NASA Technical Reports Server (NTRS) 20040110387: A Study of the Effects of Blade Shape on Rotor Noise

15 Pages·2.4 MB·English

by  NASA Technical Reports Server (NTRS)

A STUDY OF THE EFFECTS OF BLADE SHAPE ON ROTOR NOISE
Henry E. Jones
U.S. Army Joint Research Program Office/AFDD
Casey L. Burley
NASA Langley Research Center
Abstract c rotor aerodynamic reference blade chord
A new rotor noise prediction system called the Til- c rotor blade chord at tip
tip
trotor Aeroacoustic Code (TRAC) has been developed
under the Short Haul (Civil Tiltrotor) program
T
between NASA, the Army, and the U.S. helicopter C rotor thrust coefficient, ------------------------------
industry. This system couples the comprehensive T rp W ( R)2R2
rotorcraft code CAMRAD.Mod1 with either the high
resolution sectional loads code HIRES or the full r spanwise coordinate along blade
potential CFD code FPRBVI to predict unsteady blade
DNW Duits-Nederslandse Windtunnel
loads, which are then input to the noise prediction
program WOPWOP. In this paper, HIRES will be used FW-H Ffowcs Williams-Hawkings equation
to predict the blade-vortex interaction (BVI) noise
Y spanwise location of vorticity, measured
trends associated with blade shape. The baseline BAR
from tip
shape selected was a 17% scale model of a contempo-
rary design 4 bladed rotor. Measurements for this BPF rotor blade passage frequency; 96.17 Hz
rotor were acquired in the Duits-Nederslandse Wind-
BVISPL sound pressure level obtained by
tunnel (DNW). The code is used to predict noise for
integrating from the 7th through 30th
the base configuration and the results are compared to
harmonics of the BPF (nominally 673
the measured data. This provides a firm foundation
Hz to 2885 Hz), dB
for investigating the BVI noise trends associated with
blade shape. The shapes selected for study are based l blade sweep angle, deg
on variation of sweep and taper which reflect plausi-
b ’ observer elevation angle, deg
ble “passive” design concepts. Comparisons of power
required, integrated noise, and aerodynamics are Y ’ observer azimuthal angle, deg
made and important trends are noted.
p pressure variable in WOPWOP model
max
Nomenclature N force, newtons
a freestream sound speed, m/sec m meters
0
V freestream velocity, m/sec
Background
b number of blades
Impulsive rotor noise is caused by rapid
c blade mean chord, m
changes in the local aerodynamic environment of the
R rotor radius, m/sec blade. Until recently, computing these effects has been
beyond the state of the art, and hence experimental
r air density, kg/m3
studies were used to investigate the relevant phenom-
W rotational speed, rpm ena. Brooks1 provides an excellent review of much of
this work. Recently, a new rotor noise prediction sys-
s blade solidity ratio, bc/p R
tem called the Tiltrotor Aeroacoustic Code (TRAC)
m rotor advance ratio, V/W R has been developed under the Short Haul (Civil Tiltro-
tor) program between NASA, the Army, and the U.S.
helicopter industry. The code system has, to date,
been successfully demonstrated on a model tiltrotor
configuration by Burley2; furthermore, subsets of the
Presented at the AHS Technical Specialists’ Meeting for system have also been presented by Brooks3,
Rotorcraft Acoustics and Aerodynamics; Williamsburg, Visintainer4, and Brentner5. The current effort contin-
VA, October 28-30 1997
ues the exploration of the application of this new sys-
tem. Here, for the first time, the system as a whole is
1
used to exploreconventional rotors within the context essential methodology of the original model is
of a geometric parameter sweep. The goal is to test the retained; however, a number of key enhancements
ability of TRAC, with all of its modeling assumptions, and modifications have been made, including
to predict noise trends for a family of rotors. A com- increased azimuthal resolution, swept planform
plete study of all the issues associated with the solu- effects, indicial aerodynamic blade response model-
tion of this problem would be beyond the scope of this ing, and vortex roll-up modeling. The resulting
paper, hence we will restrict the discussion to wake method is much more suited to the prediction of noise.
