NASA Technical Reports Server (NTRS) 20030000915: Time-Accurate Simulations and Acoustic Analysis of Slat Free-Shear-Layer. Part 2

14 Pages·2 MB·English

by  NASA Technical Reports Server (NTRS)

• .m at, •
AIAA-2002-2579
TIME-ACCURATE SIMULATIONS AND ACOUSTIC ANALYSIS OF
SLAT FREE-SHEAR-LAYER: PART II
Mehdi R. Khorrami*
Bart A. Singer*
David P. Lockard**
NASA Langley Research Center
MS 128,Hampton, VA
1. Introduction
Abstract
Experimental studies in both the US and
Europe by researchers such as Meadows et al., 1
Unsteady computational simulations of a
multi-element, high-lift configuration are Hayes et al.,-_Dobrzynski et al.,3'4and Davy and
Remy 5have verified that the high-lift devices are
performed. Emphasis is placed on accurate
significant contributors to airframe-radiated
spatio-temporal resolution of the free shear layer
in the slat-cove region. The excessive dissipative sound in the mid- to high-frequency range. In
effects of the turbulence model, so prevalent in particular, the leading-edge slat and the side
previous simulations, are circumvented by edges of the flaps have been identified as
dominant noise sources. NASA, in collaboration
switching off the turbulence-production term in
the slat cove region. The justifications and with industrial and academic partners, has
physical arguments for taking such a step are embarked on a major research pro_am to
explained in detail. The removal of this excess enhance our fundamental understanding of
airframe noise sources and then to apply this
damping allows the shear layer to amplify large-
knowledge to develop effective noise-reduction
scale structures, to achieve a proper non-linear
technologies that do not compromise
saturation state, and to permit vortex merging.
The large-scale disturbances are self-excited, and aerodynamic efficiency. The present
unlike our prior fully turbulent simulations, no computational study is part of a larger effort to
understand the relevant aeroacoustic flow
external forcing of the shear layer is required. To
obtain the farfield acoustics, the Ffowcs Williams features associated with a leading edge slat in
high-lift settings. Specifically, attention is
and Hawkings equation is evaluated numerically
directed towards accurate numerical simulations
using the simulated time-accurate flow data. The
present comparison between the computed and of the complex unsteady flow field in a slat-cove
measured farfield acoustic spectra shows much region.
better agreement for the amplitude and frequency
In a series of in-house experiments conducted
content than past calculations. The effect of the
angle-of-attack on the slat's flow features and in the Low Turbulence Pressure Tunnel (LTPT)
radiated acoustic field are also simulated and at NASA Langley Research Center (LaRC),
presented. aeroacoustic studies of a generic, Energy
Efficient Transport (EET) wing were performed.
The three-element, high-lift configuration was
comprised of a slat, a main element, and a flap
(Fig.l). The model's stowed chord was
*Research Scientist, Computational Modeling and Simulation Branch, Associate Fellow AIAA
+Assistant Branch Head, Computational Modeling and Simulation Branch. Senior Member AIAA
**Research Scientist, Computational Modeling and Simulation Branch, senior Member AIAA
Copyright ©2002 bythe American Institute of Aeronautics and Astronautics, Inc. No copyright isasserted inthe
United States under Title 17.U.S. Code. The U.S. Government has aroyalty-free license to exercise allrights under
the copyright claimed herein for Governmental Purposes. All other rights are reserved bythe copyright owner.
1
American Institute Of Aeronautics And Astronautics
approximatetelynpercenrtelativetothemean
increase as the angle of attack is decreased. PIV
aerodynamcihcordofaBoeing757.Acoustic
measurements of differing slat flow fields by
measuremewnetsreobtaineudsingamicrophone Paschal et al.1°and Takeda eta]. 9'll also exhibited
array.Detaileddiscussioonfthearraylocation
marked changes with variations in the angle of
andorientatioinntheLTPTfacilitycanbefound
attack. More precisely, they found that at
in reference6sand7.Theexperimenwtsere
relatively high angles of attack, the fluctuating
carriedoutatseveradlistincatngleosfattacka,
velocity field emerging through the slat gap is
flapdeflectioanngleof30°andaslatdeflection
confined to a region of narrow spatial extent that
angleof30°.Duringwindtunneelntrieisn1998
is adjacent to the lower segment of the slat's
and 1999,both aerodynamiacndacoustic
wake. At low angles of attack, the fluctuating
measuremewntesreobtainedT.he1998entry field becomes more energetic and is spread over
involveda part-spafnlapthathada slightly
most of the gap width. The energetic structures
differenptrofilethanthefull-spafnlapemployed pumped through the gap appear to be mostly
in the 1999 entry. However,important
large-scale co- or counter-rotating vortex pairs.
