The Impact of Structural Vibration on Flying Qualities of a Supersonic Transport

12 Pages·2001·1.1 MB·English

by  Raney, David L.

AIAA 2001-4006
The Impact of Structural Vibration
on Flying Qualities of a Supersonic
Transport
David L. Raney, E. Bruce Jackson,
Carey S. Buttrill, and William M. Adams
NASA Langley Research Center
Hampton, VA 23681-2199
AIAA Atmospheric Flight Mechanics
Conference
6-9 August 2001
Montreal, Canada
For permission to copy or to republish, contact the copyright owner named on the first page. For AIAA-held copyright,
write to AIAA Permissions Department, 1801 Alexander Bell Drive, Suite 500, Reston, VA, 20191-4344.
The Impact of Structural Vibration on Flying Qualities
of a Supersonic Transport
David L. Raney*, E. Bruce Jackson _,Carey S. Buttrill*, and William M. Adams*
NASA Langley Research Center
Hampton, VA 23681-2199
ABSTRACT in the design of the Boeing High Speed Civil
Transport (HSCT). Constraints imposed by flight
A piloted simulation experiment has been at supersonic speeds resulted in a very large but
conducted in the NASA Langley Visual/Motion relatively light and slender vehicle design that
Simulator facility to address the impact of exhibited unusually low-frequency aeroelastic
dynamic aeroelastic effects on flying qualities of modes. The resulting low frequency cockpit
a supersonic transport. The intent of this vibrations had significant potential to negatively
experiment was to determine the effectiveness of influence the pilot's ability to maneuver the
several measures that may be taken to reduce the aircraft, not only due to the degradation of ride
impact of aircraft flexibility on piloting tasks. quality but also due to adverse coupling between
Potential solutions that were examined included the human pilot's control dynamics and the
structural stiffening, active vibration configuration's aeroelastic dynamics.
suppression, and elimination of visual cues
A piloted simulation experiment was
associated with the elastic modes. A series of
therefore conducted in the Langley
parametric configurations was evaluated by six
Visual/Motion Simulator (VMS) facility to
test pilots for several types of maneuver tasks.
address the impact of dynamic aeroelastic effects
During the investigation, several incidents on flying qualities of the High-Speed Civil
were encountered in which cockpit vibrations Transport. The primary objective of this
due to elastic modes fed back into the control
investigation was to determine the effectiveness
stick through involuntary motions of the pilot's of measures that may be taken to reduce the
upper body and arm. The phenomenon, referred impact of aircraft flexibility on piloting tasks for
to as biodynamic coupling, is evidenced by a the HSCT. The secondary objective was to
resonant peak in the power spectrum of the establish preliminary guidelines for designing a
pilot's stick inputs at a structural mode structural mode control system for an HSCT
frequency. concept.
The results of the investigation indicate that An earlier motion-based simulation study
structural stiffening and compensation of the using a dynamic aeroelastic HSCT model in the
visual display were of little benefit in alleviating Langley VMS facility revealed a significant
the impact of elastic dynamics on the piloting reduction in the ease with which piloted
tasks, while increased damping and elimination approach and landing tasks were performed
of control-effector excitation of the lowest when dynamic elastic modes were included in
frequency modes offered great improvements the model. 1 An even earlier investigation by
when applied in sufficient degree. Schmidt and Waszak had also been conducted in
INTRODUCTION the Langley VMS, illustrating the potentially
detrimental effects of dynamic elasticity on the
As commercial transport aircraft designs flying qualities of a flexible B1 aircraft
become larger and more flexible, the impact of simulation in 1985) The potential for such
aeroelastic vibration on the vehicle's flight effects had been noted in a 1983 report by
dynamics, flight control, and flying qualities Ashkenas, Magdaleno and McRuer that
increases in prominence. The consideration of cautioned of implications of structural flexibility
such effects assumed unprecedented significance for the flying qualities of large aircraft. 3
*Research scientist, Dynamics and Control Branch, MS 406, Senior Member AIAA.
Research scientist, Dynamics and Control Branch, MS 149, Senior Member AIAA.
t Research scientist, Systems Development Branch, MS 125B, Senior Member AIAA.
$Distinguished Research Associate, Dynamics and Control Branch, MS 406, Senior Member AIAA.
Copyright © 2001 by the American Institute of Aeronautics, Inc. No copyright is asserted in the United States under
Title 17, U.S. Code. The Government has royalty free license to exercise all rights under the copyright claimed herein
for government puqooses. All other rights reserved bythe copyright owner.