modeling, rotor loads, and blade-vortex interaction
(BVI) noise. The emphasis will be on the use of TRAC Computational results from CAMRAD.Mod1 pro-
and the predicted results and not on the relative mer- vide the spanwise blade section loading, performance,
its of a particular design. and trim solution for the rotor. Additional processing
is required to produce the high resolution solution of
The paper is divided into three sections: an intro- the rotor wake, blade motion, and sectional loading
duction to the system of codes, application of the solution which is necessary for noise prediction.
method to a single configuration and a brief correla-
tion with experimental data, and the blade geometry Compact Section Loading
study.
HIRES is an extension to the CAMRAD.Mod1
The Tiltrotor Acoustics Code System: TRAC which is executed following the trim calculation.
HIRES is typically used to compute blade loads at .5
TRAC consists of a set of computational tools
degrees azimuthal step and at 100 radial stations.
which, taken together, provide a prediction method
These high temporal and spanwise resolution loads
for the complex multi-disciplinary problem of rotor
are obtained by recomputing the wake velocity influ-
acoustics. Figure 1 depicts the flow of data between
ence coefficients for recalculated blade motion and
the various codes within the system. For the current
interpolated wake geometries at each azimuth step.
effort, the CAMRAD.Mod1-HIRES-WOPWOP
The aerodynamic calculations in HIRES are accom-
method is employed.
plished in a post-processing program which uses the
high-resolution definition of the induced velocity over
High Resolution Acoustics the rotor disc. The resulting high resolution lift and
Low Resolution Blade Loads drag are then available for use with the acoustic pre-
Harmonic, Impulsive, diction code.
Broadband Noise and
FPRBVI
propagation Acoustics
CFD
Full Surface
After the TRAC modules have produced the high
Loading
Rotorcraft CAMRAD.Mod1 WOPWOP resolution airloads, the acoustics propagation model
Operating Performance WOPWOP+ WOPWOP7,8 is employed to predict the noise level at
Condition Trim/Wake Code specified microphone locations. WOPWOP is the cod-
KIRCHHOFF
ing of acoustic formulation 1A of Farassat7, which is a
HIRES BARC
time domain solution to the FW-H equation, exclud-
Non−CFD RNM
ing the volume source or “quadrupole” term. It was
Compact
ROTONET
Section developed as a method to predict the discrete fre-
Loading quency noise of helicopter rotors, and is valid for arbi-
trary blade motion, geometry, and observer location.
FIGURE 1. TRAC flowchart. The HIRES option of TRAC provides high-resolution
section blade loads which are converted into equiva-
Rotorcraft Trim lent “compact” pressures acting at a specified chord
location (usually the local quarter chord).
The CAMRAD.Mod1 computer code is capable of
predicting the dynamics, aerodynamics, loads, and
Test Data
performance of a trimmed rotorcraft system in a wind
tunnel or in flight. The basic method makes use of A limited correlation with test data will be
lifting line theory coupled with airfoil section force conducted in order to provide a firm foundation for
and moment data tables to compute the load on the the planform study. The data will be compared to
blade. The method is described in some detail by acoustic data from the 1989 aeroacoustic test con-
Brooks.3 Basically, CAMRAD.Mod1 is a highly modi- ducted in the DNW.9 The test was part of the U.S.
fied version of the original CAMRAD6 model. The Army Aerodynamic and Acoustic Testing of Model
2
Rotors (AATMR) Program and involved U.S. Govern- and 1500, in contrast to the smoother changes which
ment agencies and United Technologies Corporation occur on the retreating side of the disc. The secondary
(UTC). vortex features much more rapid variations in Y
bar
and is not present on the retreating side. It will later be
Application of the Method shown that the primary BVI for this configuration
occurs near the azimuth angle of about 750, just before
In this section, a set of results for a base configura-
the sharp change in Y for both vortices. The source
tion will be used to establish formats, to introduce the bar
of the vortex is at aboutY = 1200. Another interesting
use of the system, and for a limited correlation. Predic-
issue related to the rotor wake is the loading distribu-
tions will fall into three categories: (1) wake geometry;
tion during the actual BVI which is presented in figure
(2) rotor airloads, and (3) rotor acoustics.
