary conditions of the
relative movement between vehicle and
road as well as on the rotation of the wheel.
Thereby only a small part of the vehicle was
design-engineered and meshed for the sim-
ulation. This enables focusing the mesh on
the main area of influence and an efficient
calculation.
To verify the boundary layers and the nu-
merical simulation under conditions close
to reality the boundary conditions, the cal-
culation area and the velocity distributions
are shown in Figure 5.
Figure 2: Bifid air dam of the model BMW E92 [3]
Figure 4: Surface microphone (left) and its positions on the vehicle (right, air dam not mounted),
compare with Figure 8
Figure 3: Schematic view of the longitudinal vortex in the area of the wheel housing with
rotating wheel and with moving street – view from the side (a) and from behind (b) [5]
RESEARCH
ATZ 07-08I2009 Volume 11156
Aerodynamics
4 Measurement Results
of the Pressure Fluctuations
Figure 6 shows a comparison of the sound
pressure spectra of the vehicle interior at
a velocity of 150 km/h with and without
an air dam. Harmonic tonal frequency
components of the rotational wheel speed
occur only without the air dam. Without
the air dam, even more tonal frequency
components are clearly transferred in the
interior of the vehicle. Differences at low
frequencies in the signals of the surface
microphones mounted in the wheel hous-
ing occur between the spectra with and
without air dam, Figure 7.
5 Correlation of the Wall Pressure
Fluctuations
By using a correlation analysis it is possi-
ble to differ between acoustic pressure
fluctuations with the speed of sound, as
their propagation velocity and aerody-
namic pressure fluctuations which are
not obligatory relevant for noise. Impor-
tant for the interpretation of the phase
characteristics in accordance to [7] is the
microphone placement in the wheel
housing, Figure 4.
With the knowledge of the distance Δx
between the microphone positions, it is pos-
sible to calculate the speed of propagation
of the pressure fluctuation from the phase
angle of the cross-spectrum. With a suffi-
cient level of the coherence of two adjacent
pressure transducers, the linear phase shift
at constant speed of propagation (non dis-
persive propagation process) can be allocat-
ed. In Figure 8 the coherence and the phase
angle of the microphone signals on posi-
tion 2 to the reference channel above the
wheel housing 4 is shown (blue: measure-
ment with air dam, red: measurement with-
out air dam).
The calculation of the time delay results
by means of the phase slope [8]. The time de-
lay yields from:
tverz =
Δϕ
___
ω
=
Δϕ°
____
360°
· 2π
_______
Δf · 2π
Eq. (1)
With tverz and the distance of the micro-
phones the velocity v = Δx/tverz results. The
speeds of propagation from the phase of
the characteristics shown in Figure 8 rep-
resents between 500 and 800 Hz with Δ
ϕ
= 180°, Δf = 300 Hz and Δx = 0.52 m ap-
proximately v = 324 m/s (the speed of
sound at 5 °C is 334.5 m/s) and between
1700 and 1800 Hz with Δ
ϕ
= 360°, Δf = 60
Hz and Δx = 0.52 m approximately v = 31
m/s (driven speed: approximately 42 m/s).
Figure 5: Boundary conditions of the CFD calculation
Figure 6: Sound pressure levels in the vehicle interior with (black) and without
(green) air dams, road drive, v = 150 km/h
Figure 7: Wall pressure fluctuation levels in the wheel housing with (black) and without
(green) air dams, test drive, v = 150 km/h
ATZ 07-08I2009 Volume 111 57
Hence the speed of propagation is well-
defined between high values near the
speed of sound and lower levels in the
range of the travelling speed of the vehi-
cle. On the one hand, a steep slope of the
phase angle indicates a slow speed of
propagation, in comparison to the speed
of sound. It will be generated by the flow
itself respectively by turbulent pressure
fluctuations.
On the other hand, a minor slope in-
dicates that it concerns with noise prop-
agation in terms of acoustic pressure
fluctuations. In addition the propaga-
tion direction can be detected by means
of a positive or negative slope: With a
positive slope the sound arrives with a
certain delay at the second microphone.
The propagation occurs from the rear to
the front – opposite to the outer f low
around the vehicle and therefore is
dominated by the rotating wheel. It is
not possible to measure this phenome-
non in a wind tunnel with a non-rotat-
ing wheel. The velocity of the vortex rope
matches the slip with 74 % compared to
the circumferential speed of the wheel.