American Institute of Aeronautics and Astronautics
Theapproacohfthepresenintvestigation
trailing-edge flaperons (four per wing), and
wastovarycertaipnarameteinrstheaeroelastic
leading-edge flaps.
modetloprovideasimplifierdepresentatoiofn
severaplotentimaleanosfalleviatintgheimpact The vehicle is designed to cma2¢ 300
ofdynamiaceroelasticoitnypilotingtasksT.he passengers a distance of approximately 5,000
potentiaslolutionthsatwereexaminecdonsisted nautical miles. The aircraft has an operating
of 1)increasintghefrequencoyftheelastic empty weight of 280,000 lb and a maximum
modes2,)increasintghedampinogfvarious takeoff gross weight of 650,000 lb. Final cruise
combinatioonfselastimc odes3,)eliminatioonf weight is expected to be 385,000 lb.
controleffectorexcitationof the lowest
The length of the HSCT configuration is
frequenceylasticmodesa,nd4)eliminatioonf approximately 310 ft, with a wingspan of 130 ft.
visuaclueasssociatweidththeelastimcodes. The HSCT's length and slenderness result in very
Approximarteepresentatiofonrseachof low-frequency structural modes, with the first
thesepotentiaslolutionwseregenerateindthe symmetric bending mode occurring at approxi-
simulatiomnodeblydirectlyparameterizthineg mately 1.25 Hz. As a point of reference, the
modalfrequencieasnddampinga,swellas length of the Concorde is approximately 204 ft,
controilnputandvisualoutpuet ffects.By the length of the Tu-144 is 215 ft, and the B1 is
explorinpgarametvriacriationinseachofthese approximately 135 ft. The HSCT planform is
factorsin,formatiownasgainedregardintghe shown in Figure 1, super-imposed with a
effectiveneosfesachapproacahn,dthedegreteo planform of the B1 configuration for scale
whichitmustbeexerciseindordetroachieve comparison.
thedesirefdlyingqualitieims provemeAntt.otal
of20parametrciconfiguratiownesreevaluated
B1RideControl
bysixtestpilotsrepresentinthgeFAA(1),
Va_e(RCV)
Boein(g1),NASA(2),andCalspa(2n).
HSCT SIMULATION MODEL
This experiment used a mathematical RkteContro_Vanes
simulation of the so-called "Cycle 3" version of (RCV)andChi_ Fin
(CF)addedat nose
the "Reference-H" supersonic transport design. 1
The model was published by Boeing
Figure 1. Comparison of HSCT and B-1
Commercial Aircraft Group in the summer of
planforms.
1996 as the fourth major release in a series of
increasingly detailed math models of the
The B 1uses an active vibration suppression
Reference-H configuration. The simulation
control system to improve ride quality at the
model was based upon a combination of wind
cockpit station. Small active control surfaces
tunnel and computational fluid dynamics studies called ride control vanes are located near the
of the Reference-H design, ranging from low
pilot station of the B1to damp vertical vibrations
subsonic to Mach 2.4 supersonic wind tunnel
at the cockpit. The Boeing HSCT design is
studies.
equipped with similar devices for active
Finite-element structural models were used suppression of vertical vibrations as noted in
to predict the effect of steady flight loads upon Figure 1. Additionally, a vertical "chin fin" is
aerodynamic stability derivatives, referred to as included in the HSCT design for active
quasi-static aeroelastic effects (QSAE). A key suppression of lateral cockpit vibrations.
feature of the math model was the inclusion of
dynamic aeroservoelastic effects (DASE), which Dynamic Aeroelastic Effects
required additional states to represent the flexible
The dynamic aeroelastic portion of the
modes of the aircraft structure.
model used in this simulation experiment
contained six flexible aircraft modes, 3
General Configuration Description
symmetric and 3 antisymmetric. The mode
The Reference-H vehicle design has a shapes and their associated in-vacuo frequencies
cranked-arrow planform with a conventional aft are shown in Figure 2. The model was generated
tail and four under-wing engines. The control for a flight condition of Mach 0.24 at aweight of
devices include an independently actuated 384,862 lb and a cg location at 53.2% of the
horizontal stabilizer and elevator, a three- mean aerodynamic chord, which constitutes the
segment rudder on a fixed vertical fin, eight landing condition.