3. Note especially, the multiple loading peaks for this
The CAMRAD.Mod1 model provides the user condition, most probably a result of predicted multi-
with a rich choice of options for use in predicting rotor ple interactions. It should be noted that the actual BVI
performance, loads, and acoustics. These models was not measured or observed during the test but
include extensive treatments of the rotor structural measured blade loads indicate similar trends.
and aerodynamic environments. Before proceeding, it
is, therefore, important to establish the modeling
options to be used. This will provide a frame of refer- TipVortexShedding&Rollup
ence within which comparisons can be made. The fol-
1
lowing assumptions are employed for the study:
1. an isolated main rotor in a wind tunnel
0.9
a. R = 1.428 m
)R TipVortex
b. a0 = 338.19 m/sec -yBA SecondVortex
1
c. r = 1.226 kg/m3 n,( 0.8
o
d. W = 1442.5rev/min ati
c
o
e. V = 215.7 m/sec L 0.7
TIP al
f. l = 200 (at r=.94 for base adi
R
configuration)
0.6
2. shaft axes at 5.150 (modeling descent)
3. C /s = 0.0714
T 0.5
0 60 120 180 240 300 360
4. m =.151
AzimuthalLocationofShedding
5. constant blade area for each configuration
FIGURE 2. Predicted spanwise emission locations
6. linear blade twist of -120 (except for baseline) for both the tip and secondary vortices
8. The full free-wake model (3 iterations)
Rotor Airloads
with multiple vortex trailers and a
multi-core (9) vortex model Rotor airloads will be presented as contours of
loading (in N/m) over the rotor disc of the blade as in
9. The blade structural response will be modeled
figure 4 below. Of particular interest in this figure are
with six bending modes and 1 torsion mode.
the regions of dense “bands” which indicate high tem-
poral gradients. When these bands are near parallel to
Rotor Wake
the rotating blade, BVI impulsive noise results. It is
Performance and wake data are presented to sup- seen that this occurs in the first quadrant between azi-
port the acoustic results. Of particular interest here is muth angles of 600 and 900 and on the retreating side
the vortex roll-up behavior. Figure 2 is a plot of the between about 2900 and 3100.
predicted tip and secondary vortex spanwise emission
locations as a function of azimuth. Note the presence
of the secondary vortex which is deposited into the
wake beginning at 300 azimuth and existing until
about 1900. Also note the sharp decrease in Y at800
bar
3
loads, the input WOPWOP subroutines were adapted
to utilize such a loading description.
SpanwiseLoading
CAMRAD-Mod In WOPWOP the pressure on the lower surface is
800
Y =700 assumed to be zero and the upper surface pressure is
given by:
0 (x/c) <.2
600
p = 20((x/c) -.2)p .2< (x/c) <.25
max
20((x/c) -.3)p (x/c) >.3
max
N
d, 400
a where p is set such that the total section lift is the
o max
L same as that produced by the distributed load. The
WOPWOP blade motion is specified using the pre-
dicted results from CAMRAD.MOD1. The first and
200
second harmonics of rigid blade flap, lead-lag and 1st
torsion are considered.
VortexHits
Figure 5 compares the measured and predicted
0 noise time history at a microphone location 250 below
0 0.2 0.4 0.6 0.8 1
RadialStation the advancing side. The figure shows that the TRAC
system is capable of computing the BVI noise ampli-
FIGURE 3. Spanwise blade load during vortex tude including important blade passage features and
encounter, Baseline configuration (HIRES) BVI events, but that many of the details, such as blade
to blade variations of the waveform are still beyond
the ability of the system to capture.