parametesruschastheslat'sgeometrsye, ttings, Clearly, a significant change in the dynamics of
andaerodynamloicadingwereunchangefodr
the slat-cove flow field occurs as the angle of
bothentriesa, ndthecorrespondinragdiated
attack is reduced. The underlying intricacies
soundfields(exceptfor minordifferences)
responsible for this flow alteration as well as any
remainendearlyidentical.
links that they may have to the increase in the
radiated acoustic field still remain unresolved and
Samplme icrophonaerraymeasuremeonfts
deserve a careful study.
theEETslatareshowninFig.2. Thedisplayed
spectraarein 1/12th-octabvaends.Theflow
The present paper continues our effort
Machnumbeirs0.2,corresponditnogatypical towards understanding noise sources associated
approacchonditiona,ndtheReynoldnsumber
with a leading-edge slat in a high-lift setting.
(baseodnthestowedchordi)s7.2million.The
Our previous research focused on accurate
measuremesnhtsowhighsoundlevelsin the
Unsteady Reynolds Averaged Navier-Stokes
lowerfrequencryangefollowedby a gradual (URANS) simulations of the slat-cove flow field
dropinthelevelsasthemid-frequenrcayngeis
and the computations of the resulting acoustic
approacheIdn.thevicinityof50kHz, the spectra farfield. 7These simulations were performed for a
display a rise in sound amplitude that seems to be single angle of attack of 8degrees. In reference 7
tonal in nature. This high-frequency tonal feature (hereafter referred to KSB), the amplification of
was studied in detail by Khorrami, et al.8 and instabilities by the slat free-shear layer was
Singer et al.6 who found vortex shedding from shown to produce the radiated noise in the lower
the finite-thickness trailing edge to be its source. frequency range. Those earlier simulations,
The occurrence of vortex shedding at a slat conducted in a fully turbulent mode, proved to be
trailing edge was confirmed by Takeda et al.9 overly dissipative, and, thus, artificially
using the Particle Image Velocimetry (PIV) decreased the radiated acoustic signature. In the
technique to map the flow field of a European current study, we present a simple strategy to
designed high-lift model. Although this tone is partially remedy this shortcoming. The
significant at model scale, its importance for a effectiveness of the approach is shown via
full-scale aircraft is likely to be diminished numerical simulations and a comparison between
because of the smaller trailing-edge bluntness the computed sound field and acoustic array
relative to boundary-layer thickness on a full- measurements. Here we also extend the previous
scale slat. Even if the tone exists at full scale, its simulations, explaining the effect of angle-of-
source is now understood and strategies for its attack variation on the shear-layer instability
elimination are straightforward. modes, the dynamics of the cove flow field, and
the radiated acoustic field.
More central to our current effort is the
behavior of the sound sources in the 1,000Hz to
10,000Hz frequency band range. In this range,
the sound levels exhibit a gradual but noticeable
American Institute Of Aeronautics And Astronautics
2. Computational Procedure where turbulence production is turned off. We
emphasize that convection of the existing
Based on the EET model tested, the current turbulence field into, out of, or within this region
simulations are performed for two-dimensional is uninhibited. A short justification for our
flows. A complete discussion of the numerical approach follows.