2
American Institute of Aeronautics and Astronautics
Symmetric Modes (SideView) Antisymmetric Modes (TopView) variation probably extends beyond the conditions
Mode AN1: 1.39Hz that would be physically practical for this design
Mode SY 1:1.25 Hz
because of the weight penalties associated with
producing the stiffer structure. Aircraft weight
L Mode SY 2:2.01 Hz Mode AN 2:2.13 Hz was not increased in the simulation as the modal
frequencies were increased.
IL. Mode SY3:2.70 Hz Mode AN 3:2.82 Hz The "stifl" condition of 1.45 Hz for the first
symmetric mode was considered to be most
representative of the actual configuration, and so
was used as the modified baseline configuration
Figure 2. Aeroelastic mode shapes and in-vacuo for all other parametric variations. Acceleration
frequencies used in simulation. time histories from the motion-based simulation
were used to verify that this parameterization
Dynamic aeroelastic modes contained in the
method produced the desired effect.
model could be excited by turbulence and by
control effector movements. No inputs from Configuration Frequency Stiffness 1stSYMode
Ratio Increase Frequency
landing gear reaction forces or engine pylon baseline 1.00 -- 1.25Hz
reaction forces were included in the model. still 1,1_} -.35% 1,45H_t
stir2 t,30 ,-85% t,80Hz
Visual effects of the structural flexibility stir3 t,_0 ,-150% 2,00}-tz
were provided in the simulation. The out-the-
0.10 0.08 0.06 0.04 002 -- damping ratio
window scene that was presented on the cockpit
monitors moved in relation to the Head-Up
Display (HUD) to represent local perturbations
in pitch and yaw at the pilot station. The overall
effect was that the out-the window scene
appeared to bounce slightly both vertically and
laterally in response to elastic excitation. These Mode 4 (AN)
visual perturbations were typically on the order
of + 0.1 degrees during maneuvers performed
........ t
with dynamic aeroelastic effects.
Mode 2 (AN)
Parameterized Aeroelastic Characteristics
The baseline aeroelastic model described
above was parameterized to provide the ability to
systematically vary several characteristics of the
0 .......... -......... -......... •....... X:..X........... X.....
piloted simulation. These modifications allowed
::::::::_::::::: ............................... ::!i::i::::.........
the impact of structural stiffening, modal ...":::::'.'.'==
damping, modal cancellation and visual cues to -5 "'i ......... i......... i.... ..... i.... i" ........ i
0.5
-3 -z_ Real Axis -o_ o
be evaluated from a piloted control standpoint.
Variation of Structural Stiffness Figure 3. Migration of elastic mode poles with
The variation of structural stiffness was structural stiffness variation.
represented by multiplying the frequencies of all
Variation of Modal Damping
six elastic modes by a given frequency ratio. The
representation of structural stiffening by directly The portion of the test matrix that varied the
manipulating the model in this fashion is clearly modal damping level actually targeted several
approximate, but appeared sufficient to capture issues. The first issue was the effect of the
the basic effect. Frequency ratios of 1.0 amount of damping that was applied to the
(baseline), 1.16 ("stifl" configuration), 1.36 dynamic elastic modes. Damping ratios of 0.07,
("stif2") and 1.60 ("stif3") were chosen. This 0.15, and 0.30 were selected based on feedback
selection produced first symmetric bending obtained during an HSR Dynamic Aeroelastic
mode frequencies of 1.25 Hz, 1.45 Hz, 1.80 Hz Model Working Group meeting that was held in
and 2.0 Hz. Aeroelastic simulation models were August of 1997.
produced for each of these conditions. The second issue dealt with the frequency
Figure 3 illustrates the migration of the range of the modes to which damping was
dynamic elastic poles that occurred as the applied. In the first variation, damping was
stiffness level was varied. The total range of applied only to elastic modes with frequencies
3
American Institute of Aeronautics and Astronautics
lessthan2Hz. Thisrangeincludetdhefirst
Nz ps / Elevator (g/deg)
symmetrimcodeat 1.45Hz andthefirst
: : ': /
antisymmemtriocdeat1.62Hz.Inthesecond
variationd,ampinwgasapplietdoelastimcodes o_..................._ base!ine
withfrequencileesssthan3 Hz. Thisrange
/- oo7 : /
includetdhefirsttwosymmetraicndthefirsttwo .....,.............. ............
antisymmemtriocdeTs.heintenwt astoexamine
theeffectof dampinognlythefirstfuselage
bendinmg odesascomparetodthefirstand
seconfduselaghearmonischsowinnFigure2.