BaselineBlade
ShaftAngle=5.15
CT/S=.0714
400
600 AcousticWaveform
600
600 600 640000 25 PREDICTED
20 MEASURED
800 600 400 200
400 15
600
600800400600 400 204006000 800800600 600600 240006000 600400400 tot-pa 1050
400
800
600 600400
-5
-10
FIGURE 4. Contour of blade loading for base
configuration. (HIRES) -15
0.25 0.5 0.75 1
TIME
Rotor Acoustics
FIGURE 5. Sound pressure level for base
Acoustic Signal configuration at a specified microphone location
compared with test data. (HIRES)
As stated above, the WOPWOP code predicts the
noise time histories and sound pressure level (SPL)
Noise Directivity
generated by the blade at a user specified microphone
location. In the WOPWOP code, the loading must be It has long been recognized that the impulsive
given as a force distribution over the blade surface. loading of the BVI generates a highly directional
Since CAMRAD.Mod1/HIRES computes sectional acoustic field. It follows that a large number of
4
observer locations must be computed to capture this
effect. To study this characteristic, Brentner5 proposed HIRES MODEL
BaseBlade
that the sound be predicted on a sphere of 1.5 rotor NoTorsion-Rollup ShaftAngle=5.15 7th-30thHarmonics
dltYohi’cae aa mrtneiedogtneibosr’n sla i -erc6e ea0 ntt0h t 1e<e5r e0aY dzini ’m co<rnu e6 tm0ht0h e aeann ntdsrdo o et3nol0e r0tv h a<hetu i sobbpn’. h <aTen-rh2ege1 l s0euo0s,br sfwsahecohrevwe irenner 300 83 897185999133 8995 95 97 919389 111000975
99 103
in figure 6. -30 101
99
ObserverLocationson 101 99 95 97
Y / -60 95
ComputatinalSphere 95 93
hi-90 97 97 9819
p 93 95 87
85
Z -120 101 99 887319
X Y -150110031 105 9973 95 7775
-180 95101 9997 9591 89 89
b / -210 103
-60 -30 0 30 60
psip
FIGURE 7. Directivity pattern for predicted mid-
range noise of baseline blade. (HIRES)
Blade Planform Study
The remainder of the paper will deal with the
blade planform study. The shapes selected for study
are based on variation of sweep and taper which
FIGURE 6. Geometric distribution of observers in a reflect plausible “passive” designs of a family of con-
“bowl” around the rotor.
ventional single main rotor configurations. Figure 8
The noise predictions are presented as contours of provides a tabular and visual summary of the configu-
the midrange noise metric which is the sum of the 7th rations. Each configuration is assumed to have the
through 30th harmonic of the blade passage frequency same mass and stiffness distributions. Furthermore,
of the overall sound pressure level (OASPL) on theY- airfoil section shape and distribution are held constant
b plane. Figure 7 gives this result for the baseline as is the sweep angle of 200. Chord and sweep location
rotor. are the only variables. Chord is changed via a tapering
of the blade and by moving the location of the taper
The figure shows two regions of focused high
initiation point. The sweep initiation point is also var-
noise: a large region centered around (Y,b) = 200,-600
ied. For each of the configurations, the root chord has
at about 101 dB.; and a smaller but more intense
been modified in order to maintain a constant blade
region centered at (Y,b) = (-450,-1500) at about 105 dB. area. Twist is held constant at -120, except for the base-
Both these regions can be related to the high loading
line which has a highly non-linear twist distribution.
gradients seen in figure 4.
Finally, a five bladed version of the last swept-tapered
configuration is examined. The configurations can be
viewed as being “grouped” with the LinTwt, Lamda1,
and Lamda2 group representing taper changes, and
LamdaSwp1, LamdaSwp2, and LamdaSwp3 repre-
senting sweep changes. The Base and LinTwt configu-
rations show the effect of twist and the LamdaSwp3
and Blds configurations show the effect of number of
blades.