6 Presentation of Vortex Structures
with the Q-Criterion
Figure 9 shows the extensive velocity dis-
tribution around the rotating wheel
both with and without air dam. In order
to compare vortex pattern from the nu-
merical CFD calculation with the topolo-
gy of the flow in accordance to the sche-
matic sketch of Figure 3, an extensive
vector analytical computation of the
flow velocity gradients is necessary.
In order to visualise vortices, the Q cri-
terion in accordance to [9] was imple-
mented in the Ansys CFD post process-
ing, so that coherent vortex structures of
the velocity gradients are visible [10].
Generally, a shear strain is specified as a
gradient of the velocity:
u
___
x
v
__
x
w
___
x
grad u =
u
___
y
v
__
y
w
___
y
=
ui
___
xj
u
___
z
v
__
z
w
___
z
Eq. (2)
This gradient can be converted with fol-
lowing formulas:
S
_ =
1
__
2
· (grad u + grad Tu) and
=
1
__
2
· (grad u – grad Tu) Eq. (3)
to the symmetric shear rate tensor S
_ and
the antimetric vorticity tensor
:
grad u = S
_ +
Eq. (4)
The Q criterion is derived out of both of
these tensors:
Q =
1
__
2
·
(
|
2| – |S
_2|
)
> 0 Eq. (5)
Q =
1
__
2
·
[
(
u
___
x
)
2 +
(
v
__
y
)
2 +
(
w
___
z
)
2
]
[
v
__
x
u
___
y
+
w
___
x
u
___
z
+
v
__
z
w
___
y
]
Eq. (6)
Figure 10 shows the vortex patterns calcu-
lated with the Q criterion. In addition to
the schematic illustration of Hucho [5],
Figure 8: Coherence and the phase angle of the cross-spectrum of the wall pressure
fluctuations position 2 towards position 4 (compare with Figure 4); with (blue) and
without (red) air dam measurements
Figure 9: Velocity distribution wit and without air dam of the simplified wheel housing model
 
 
 
 
 
 
 
 
Figure 10: By Q-criterion calculated vortex pattern with and without air dam
RESEARCH
ATZ 07-08I2009 Volume 11158
Aerodynamics
which shows the vortex system without
an air dam in Figure 3, Figure 10 displays
the vortex system with an air dam. With
an air dam, the vortex is pushed from
the upper side of the wheel down to the
driving surface and produces here fore
less pressure fluctuations at the carriage
of the vehicle. This has a positive effect
on the acoustics.
Whilst the flow topology with and
without the air dam is almost the same
in the area of the undercarriage, there
are obvious differences at the exterior be-
hind the wheel housing. At this area a
longitudinal vortex develops with a de-
mounted air dam. At the shown simula-
tion this vortex is not present. Probably
in consequence of the diverted air flow
by the air dam, which makes sure that
only a small flow rate is conducted
through the gap between the wheel and
the carriage at the wheel housing.
7 Summary
Within an experimental and numerical
investigation by the University of Ap-
plied Sciences Duesseldorf and BMW
the acoustic effect of the air dam of a
BMW 3-Series vehicle was investigated.
Both realistic drive tests on public roads
as well as numerical calculations in the
area of the front rotating wheel were
performed. One of the major aspects
were on the one hand the correlation of
the pressure fluctuations at the exterior
and the wheel housing with noise in-
side the car and on the other hand the
illustration of the flow field with CFD
calculations. To show the influence of
the f low velocity and to gain better un-
derstanding of the noises and the flow
induced vibration directly at the source,
the vehicle was equipped with seven
surface microphones at the exterior and
two measurement microphones in the
interior. The drive tests were performed
at various velocities with and without a
mounted air dam.
At the spectra between 450 and 600
Hz as well as between 1200 and 1400 Hz
without an air dam, an increased level of
3 to 5 dB was measured. The frequency
components appear in a distance of 21
Hz (matching with the wheel speed or-
der), which were not-existant with a
mounted air dam. The reason for these
frequency components is the incident air
flow of the tire profile. The mount of the
air dam just in front of the wheel hous-
ing makes sure that the air f low is de-
flected away from the wheel and the un-
dercarriage. Through this air flow deflec-
tion, different pressure distributions re-
sult especially at the wheel housing,
causing a lower sound pressure level for
a frequency range below 1000 Hz. A com-
pletely different situation is present at
the measurement points directly in front
and behind the air dam. The air dam rep-
resents a fluid dynamical barrier, where
the air flow dams up and at which lead-
ing edge the wake vortices separate.