Figure4 illustratetshemigrationof the 0.US
dynameiclastipcolesthatoccurreadsincreased
dampinwgasappliedin thesetwofrequency 0
rangesT.hefrequencreysponspelotsshowinn 0 Frequency, Hz 4
Figure5illustratteheattenuatioonftheelastic
Ny ps / Rudder (g/deg)
responstoecontroilnputsthatresultfsromthe
increaseddampinlgevelsapplietdomodewsith
frequencileesssthan3Hz.Verticaalcceleration 0.Z5
atthepilotstatio(nNzps)inrespontsoeelevator
inputasndlateraalcceleratiaotnthepilotstation 02
(Nyps)inrespontsoeruddeinrputasreshown.
0.1
Anadditionaislsuethatwasaddressbeyd
thisportionofthetestmatrixwastherelative 0.
importancoef suppressinsygmmetrimc odes
versuasntisymmemtriocdesI.nbothfrequency 0.05
rangesa,dampinrgatioof0.3wasapplietdothe
symmetrmicodeaslone(configuratio"ndsamp4" 0 " " _ .... ;
and"damp9a")ndthentheantisymmemtriocdes 0 Frequency, Hz 4
alone(configuratio"dnasmp5a"nd"dam1p0").
Figure 5. Variation of frequency response to
Conflquration Dampinq Ratio Modes control inputs for various damping levels.
........s..t.ir!...........................n..o..m...i.n.a!..............._.............................
A total of 10 parametric conditions were
included in the damping portion of the test
matrix. Again, since the model was directly
manipulated to produce the desired damping
levels, the representation of an active mode
suppression system is approximate and lacks
nonlinearities and additional filter dynamics that
might be present in the actual system.
Z5
........ 0.i5 - 0.0'7 _-
...... 0.30 °Mode AN3 Impact of Modal Cancellation
zo ...,. .:_:"Y_"i..-ModeSY3 Another portion of the test matrix examined
...................."::=........ i......
-Mode AN2 the impact of modal cancellation. This term
" _ : Xio. refers to the elimination of the control effector
,CO 15 ...... __- ............_:-_.x =_-:.,"Mode SY2
excitation of a particular elastic mode or modes.
10 i .--ModeAN1 Such a modification allows the control system to
b .....ii...... _ _ _-.
"Mode SY1 pitch, roll, or yaw the aircraft without exciting
....... .;.i.i.... the specifically targeted modes. It is intended to
represent the effect that would be produced by
.............:..............................
using multiple control effectors in appropriate
,_ 0 proportion (canard and elevator for instance) to
'... . ...
pitch the vehicle without exciting the first
-5 •.., i i " i ...... "w. .....,.,.,........ii fuselage bending mode. In the lateral case, it
-I Real Axis -Z 0
represents the use of rudder and chin fin in
Figure 4. Migration of the dynamic elastic poles combination to avoid excitation of the first
with damping level variation. antisymmetric mode.
4
American Institute of Aeronautics and Astronautics
Configuration ModesCanceled ModesDamped require more control effectors. This representa-
stifl none none
tion is inherently approximate since it cannot
depict the effect of nonlinearities that would be
iiiiiiiiiiiiiii iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiSpr e!seiinitii i"ini!iiainiiiiaicitiuiaiiliiiiimiioidieiiiiiisiuipipiirie1ssiioin iiii iisiy site!m iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii
such as control saturation or rate limiting.
Nz ps / Elevator (g/deg) __ THE LANGLEY VISUAL/MOTION
SIMULATOR
o._ .................. _.stJf.!.. (ba.se!j.n.e
The Langley Visual Motion Simulator
cancl (VMS), shown in Figure 7, uses a synergistic
hexapod motion system. The motion platform
0.15 ............ i.i..... L.:__./\i[-canc2 ,i, ...........
provides up to +0.6 g acceleration cues vertically
o.1 ................ _ I_..... _../.._....O_i_!!C_'. ...... within a 5.75 foot travel envelope; lateral and
longitudinal acceleration limits are similar. 4The
o.o .........i..../........_.-./.i........ :i.:.c.a..n.e..4.......
angular limits of the platform are +30/-20 ° pitch,
i i/: ::
+32 ° yaw, and +22 ° roll. (Positive pitch is in the
nose up direction.)