5
etry (miss distance), and vortex strength during the
interaction. The issue here is how the blade loading
BaseBlade LamdaSwp1Blade BladePlanformVariations
changes with shape, and how this affects the vortex
Name Taper Sweep Taper rollup. Table 1 gives a summary of the key (closest)
Starts Starts
BVI events for each blade. The major interaction for
Base1 1. .94 .94
most of the configurations is in the first quadrant
Lintwt2 1. .94 .94 between 65 and 75 degrees azimuth; the first swept-
LinTwtBlade LamdaSwp2Blade
Lamda1 .75 .94 .94 tapered blade is an unexplained exception. Another
feature of these interactions is that it would seem that
Lamda2 .50 .94 .94
each blade has at least several encounters at the same
LamdaSwp1.50 .94 .94 location. The source of these vortices all seem to be
located in the second quadrant at about 1200azimuth.
Lamda1Blade LamdaSwp3Blade LamdaSwp2.50 .87 .87
LamdaSwp3.50 .87 .75
SpanwiseLoading
Blds3 .50 .87 .75
CAMRAD-Mod
2000
1Non-LinearTwist
Lamda2Blade BldsBlade Base
2LinearTwist 1800 LinTwt
Lamda1
35Blades Lamda2
1600 LamdaSwp1
LamdaSwp2
AMSiearcfsotsiiol&SneSCct.tifiGofnn.e&DssAisP.trCriob.puAetidrotjnuies&steCSdohfnaosprteSanwCteoenpstant 1400 LBaldmsdaSwp3
FIGURE 8. Planform views of blade shapes. 1200
N
d,1000
Rotor Performance and Wake a
o
L
800
600
15
TotalRotorPowerRequired
400
200 VortexHits
0
0 0.2 0.4 0.6 0.8 1
p 10 RadialStation
h
r,
e
RotorPow 5 Baseline LinTwt Lamda1 Lamda2 LamdaSwp1 LamdaSwp2 LamdaSwp3 5Blds FFIilgGoucUraReti Eo10 n1 0pfo.r rBe selaeadnctehs r ctaohdnei fiaslgp luaornaawdtiiionsneg ; baHlta ItdRheEe Slmo maadjooidnr egBl V.foIr
each configuration at the location of the principal BVI.
Please note that the “zero” for each configuration is
shifted by 200 N in order to separate the curves for
easy viewing. With the exception of the baseline rotor,
which has exceptionally large variations (or response),
0
0 1 2 3 4 5 6 7 8 9 the remaining blades show a similar loading curve,
Configuration
with the geometric changes appearing as variations in
the location of the peaks. Note that a general trend of
FIGURE 9. Rotor power required at test point.
shifting load inboard follows the location of the taper/
Figure 9 is a summary chart of the predicted rotor sweep initiation point. Also note the small variation at
power required for the test condition. The figure the location of sweep initiation.
shows a clear trend in power with shape/geometry
Figure 11 gives the vortex centroid locations, that
change. HIRES is executed following the trim compu-
is the final rolled-up position, for the various configu-
tation, and it is from its results that the details neces-
rations. Note that except for the baseline configura-
sary to predict the rotor BVI phenomena are obtained.
tion, the vortex rollup follows a smooth pattern. The
Rotor BVI effects are influenced by four wake roll-up behavior seems to “group” with geometry,
related factors: blade loading at the time the vortex is that is, LinTwt-Lamda1-Lamda2 all have similar
generated, blade loading during the BVI, wake geom-
6
behavior as does the LamdaSwp1-LamdaSwp2-
LamdaSwp3 group. BaselineBlade
ShaftAngle=5.15
CT/S=.0714
TipVortexShedding&Rollup
400
1 600
600
0.975 600 600 640000
0.95 800 600 400 200
n 400
ortexLocatio0.902.95 bLLaiansmTedwtiapt1loc 600800400600 400 204000 600 606000600 400600 600400400
TipV0.875 LLLaaammmdddaaa2SSwwpp13 600 800800 200
Blds 400
0.85 800
600 600400
0.825
0.8
0 30 60 90 120 150 180 210 240 270 300 330 360 FIGURE 12. Contour of blade loading for base
configuration. (HIRES)
AzimuthalLocationofShedding
FIGURE 11. Tip vortex roll-up characteristics.