approach such as flow solver, high-lift geometry,
grid distribution, and the expected numerical 3.2 Physical Justifications
accuracy was provided by KSB and will not be
repeated. Moreover, the computational Because of a strong favorable pressure
framework of URANS plus Ffowcs Williams and gradient and a short travel distance, the slat
Hawkings used in the present study to obtain the boundary layer between the leading edge
farfield acoustic noise was explained in detail in stagnation point and cusp experiences rapid
references 6-7, and thus omitted here. Following acceleration and is extremely thin. For all
the work of KSB, the turbulence model employed practical purposes, at the point of separation
is the 2-equation Shear Stress Transport (k-a_) (cusp), the boundary layer is laminar and in all
model of Menter 12with one major alteration as likelihood will not be fully turbulent, even up to
explained in the following sections. flight Reynolds numbers. Also, the separated free
shear layer is initially laminar. In such free shear
3. Zonal Approach flows, the overall dynamics of the flow field are
governed by the growth of the large-scale
3.1 Past Deficiencies structures that are the manifestation of the
eigenmodes of the system. The underlying
The fully turbulent simulations of KSB instability mechanism is inviscid (inflectional) in
required explicit forcing of the shear layer to nature; hence, simulation of the unsteady
excite and maintain the large-scale structures. evolution of these structures should not be
The location of the forcing was on the slat encumbered by the introduction of an eddy
surface two to three local boundary layer heights viscosity designed for statistically steady fully
away from the cusp. The amplitude of the turbulent flow. The growth rate and frequency of
forcing corresponded to 3% of the freestream the large-scale structures depend on the
velocity. Figure 3 shows a snapshot of the magnitude of the vorticity across the shear layer
spanwise vorticity field in the slat cove area for and the thickness of the layer. Although the
the fully turbulent forced case. The vorticity overall loading of the slat sets the magnitude of
contours clearly display the spatial location of the the vorticity shed at the cusp, the shear-layer
free-shear layer. Notice that the shear layer is a thickness is one of the few parameters that
good amplifier of the initial perturbations that depends on the viscosity and the state of the
grow rapidly, roll-up the shear layer and form boundary layer at the point of separation.
discrete vortices. Unfortunately, the fully Nevertheless, whether one assumes a turbulent or
turbulent computations were overly diffusive, laminar initial state, only a minor shift in the
causing rapid dissipation of the rolled-up vortices frequency of the rolled-up vortices is expected as
within a short spatial distance (see Fig. 3). To long as the mean shear layer thickness does not
circumvent these excessive diffusive effects (due change appreciably. An important aspect of the
to the turbulence model), a simple remedy based slat flow field is the recirculating zone in the
on physical arguments is advocated and pursued. cove area. The velocity magnitudes in this zone
It is argued that, in the cove region, the are relatively small. Hence, the encountered flow
established flow field is quasi-laminar but highly field has a low Reynolds number. As in the case
unsteady. Accordingly, the production term of a shallow cavity, some, if not the majority, of
associated with the turbulence model is switched the large scale structures become trapped in this
off in a limited zone that encloses the cove area.
zone, setting up an unsteady flow field. Although
Figure 4 presents the grid distribution in the the established recirculating flow may not be
vicinity of the leading edge slat. The grid lines truly laminar, it is not fully turbulent, either.
shown by a lighter color shading (mainly the Based on the above observations, we surmise that
cove area) highlight the extent of the region the cove flow field (including the free shear
American Institute Of Aeronautics And Astronautics
layer),drivenbythelarge-scalseu,bstantially
most of the vortices turn inward and get trapped
inviscidstructuresb,ehavesin anunsteady,
in the cove recirculating flow field. Once
quasi-laminmarannerI.n thesecircumstances,
trapped, these vortices move very slowly towards
theeffecot faturbulencmeodeilsneutraaltbest
the slat cusp. Their proximity to the slat bottom
anddetrimentianlmostinstanceUs.nderthese
surface cause the boundary layer to separate,
uniquesituationsth,ediffusiveeffectsof the
forming vortices of opposite sign vorticity. The
turbulencmeodeml aybebypassekdnowingfull
growing disturbances of the shear layer interact
wellthatthepertinenfetatureosftheflowfield
with the arriving vortices and thus establishing a
arenotalteresdignificantly.