0 Frequency, Hz 4
Linear Accelerations, g
Ny ps / Rudder (g/deg) Surge: + 0.6 Sway: + 0.6 Heave: + 0.8
Angular Accelerations, deg/s 2
0.25..............:.:.............:.:.............::..............
Roll: + 50 Pitch: + 50 Yaw: + 50
• I
0.1 ............. i........... i........ i..............
005 ..................
0 Frequency, Hz 4
Figure 6. Variation in frequency response to
control inputs for mode-canceled configurations.
The effect of such a mode-canceling control
system was approximated by eliminating the Figure 7. Langley Visual/Motion Simulator
elements of the B matrix in the dynamic (VMS) and its acceleration capabilities.
aeroelastic model that represent the control
effector excitation inputs to the first symmetric The cockpit configuration at the time of this
and first antisymmetric fuselage bending modes. experiment included a left seat pilot's station and
Mode-canceling configurations were generated a right seat observer's station. A four-lever
for each of the three damping levels so that the throttle quadrant was located between the two
test matrix would include direct comparisons of stations. A McFadden left-handed side stick
cancellation on and off for each of these controller was used for all maneuvers performed
damping conditions. during this experiment. The pilot station
Figure 6 illustrates the change in frequency included an armrest that was adjustable to
provide appropriate forearm support for the left
responses that resulted from this modification.
arm of the evaluation pilot.
The plots show that the first symmetric and first
antisymmetric modes can no longer be excited Refinements were made to the motion drive
by control inputs for the canceled configurations. algorithms to improve the suitability of the
The modal dynamics remain, however, and are simulator for representing the aeroelastic motion
subject to excitation by turbulence or by cues. The motion commands produced by the
coupling from other elastic modes. Cancellation dynamic elastic portion of the aircraft model
was applied only to the first symmetric and first generally bypassed the motion washout filters to
antisymmetric modes, since cancellation of avoid any attenuation or delay of the elastic
higher fuselage harmonics would probably vibration cues.
American Institute of Aeronautics and Astronautics
Figure8 wastakenfroma 1973report
completing a sufficient number of runs to rate a
whichdocumentethde frequencyresponse
particular configuration, the test pilot's
capabilitieosf the LangleyVMS Motion evaluations and comments were recorded. The
facility4.Theinput/outpaumt plitudreatiosfor
pilot's task performance, in terms of touchdown
vertica(lz)andlatera(ly)sinusoidinapl utosf1.8
parameters or flight-director tracking accuracy,
incheasreshowna,longwiththeresultinpghase
was provided on a cockpit display immediately
lag,forinpuftrequenciferosm0.1to12rad/sec.
following each run. This information provided
Alsoshowonnthespelotsisthefrequencraynge
an indication of whether desired or adequate
ofthedynamieclastimcodetshatwereincluded
performance tolerances had been achieved, thus
intheearlie(rLaRC1.)assessm1ent.
helping the pilot to navigate through the Cooper-
O Harper decision tree.
1.2_ NASA TN D-7349, 1973
In addition to Cooper-Harper Ratings, the
pilots were asked to provide an assessment of the
extent to which dynamic elastic effects adversely
impacted their control inputs and ride quality.
- 0 - 1.8 in peaks The rating scales for these assessments are
shown in Figure 9 and Figure 10.
.5 I.
o DASE INFLUENCE ON RQR
RIDE QUALITY
Cquoaclkitpyi.t vibrations do not impact ride [1]
Cockpit vibrations areperceptible but not IF ,__1
objectionable -no improvement necessary. IL2JI
Cockpit vibrations are mildly objectionable - IF Q_I
improvement desired. ILr.JI
OX Cobojcekcptiiotnvaibblreatio-nsimparroevemmoednetratwealyrranted. [4 l
30 - Oy
Cimopcrkopvietmviebnrattiorenqsuiraerde.highlyobjectionable- [5]
Cockpit vibrations cause abandonment
oftask - improvement required. IkOj
Figure 9. Evaluation scale for dynamic
Figure 8. Frequency response of Langley aeroelastic influence on ride quality.
Visual/Motion Simulator as documented in
NASA TN D-7349, 1973. DASE INFLUENCE ON PILOT'S CIR
CONTROL INPUTS
The dynamic elastic portions of the model
caused the motion platform to operate at the Preilsoutltdooefsairncoratfatltfelerxciboinlittyro.l inputs asa [1]
threshold of its capabilities. The simulator Pilot intentionally modifies control inputs r_
provides reasonable performance (0.8 amplitude toavoid excitation offlexible modes. Ik'-JI
ratio vertically and 1.0 amplitude ratio laterally,
Cockpit vibrations impact precision of r,l_
with about 15 deg of phase loss) at the lowest voluntary controlinputs.
structural mode frequencies (1.25 and 1.39 Hz).