LinTwtBlade
ShaftAngle=5.15
CT/S=.0714
Airloads
ratioAnisr laoraed p rceosnetnotuerd pinlo ftisg uforre sr e1p2-r1e7s.entative configu- 600800 460000 600 600420600000
Figures 12, and 13 present the effect of twist
change on the loading: note here, the increase in load- 800 400 200 600
iecnoxgnte fniggtr uaordfai teain oBtn V.i nIA intsh iemn cosrereecao sanepd pi naq rutehanedt rnfaoonrit s.te h Teah sle ia n serpaearsn utwwlt iisoseft 600 600600 400606000 600600 400
the increased lateral coherence is expected and shown 400 200
200
ionf ttahpee nr:e txhte s cehctainogne. sF higeurere asr e1 3m aonsdtl y1 4in s dheotwai ltsh we iethff encot 400 400 600 640080000 600 400
major impact. Figures 14 and 15 show the effect of 200
800
sweep change: here there is a noticeable reduction in
loading gradient especially in the first quadrant. Fig- 600 806000400 200
ures 16 and 17 show the change with blade number.
There is a definite reduction in the intensity of the gra-
dients on the disc. FIGURE 13. Contour of blade loading for linear
twist configuration. (HIRES)
7
Lamda2Blade LamdaSwp3Blade
ShaftAngle=5.15 ShaftAngle=5.15
CT/S=.0714 CT/S=.0714
600 400
400 400 600 200
600 600 600 200600 400 600 800 400 600 600 400
400600800804000400 606000440000 200200 800 240408000060000 600600606000400602000400 660004000600 400 400400 200 600800 600200400 600600 600200600400 200
200 600 400
600 804000 800400 400 880000 600
FIGURE 14. Contour of blade loading for second FIGURE 16. Contour of blade loading for third
tapered configuration. (HIRES) swept-tapered configuration. (HIRES)
LamdaSwp1Blade BldsBlade
ShaftAngle=5.15 ShaftAngle=5.15
CT/S=.0714 CT/S=.0714
600 804000 200 300 604050000 200
604000600402006800000 400400460000 260000 200 400 200 604600060000606000660000400460000600 400200 400350300000500200434000500500000 404030400000500550000602100003000302005000 704000502060000300545000000 400505000345005000000300400320000200
400600800 800600 2000 300 600 650300000 200
FIGURE 15. Contour of blade loading for first FIGURE 17. Contour of blade loading for five
swept-tapered configuration. (HIRES) bladed configuration. (HIRES)
8
Acoustic Predictions Figure 31 gives the change in noise with shaft
angle. Here, the “noise slope” for each configuration is
Signal Characteristics seen to “group” with geometry - the LinTwt-Lamda1-
Lamda2 all have slopes of about 3dB/deg while the
The loading predictions are utilized to predict
LamdaSwp1-LamdaSwp2-LamdaSwp3 are at about
the noise for the various configurations. Figures 18-21
2.8~2.9dB/deg. The Base slope is ~2dB/deg while the
give the comparison of the noise signal at a single
five-bladed configuration is ~2.2dB/deg. These differ-
microphone location (corresponding to the “hot” spot
ences may be due to the strength of the vortex encoun-
noted in figure 10) for the key geometric changes.
tered. Note from table 1 that LinTwt-Lamda1-Lamda2
Figure 18 shows the effect of changing twist from all have “hits” from the tip (i.e. stronger) vortex,
linear to non-linear; note here the shift in phase and whereas all others have at least on hit from a second-
the reduction in some of the peak pressures. Figure 19 ary vortex.