feedback loop. A feature worth noticing is the
absence of vorticity in the center of the
4. Results and Discussions recirculating zone. Contrary to the simulations of
KSB, occasionally a vortex does escape through
Following the work presented in KSB, the the slat gap, and the released vorticity is always
numerical results are normalized with respect to of positive sign. However, the cycle appears
the stowed chord and free stream speed of sound, random, and we have been unable to discern any
density, and molecular viscosity. Similarly, the recognizable pattern. For the present angle of
current nondimensional time step of At = attack, vortices that do escape are severely
4.116xl 0-4 (corresponding to 200 points per deformed by the accelerating local flow before
period for a 7.5kHz model-scale signal) matches leaving the gap very close to the bottom surface
the time step employed by KSB. All simulations adjacent to the trailing-edge wake. This behavior
are performed for a Mach number M- 0.2 and a closely mimics the PIV measurements of gap
Reynolds number Re = 7.2×106 . flow unsteadiness reported in references 9-11.
4.1 URANS Simulations Better insight into the general migration
path of the shear-layer large-scale structures is
To provide a direct comparison with the gained from the averaged flow field. A long-time
previous computations and to understand the average of the spanwise vorticity field is plotted
effects of angle-of-attack variation, simulations in Fig. 6. In excess of 30,000 time steps were
for three angles of 8, 6, and 4 degrees were used to generate the average. To put the length of
performed. The selected angles encompass the data record in perspective, the convective
typical aircraft approach conditions when travel time between the slat's cusp and its trailing
airframe noise is loudest. edge is approximately 1,200-1,300 time steps.
Several prominent features are worth noting in
4.1.1 EigJatDegree Case Fig. 6. The vorticity contours highlight the
presence and location of the cove free shear
A sample plot of the instantaneous layer. The layer is well defined and mildly
spanwise vorticity field from the partially laminar spreading. The large-scale (instability modes)
simulation at 8-degree angle of attack is shown in structures follow migratory paths that fall within
Fig. 5. In contrast to the fully turbulent a fairly narrow spatial band. Vortices ingested by
simulations (Fig. 3), the current computations the recirculating zone travel within a ring that
display extremely complex and highly nonlinear borders the slat's bottom surface and the shear
flow dynamics. In Fig. 5, important stages such layer interior side, leaving the zone's center
as shear layer undulation, roll-up, and the devoid of any significant levels of vorticity.
formation of discrete vortices are vividly Inside the gap, adjacent to the slat surface,
depicted. It is emphasized here that unlike the moderate levels of vorticity are indicated. The
fully turbulent case, the shear layer is self- large width of this layer precludes its association
exciting and no external forcing is used. Absent with the vorticity of the local boundary layer.
in Fig. 5 is the premature dissipation of the large- The magnitude of vorticity in the noted layer
scale structures. In a significant departure from suggests that a fair amount of cove generated
Fig. 3, the vorticity field shows the convection flow disturbances pass through the gap
and coalescence of the vortices as they approach confirming the trends deduced from the
the reattachment point. At the reattachment point, experimental measurements. 9-_ Finally, observe
American Institute Of Aeronautics And Astronautics
thatnegativesignvorticityis confinedto the
vorticity through the gap. Recall that in the 8-
boundarlyayerontheslatbottomsurfaceT.he
degree case only positive vorticity was observed
boundary-latyheicrknesinscreasemsoderatealys
to leave the gap. Figure 7 clearly shows the
theslatcuspisapproached.
ejection process for two co-rotating vortices
(carrying negative vorticity) and some positive-
4.1.2
Six Degree Case vorticity lumps. Note that the vortices are only
moderately deformed, and, unlike at the higher
An instantaneous snapshot of the vorticity angles, they do maintain their structures.
field for the 6-degree case is shown in Fig. 7. Moreover, the ejected vorticity field is spread
Although the angle of attack is reduced a mere over a significant portion of the gap width. Both
two degrees from the previous case, a remarkable of these observations tend to corroborate the PIV
shift in the underlying flow dynamics takes place. measurements obtained at angles of attack of 4
The particular frame displayed in Fig. 7 was and 5 degrees by Paschal et al._°and Takeda et
chosen to highlight some of the important al.9,11
changes. The first distinguishable feature is the
presence of a very large and strong vortex of The long-time averaged spanwise vorticity
positive sign vorticity near the center of the field for the 6-degree case is shown in Fig. 8.