Cockpit vibrations cause occasional r/l_
The next two dynamic elastic modes (at 2.0 and involuntary control inputs. Ik"JI
2.1 Hz) lie marginally within the motion
Cockpit vibrations causefrequent r_
platform's capabilities with about 25 deg of involuntary control inputs. Ik"JI
phase loss. Based on cockpit accelerometer time
histories, the facility appeared to provide a Cockpit vibrations cause sustained involun- rc:'_
tary controlinputs orloss ofcontrol. Ik"JI
reasonable representation of the modes that were
included in the dynamic aeroelastic model.
Figure 10. Evaluation scale for dynamic
EVALUATION SCALES aeroelastic influence on pilot's control inputs.
The familiar Cooper-Harper rating (CHR) These supplemental rating scales were
scale was used during pilot evaluations of the designed for this experiment and are intended to
parametric aeroelastic configurations. After target the acceptability of a particular aeroelastic
6
American Institute of Aeronautics and Astronautics
configuratioanpartfromdeficienciethsatthe
signal containing segments from various
pilotmayperceivienthenominafllightcontrol
maneuvers that had been examined in previous
system.Forinstanceif,a pilotawardsthe
HSCT simulations. These segments included a
landintgaskaCHRof4,butprovideasDASE
localizer capture, a glideslope capture, a
ControInlfluencReating(CIRo)f1andaDASE
descending turn, and a rapid pull-up as
RideQualityRating(RQRo)f1,thenwemight
performed during an aborted landing task. Flight
concludtheatthedeficienCtHRisduetopilot
path and track angle command segments from
dissatisfactiwoinththebaselinfelightcontrol
these tasks were combined with varying order
systema,ndnotwiththedynamiaceroelastic
and sign to produce a flight director behavior
characteristics
of that particular configuration. that was not easily anticipated, but was still
representative of actual flight maneuver tasks.
The CIR Scale shown in Figure 10 bears
further discussion. The scale was developed Each of the parametric aeroelastic
based on pilot comments from the earlier configurations were evaluated using each of the
(LaRC. 1) piloted assessment of dynamic 3 maneuver tasks described above. All tasks
aeroelastic effects. Pilots sometimes indicated were performed in the presence of mild
that they were "reducing the gain" or "backing turbulence that was produced using a Dryden
off'' on their control inputs to avoid excitation of spectra turbulence model.
the elastic modes. Several time histories from
the LaRC.1 test suggested that the cockpit RESULTS AND DISCUSSION
vibration environment had sometimes corrupted
the precision of pilot control inputs, or even Configuration Preference Ranking
caused occasional involuntary stick inputs that
have been referred to as biodynamic feedthlT_. 3 Figure 11 presents a ranking of the 20
The Control Influence Rating Scale was parametric aeroelastic configurations based on
developed to specifically address this issue apart the average of the DASE Control Influence
from the pilot comfort or ride quality issue. Pilot Ratings and Ride Quality Ratings that were
feedback was incorporated during the design of assigned by all the pilots for all three maneuver
the CIR and RQR scales prior to this experiment. tasks performed with a given configuration.
Cooper-Harper Ratings did not seem to
EVALUATION MANEUVERS discriminate among configurations as clearly as
the CIR and RQR ratings which specifically
The tasks that were evaluated during this
targeted DASE effects, and their impact has not
investigation included an up-and-away flight
been included in this ranking.
director tracking task, a nominal approach and
landing, and a lateral-offset landing task. The .......w...L
offset landing was the most challenging of the
three evaluation maneuvers. This task was
initiated at an altitude of 750 ft with a 300-ft
lateral offset and 580 ft longitudinal offset of the g 4.5 cancellation&
instrument landing system (ILS) approach glide- c
slope from the nominal approach path. The pilot rr"
LU 3.s
was directed to fly down the offset glideslope to C9
<
an altitude of 250 ft. At this point, the test
46
conductor called "Correct," and the pilot
_o_ 2.5
executed a lateral correction and vertical descent
©
to re-acquire the runway centerline. The pilot
then executed a flare and attempted to achieve 1.5
touchdown within the tolerances required for qsaeO __
desired performance. The task required an
o.s
aggressive lateral maneuver due to the low 1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 1_ 2_
altitude at which the correction was initiated. PilotPreferenceRanking
The lateral offset landing is a standard evaluation
maneuver used in Calspan's Total In-Flight Figure 11. Pilot preference ranking of
Simulator (TIFS) aircraft flight tests. configurations based on DASE ratings.