gives the effect of taper, which appears to have little or
Figure 32 is taken from reference 1 and represents
no effect. Figure 20 gives the effect of sweep location,
the only source of data which can be used to assess the
which appears to be mostly a phase change. Finally,
validity of figure 31. These data were taken for a vari-
figure 21 gives the effect of number of blades; shown
ety of blades and are operating at different conditions.
here is the obvious change in peak location due to the
Also the data-base was sparse. In spite of this, the gen-
shift in blade-passage frequency, but also a general
eral comparison between these figures, qualitatively
reduction in the amplitude.
shows similar trends.
Noise Directivity
Comments and Observations
Noise directivity contour results are shown in fig-
Rotor noise prediction is a complex problem. In
ures 22 to 29. The acoustic results using the HIRES
order to meet this challenge, the TRAC system has
blade loading show a noise ‘hot’ spot on both the
evolved from a number of different disciplines
advancing and retreating sides, with the most intense
(dynamics, aerodynamics, and acoustics) into an inte-
region on the retreating side. The blade loading con-
grated tool.
tours (figures 12-17) for these conditions, indicate high
loading levels and strong gradients on the retreating The following observations can be made:
side. The wake geometry results predict a single
1. TRAC is capable of predicting the basic
strong tip vortex, with a small miss distance for this
noise level of a conventional rotor even with geometri-
region, whereas on the advancing side there is typi-
cally complex features such as non-linear twist and
cally two vortices of lesser intensity. In general the
sweep.
advancing side ‘hot’ spot consistently was at a level of
101 dB for all the blade configurations, except for the 2. Noise can be affected using “passive” blade
5-bladed rotor, where the level decreased significantly designs.
by 5 dB. This is expected since the load is distributed
3. Noise effects seem to “group” within Archi-
among 5 blades rather than 4, and the resulting tip
tectures (i.e. shape, twist, or number of blades).
vortices should be weaker.
4. Architectural changes, especially step
Several changes in the noise directivity and level
changes, have a greater effect on noise than do para-
as a function of twist and sweep are of note. The linear
metric changes (i.e. taper).
twist results compared to the baseline (and all others)
shows a much larger advancing side ‘hot’ spot. The 5. The trends predicted for overall noise as a
HIRES airloads produce a considerable reduction in function of shaft angle are similar to those obtained
the size of the advancing side ‘hot’ spot for a change in from experimental studies.
the sweep location.
Issues which remain for future study include:
Integrated Acoustics 1. Addition of a surface loading model to the
CAMRAD.Mod1 system.
Figure 30 gives the integrated noise result in bar
chart format. Note here especially the drop in noise 2. Improved wake modeling within CAM-
between the linear and non-linear configurations RAD.Mod1,
Base-LinTwt. Also note the sharp drop for the five-
3. Continued investigation using the FPRBVI
bladed configuration. The other configurations show a
CFD post processor.
steady drop in noise with taper and sweep location
which parallels the trend seen in power.
9
15 15
Base Lamda2
LinTwt LamdaSwp2
10 10
a a
p 5 p 5
e, e,
r r
u u
s s
s s
e e
r 0 r 0
P P
-5 -5
-10 -10
0 0.05 0.1 0.15 0.2 0.25 0 0.05 0.1 0.15 0.2 0.25
TIME TIME
FIGURE 18. Effects of non-linear twist on noise FIGURE 20. Effects of sweep location on noise
signal. signal.
15 15
LinTwt LamdaSwp3
Lamda2 Blds
10 10
a a
p 5 p 5
e, e,
r r
u u
s s
s s
e e
r 0 r 0
P P
-5 -5
-10 -10
0 0.05 0.1 0.15 0.2 0.25 0 0.05 0.1 0.15 0.2 0.25
TIME TIME
FIGURE 19. Effects of taper on noise signal. FIGURE 21. Effects of blade number on noise signal
10