recirculating zone. Unlike the 8-degree case (Fig. More than 35,000 time steps were used in the
5), within this zone, some of the trapped shear- averaging process. Figure 8 suggests that the
layer vortices amalgamate and emerge as the prominent features of the averaged field are
center vortex. Once established, the vorticity in different than those displayed in Fig. 6. Except
this vortex is constantly replenished by the newly for near the slat cusp, the vorticity contours
ingested shear-layer vortices. The recirculation provide evidence of a much diffused shear layer.
vortex possesses a somewhat jittery motion Due to large amplitude and highly erratic motion
resulting in very irregular annular paths. The of the recirculation vortex, the paths of the shear-
presence and the movement of the center vortex layer vortices are no longer confined to a band of
significantly modify the shear-layer limited spatial extent. Moreover, a fair amount of
development. A prominent effect is a more negative sign vorticity is released at the cusp; this
severe boundary layer separation on the slat negative vorticity then becomes entangled within
bottom surface with the consequence of forming the shear layer. Therefore, in a long-time
larger and stronger vortices of opposite sign averaging process, substantial spreading and
vorticity. In most instances, the passage of these cancellation of the shear layer vorticity field
vortices over the cusp region disrupts the normal occurs. Similar arguments can be applied to
development of the shear layer causing instant explain the lack of high vorticity levels inthe gap
roll-up of the separated layer. In other words, the region. On the other hand, the concentrated
mechanism for generating the shear-layer discrete region of vorticity at the center of the
vortices is toggled between promoting growth of recirculating zone (Fig. 8) maps the extent of the
convective instability modes and forcing the area where the center vortex is most active.
instant roll-up of the separated boundary layer at Lastly, Fig. 8 shows lifting of the negative
the cusp (Fig. 7). vorticity layer from the slat surface as the slat
cusp is approached. This lifting is due to more
Once released at the cusp, typically a pronounced boundary layer separation at an
negative sign vortex is paired up with an opposite earlier stage, formation of large negative sign
sign vortex while moving upward towards the
vortices, and subsequent generation of secondary
reattachment point. Although not depicted in Fig. and tertiary vortices.
7, occasionally, a significant incursion into the
freestream by the discrete vortices takes place. Analysis of the simulation for the 4-degree
However, careful frame-by-frame scrutiny of the angle-of-attack produced results that are
unsteady cove flow failed to produce any somewhat similar to the 6-degree case described
evidence of vortices colliding with the main above. Although a few differences exist, no
element leading edge. Another distinguishing dramatic changes in the cove flow dynamics
feature of Fig. 7 is the pumping of negative sign (such as those observed between 8-degree and 6-
American Institute OfAeronautics And Astronautics
degreceasesw)erefoundO. veraltl,helarge-scale
generation mechanisms and not on producing a
structureisn thecoveflowareslightlymore
perfect prediction technique.
energetaicndthemovemeonftthevorticemsore
chaoticcomparetodthe6-degrereesults.
A careful examination of Fig. 9 reveals several
similarities between the experimental acoustic
4.2
Acoustic Analysis and computational acoustic spectrums. For
instance, both suggest a noise minimum in the
Unsteady flow data on a surface enclosing vicinity of 20,000 Hz. In the low- to mid-
the wing elements and the slat-cove region are frequency range, both computed spectra exhibit
used as input to the code described by Lockard 13 some of the proper measured trends, but the
for the solution of the Ffowcs Williams and partially laminar spectrum is on average 5 to 6
Hawkings 14 equation to calculate the noise dB higher in amplitude. In particular, a noise
radiated below the airfoil. Full discussions of the
peak in the 3,000 to 3,500 Hz range is apparent.