These separations probably represent the
reason for the higher pressure f luctua-
tions at this area. On the basis of the
spectograms shown by Wagner [3] you
can draw the conclusion, that not only
rotation speed dependent activities, but
also structural acoustic effects in the
wheel housing are responsible for the
noise generation mechanisms. Distinct
differences between the measurements
with and without air dam, at least for
the area of the wheel housing, are evi-
dent.
The observation of the correlation
analysis provides for neighbouring meas-
urement points coherences which par-
tially raise to 0.9 for a frequency range
between 700 and 1700 Hz. With the slope
of the phase angle, conclusions can be
drawn on the kind of sound propagation.
The measured pressure fluctuations are
composed by turbulent (caused by the air
flow) and acoustic pressure fluctuations.
Due to the determined slip of 74 % of the
driving speed and the propagation direc-
tion the measured pressure fluctuations
can be allocated to the rotating wheel.
Measurements in the acoustic wind tun-
nel with a non-rotating wheel cannot
cover these effects.
With the aid of the simulation tool
Ansys CFX the flow topology around the
unrolling wheel was pictured. The air
dam makes sure that the air f low is de-
flected. The stagnation point is dis-
placed to a position further away from
the wheel just in front of the air dam.
Herewith, a direct f low against the
wheel is avoided. The effect can be
shown on the basis of f low lines. Beside
the velocity distribution the visualisa-
tion of vortex pattern with the appro-
priate allocation of velocity gradients is
expedient. The results show clear analo-
gies to the theory of Hucho [5]. At both
vehicle configurations horseshoe vorti-
ces can be found. At the simulations
without an air dam a longitudinal vor-
tex develops on the exterior behind the
wheel housing. Under circumstances
this longitudinal vortex provides addi-
tional fluid structure interactions,
which can inf luence the sound in the
vehicles interior negatively.
References
[1] Kaltenhauser, A.; Kolb, S.; Ullrich, F.; Polansky, L.:
Neue Modulwaage im Akustikwindkanal von
BMW. In: ATZ 108 (2006), Nr. 12, S. 1026–1037
[2] Ullrich, F.: Aeroakustik: Neue Potenziale für die
Innengeräuschoptimierung. Haus der Technik
Essen, 4. Tagung Aeroakustik, Wildau, 2006
[3] Wagner, Th.: Experimentelle und numerische
Untersuchung zur Strömungsakustik der Staulippe
eines 3er BMWs. Master Thesis, Fachhochschule
Düsseldorf, 2008
[4] Garrone, A.; Masoero, M.: Car Underside, Upper-
body and Engine Cooling System Interactions and
their Contribution to Aerodynamic Drag. SAE-Paper
860212, Warrendale, Pa., USA, 1986
[5] Hucho, W. (Hrsg.): Aerodynamik des Automobils.
5. Auflage, Vieweg-Verlag, Wiesbaden, 2005
[6] N. N.: Product Data Sheet, Automotive Surface
Microphones – Types 4949 and 4949B. ATZ/MTZ-
Konferenz Akustik – Akustik zukünftiger Fahrzeug-
und Antriebskonzepte, Stuttgart, Mai 2006, http://
www.bksv.com/doc/bp2055.pdf
[7] Bendat, J. S.; Piersol, A. G.: Engineering Applica-
tions of Correlation and Spectral Analysis. New
York, USA, 1993
[8] Kameier, F.: Experimentelle Untersuchung zur
Entstehung und Minderung des Blattspitzen-
Wirbellärms axialer Strömungsmaschinen. Disser-
tation, TU Berlin, VDI-Fortschritt-Berichte, Reihe 7,
Nr. 243, VDI-Verlag, Düsseldorf, 1994
[9] Spille-Kohoff, A.: Das Q-Kriterium. CFX Berlin
Software GmbH, interner Bericht, Berlin, Juni 2006
[10] Fröhlich, J.: Large-Eddy-Simulation turbulenter
Strömungen. Teubner-Verlag, Wiesbaden, 2006
ATZ 07-08I2009 Volume 111 59
... Wagner nutzt in [2] eine CFD-Simulation zur Identifikation der optimalen Messpositionen für Brüel&Kjaer Oberflächenmikrofone Typ 4949 für eine experimentelle Untersuchung zur Strömungsakustik der Staulippe an einem 3er BMW. Die in Abb. 3 gezeigte Staulippe stromauf der Radkästen ist ein akustisches Bauteil, deren Funktion erst durch die Arbeit von Wagner vollständig geklärt wird [2], [4]. ...