The intent of the flight director tracking task The ranking exhibits a number of trends that
was to cause the pilot to exercise awide range of would appear to make intuitive sense. First, the
control input frequencies. The flight director baseline aeroelastic configuration (base0), with
presented on the pilot's HUD was driven with a no structural stiffening or active mode
7
American Institute of Aeronautics and Astronautics
suppressionw,as rankedas the worst
characteristics. On the basis of the average
configurationA.lso reassuringlyth,e rigid
ratings, the first four configurations (qsae0,
(quasi-staetifcfectsonly)configuratiwonithout
canc4, canc3, and damp8) appear to be in or on
anydynamicaeroelasteicffects(qsae0w)as
the border of the acceptable ride quality region.
ranketdhebest.
0.5
Differentiatioanmongconfigurationiss
greateasttthestartoftherankingin, themost
1.5
desirabrleegiona,ndtaperosfftonear-tieastthe
undesirabelendofthespectrumTh.estiffened
casewsithouatctivemodesuppresswioenreall 2.5
rankepdoorlys,uggestitnhgatthisapproacwhas
noteffectivaetreducintgheimpacotfcockpit 3.5
vibrationonpilotingtasks.Theconfiguration
withthegreatensutmbeorfmodesdampe(d4) 4.5
atthe greatesletvel(0.3)andwithmodal
cancellatio(cnanc4w)asratedthebestofthe
5.5
flexibleconfigurationbsu,t wasstill very
noticeablydifferentfromtherigid aircraft
6.5
(qsae0T).henext-becsatsewasidenticatolthis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
onewiththeexceptioonfdampinlgevelw, hich _*--_: Pilot Preference Ranking worst
wasreducetdo0.15(canc3T).hefourthranked
configuratihoand0.3dampinogn4modesb,ut Figure 12. Average ride quality rating vs.
hadnocancellati(odnamp8). preference ranking of parametric configurations.
The orderof rankingprovidessome
Also shown on the plot are the maximum
interestinignsightsregardinpgotentiatlrades
and minimum ride quality ratings assigned to the
betweemnode-canceclionngtroalndadditional
configurations. On the basis of the maximum
dampingA.notheirnsighits gainedwhenwe
rating, even the most highly mode-suppressed
compatrheerankingosftheconfiguratiwonhich
configuration (canc4) provides only marginally
hadonlysymmetrimcodesdamped(damp9,
acceptable ride quality at the pilot station. It
ranked18th)withtheconfiguratiownhichhad
should be noted that these ratings were provided
onlyantisymmetmricodesdamped(dampl0,
during tasks that were performed with mild
ranke9dth).It iscleatrhatthepilotsfoundthe
turbulence, and that the ride quality acceptability
undampeadntisymmetmricotionstobemore
will probably vary with turbulence level. This
problematthicanundampseydmmetrmicotions.
plot provides a subjective basis for the judgment
A comparisoonfthedisp0configuration of an acceptable level of mode suppression from
withthedampcl onfiguratioinndicatethsatthe the pilot's ride-quality perspective.
eliminatioonfvibration-indupceerdturbatioinns
thevisualscenperovideldittleornobenefit Control Influence Ratings vs. Pilot Preference
accordintogthisrankingT.heseconfigurations
This chart provides an analogous ranking to
wereidenticainl allrespecotsthetrhanthelack
the previous chart in terms of the Control
ofvibration-indupceerdturbatioinntsheout-the-
Influence Rating instead of Ride Quality Rating.
windowscenfeorthedisp0configuration.