acoustic procedure and the orientation of the data Because of the 2D nature of the simulations, the
surfaces are provided by KSB. For each angle of predictions should be significantly higher in
attack, in excess of 35,000 time steps are amplitude than the experimental results. Hence,
simulated of which (after discarding the initial even though the turbulent simulation results
transient part) the last 32,768 time steps are kept appear to be in reasonable agreement with the
for processing purposes. Except for the 8-degree experiment, they are really too low. We are
case, each data record is sub-divided into four working to quantify the influence of the spanwise
equal segments of 8,192 time steps. Each correlation on the results.
segment is run through the acoustic solver, and
the four outputs averaged before converting to A Comparison between the computed and
1/12th octave bands. Unfortunately, for the 8- measured acoustic spectra for the 6-degree case is
degree angle, the first two segments of the input plotted in Fig. 10. At this angle of attack, the
data record were corrupted during processing. peak in the computed spectrum is shifted to lower
Therefore, only the last two segments of 8,192 frequencies and occurs between 1,500Hz and
could be used. As a result, the averaged data are 2,500Hz. The decay with frequency and the
constructed using the remaining two available frequency of the noise minimum are very similar
segments. between the computation and the experiment.
The three computed spectra are displayed in Fig.
Figure 9 shows the comparison between 11. In accordance with the measured trend (Fig.
the computed and measured noise spectra for the 2), decreasing the angle-of-attack from 8 degrees
8-degree angle of attack case. Also shown in the to 6 degrees produces higher acoustic amplitude.
figure is our previously obtained spectrum from In lowering the angle further, no substantial
the fully turbulent simulation. The prominent change in the acoustic amplitude is observed.
tonal feature centered near 38,000Hz (due to This is not totally unexpected. Recall that the
vortex shedding at the slat trailing edge) was analysis of the simulated flows showed only
previously the subject of extensive studies. 6,8 minor effects in the time evolution of the cove
Accordingly, we limit our discussion to flow when angle of attack was lowered from 6
frequencies below 20,000Hz. A cautionary note degrees to 4 degrees.
is warranted in the interpretation of Figures 9-11.
5. Conclusions
As pointed out earlier, the URANS simulations
are conducted in a 2D fashion. To compute the
An aeroacoustic analysis of a generic, three-
farfield noise, a perfect spanwise correlation in
element high-lift configuration was undertaken.
the nearfield unsteady signal is assumed. No
As a continuation of our past effort, the present
doubt, in an actual experiment, three-dimensional
research was focused on the time-evolution and
effects are present. Under such conditions, the
dynamics of the slat's cove flow and the
spanwise correlation is less than perfect.
corresponding separated free shear layer. New
Therefore, a 2D acoustic computation potentially
two-dimensional, unsteady Reynolds Averaged
overestimates the noise significantly. Our main
Navier-Stokes simulations were performed in an
emphasis has been on identifying the noise
American Institute Of Aeronautics And Astronautics
attemptto remedythe shortcomingosf our
Prediction," AIAA Paper 2001-2158, May
previoussimulationasn,dalsotounderstatnhde
2001.
effectsof angle-of-attacvkariationon the
radiatedslatnoiseT.ocircumventhteexcessive 5 Davy, R. and Remy, H., "Airframe Noise
dissipativeeffectsoftheturbulencmeodeilnour Characteristics on a 1/11 Scale Airbus
pastinvestigationa, simplezonalapproach Model," AIAA Paper 98-2335, 1998.
(wherebytheturbulenceproductiontermis
6. Singer, B. A., Lockard, and D. P., Brentner,
turnedoff)wasadvocateadndimplementeTdh.e
K. S., "Computational Aeroacoustic Analysis
removaolfthisexcesdsampinagllowetdheshear
of Slat Trailing-Edge Flow," AIAA J. Vol.
layertoamplifylarge-scasletructureasc,hievae
38, No. 9, September, pp. 1558-1564, 2000.
non-lineasraturatiosntatea, ndpromotevortex
mergingIt. alsoallowedthecovetodeveloap 7. Khorrami, M. R., Singer, B. A., and
richerandmorecomplerxecirculatinflgowfield Berkman, M. E., "Time-Accurate
andto establisha feedbackloopwithinthe Simulations and Acoustic Analysis of Slat
recirculatioznone. Free-Shear Layer," AIAA Paper 2001-2155,
May 2001.