... Wagner nutzt in [2] eine CFD-Simulation zur Identifikation der optimalen Messpositionen für Brüel&Kjaer Oberflächenmikrofone Typ 4949 für eine experimentelle Untersuchung zur Strömungsakustik der Staulippe an einem 3er BMW. Die in Abb. 3 gezeigte Staulippe stromauf der Radkästen ist ein akustisches Bauteil, deren Funktion erst durch die Arbeit von Wagner vollständig geklärt wird [2], [4]. Abb. 5 zeigt die Strömungssimulation einer Zylinder-PlatteKonfiguration, wie Sie auch von [4], [5], [6] behandelt wird. ...
Different vehicle manufacturers mount an air dam on the vehicle undercarriage just in front of the wheels. For better understanding of the noise generation mechanisms and flow induced vibrations directly at their source region, the acoustical effects in the interior of the vehicle will be correlated by the University of Applied Sciences Duesseldorf and BMW with the pressure fluctuations of the exterior of the vehicle. Steady state CFD calculations in the area of the rotating front wheel will be carried out for the positioning of the surface microphones and for a better understanding of the flow topology.
Conference Paper
Relevant results are shown from wind tunnel tests on a 1:1 scale model equipped with internal balances which allow measurement of the individual drag contributions from the car underside and upper body. Test results also show effect and interactions due to configuration changes. The model's aerodynamic flow field was examined through surface pressure distributions and flow visualization. Wake surveys, airflow between the model underside and the wind tunnel floor, and base pressures are also investigated. (from authors' abstract)
Article
Das Windgeräusch spielt eine entscheidende Rolle beim Innengeräuschkomfort. Bis jetzt wird durch Feinschliffmaßnahmen das Windgeräusch so optimiert, dass keine subjektiv wahrnehmbaren Störgeräusche auf die Insassen wirken. Um in der Konzeptgestaltung eines Fahrzeugs die entscheidenden Komponenten auszulegen, ist der erweiterte Einsatz von neuen Versuchs- und Simulationsmethoden erforderlich. Für den Akustikwindkanal der BMW Group wurde dazu in Zusammenarbeit mit Horiba ATS eine Modulwaage entwickelt und installiert, die für die Lösung von aeroakustischen Aufgaben völlig neue Möglichkeiten eröffnet.
Article
The book examines applications of correlation and spectral analysis bridging the gap between the engineering measurements and theoretical results from analytical models. Basic principles of correlation and spectral density analysis based on calculus, Fourier series, and the complex variable theory; procedures for analyzing single input/output relationships; time delay and phase lag estimates; and identification of multiple propagation paths and velocities for dispersive and nondispersive media are presented. Finally, the analysis of multiple input/output applications of multiple and partial coherence functions is given along with the practical statistical error analysis formulas for computing spectral density functions, coherence functions, and frequency response functions.
aeroakustik: Neue Potenziale für die Innengeräuschoptimierung
  • F Ullrich
Ullrich, F.: aeroakustik: Neue Potenziale für die Innengeräuschoptimierung. haus der Technik essen, 4. Tagung aeroakustik, Wildau, 2006
Experimentelle und numerische UUntersuchung zur Strömungsakustik der Staulippe eines 3er BMWs
  • Wagner
  • Th
Experimentelle Untersuchung zur Entstehung und Minderung des Blattspitzen-Wirbellärms axialer Strömungsmaschinen
  • F Kameier
  • Kohoff Das
spille-Kohoff, a.: Das Q-Kriterium. cFX Berlin software Gmbh, interner Bericht, Berlin, Juni 2006
  • F Kameier
Kameier, F.: experimentelle Untersuchung zur entstehung und Minderung des Blattspitzen-Wirbellärms axialer strömungsmaschinen. Dissertation, TU Berlin, VDI-Fortschritt-Berichte, Reihe 7, Nr. 243, VDI-Verlag, Düsseldorf, 1994