Simulationfosrangleosfattacokf8,6,and
8. Khorrami, M. R., Berkman, M. E., and
4degreewsereconducteCda.refualnalysoisfthe
Choudhari, M., "Unsteady Flow
simulateddatabasreevealesdignificancthanges
Computations of a Slat with a Blunt Trailing
intheprominenfetatureosfthecoveandgap
Edge," AIAA J., Vol. 38, No. 11, November,
flowfieldswithreductioninstheangleofattack,
pp. 2050-2058, 2000.
corroboratinpgastPIVmeasuremeUntssi.ngthe
unsteadRyANSdataasinput,solutionosfthe 9. Takeda, K., Zhang, X., and Nelson, P.A.,
FfowcsWilliamsandHawkingesquatiownere "Unsteady Aerodynamics and Aeroacoustics
calculatedto yield the farfieldacoustics. of a High-Lift Device Configuration," AIAA
Comparisonsbetweenthe computedand Paper 2002-0570, January 2002.
measureadcoustiscpectrashowedmuchbetter
10. Paschal, K., Jenkins, L., and Yao, C.,
agreemenfotr the amplitudeandfrequency "Unsteady Slat Wake Characteristics of a 2-
contentthanprevioucsalculationsO.veraltl,he
D High-Lift Configuration," AIAA Paper
computeadcoustiscpectrlaeavelittledoubtthat
2000-0139, January 2000.
theslatfreeshealrayeristheprominenntoise
producinmgechanisinmthelowerfrequencies. 11. Takeda, K. Ashcroft, G.B, and Zhang, X.,
"Unsteady Aerodynamics of Slat Cove Flow
References in a High-Lift Device Configuration," AIAA
,,
1. Meadows, K. R., Brooks, T. F., Humphreys, Paper 2001-0706, January 2001.
W. M., Hunter, W. H., and Gerhold, C. H.,
12. Menter, F., "Improved Two-Equation k-t_
"Aeroacoustic Measurements of a Wing-Flap
Turbulence Models for Aerodynamic Flows,"
Configuration," AIAA Paper 97-1595, 1997. NASA TM 103975, 1992.
. Hayes, J. A., Home, W. C., Soderman, P. T., 13. Lockard, D. P., "An Efficient, Two-
and Bent, P. H., "Airframe Noise
Dimensional Implementation of the Ffowcs
Characteristics of a 4.7% Scale DC-10
Williams and Hawkings Equation,"
Model," AIAA Paper 97-1594, 1997.
Journal of Sound and Vibration, Vol 229,
3. Dobrzynski, W., Nagakura, K., Gehlhar, B., No. 4., pp. 897-911, 2000.
and Buschbaum, A., "Airframe Noise Studies
14. Ffowcs Williams, J. E. and Hawkings, D. L.,
on Wings with Deployed High-Lift Devices,"
"Sound Generation by Turbulence and
AIAA Paper 98-2337, 1998.
Surfaces in Arbitrary Motion," Philosophical
4. Dobrzynski, W. and Pott-Pllenske, M., "Slat Transactions of the Royal Society of London
Noise Source Studies for Farfield Noise A, Vol 342, pp. 264-321, 1969.
American Institute Of Aeronautics And Astronautics
_Ostowed(= Ccruise)
_s__ "_-_ f
j-of
Fig. 1Cross-sectional view of three-element EET high-lift system. Model stowed chord is 0.55mm.
100 -
_
6 deg
_ 9 (leg
_
90- 15 (leg
_80 -
0
_.,__._ __ /'_\1
_70 -
m
"0
..f
o-60
50
' 'l I1ll0_ I I I , , ,_L1I0I , , , , , iii10IS
Frequency, Hz
Fig. 2 Measured acoustic spectra for slat in 1/12'h-octave bands. Test Parameters are: slat deflection angle of
30°, flap deflection angle of 30°, Mach number of 0.2.
Fig. 3 Simulated instantaneous spanwise vorticity field for fully turbulent 3% forced case.
American Institute Of Aeronautics And Astronautics