The CIR assigned by the pilots is shown adjacent
to the control influence axis, along with shading
Ride Quality Ratings vs. Pilot Preference to indicate the transition from acceptable to
marginal to unacceptable configuration charac-
The ranking in the previous figure provides
teristics. Based on discussions with the test
insight regarding the order of pilot preference for
pilots prior to the experiment, it was decided that
the 20 parametric configurations, but does not
the unacceptable threshold should be placed at
indicate the point in the ranking at which
the point at which cockpit vibrations impact the
dynamic elastic characteristics make the
precision of voluntary control inputs.
configuration unacceptable. Figure 12 shows the
average Ride Quality Rating assigned by the On the basis of the average control influence
pilots for each configuration plotted against the ratings, the first four configurations (qsae0,
Pilot Preference Ranking from the previous canc4, canc3, damp8) again lie within the
figure. The RQR assigned by the pilots is shown acceptable region. Each of the acceptable elastic
adjacent to the ride quality axis, along with configurations applies a damping level of 0.3 to
shading to indicate the transition from accept- the first two symmetric and first two
able to marginal to unacceptable configuration antisymmetric modes. On the basis of the
American Institute of Aeronautics and Astronautics
maximum rating, the most highly mode- accelerations in g's (dashed line) and lateral stick
suppressed configuration (canc4) again lies in deflections (solid line). Although the units on
the marginal control influence region. the two quantities differ, the scaling of +1 is
0.5 convenient since it represents the maximum
throw for lateral stick deflection and since lateral
g's commanded by the simulation remained in
1.5
the range of + 1 g. The plot in the lower left of
the figure shows the power spectral density of
2.5 lateral accelerations and lateral stick deflections
applied to a 6-second segment of the time history
3.5 (from 37 to 43 seconds).
The frequency spectrum of the pilot's
4.5
voluntary control inputs during this period lies
primarily below 1 Hz. The frequency spectrum
5.5 of the lateral accelerations at the pilot station
shows some content at the first antisymmetric
6.5 mode frequency of 1.6 Hz and the second
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
antisymmetric mode frequency of 2.5 Hz due to
m:_ Pilot Preference Ranking [Wo_
minor turbulence excitation of these structural
modes. (Recall that the frequencies shown in
Figure 13. Average control influence rating vs.
Figure 2 were multiplied by a factor of 1.16 to
preference ranking for parametric configurations
produce the "stifl" modified baseline.)
The border between acceptable and marginal
control influence on this plot is somewhat Time History: Offset Landing Maneuver Task, stir I Configuration
arbitrary, since it might be perfectly acceptable
for the pilots to intentionally modify their control
inputs to avoid excitation of the dynamic elastic
modes as long as their ability to precisely control
the aircraft is in no way hindered by this ,l
practice. However, recorded time histories of
35 40 45 50 time, sec
pilot stick deflections indicate that pilots were Power Spect=ra
sometimes unaware that cockpit vibrations were L ]
........
in fact impacting their control inputs. Sustained
I Feedback
The occurrence of biodynamic feedthru of __L_ of Cockpit
[_T_ Vibrations
cockpit vibrations through the pilot's arm and
back into the control stick is involuntary and zI: ] [",--"LateralStictk1
,=="LateralAccel
therefore may indeed be unnoticed by the pilot in Frequency, Hz I,[ _,'/\: I
minor instances. Use of the Control Influence
Voluntarylnputs 0 Frequency,2 Hz 4 IO_o!y''y/
Rating scale shown in the figure requires the
Input Contamination [ Frequency, Hz
pilot to be aware of the occurrence, and therefore
the CIR ratings may sometimes be optimistic. Figure 14. Power-spectrum analysis of
There were, however, a number of profound biodynamic coupling incident for pilot B.
instances of frequent or sustained biodynamic
The power spectrum of a later 6-second
feedthru of cockpit vibrations into the control
segment of the time history (from 41 to 47
stick as indicated by the maximum CIR ratings
seconds) indicates the bulk of the pilot's input
shown on the figure. Frequent or sustained
spectrum remains below 1 Hz, but it also shows
feedthru of cockpit vibrations through the pilot's
arm and back into the stick will be referred to as some frequency content of the pilot's inputs in
the range of the lateral elastic modes. Once the
Biodynamic Coupling (BDC) in this report.
pilot begins to move the stick at the resonant
Example of Biodvnamic Coupling Incident frequency of the first antisymmetric structural
mode there is tremendous potential for the lateral
Figure 14 presents a power-spectrum mode to be excited by the control inputs,
analysis of a lateral offset landing run in which producing larger lateral accelerations at the pilot
the pilot experienced biodynamic coupling while station. These lateral accelerations can move the
flying the stifl configuration with no additional pilot's frame in a fashion which produces
damping or cancellation. The time history at the involuntary control inputs that further excite the
top of the figure shows lateral cockpit structural mode.
9
American Institute of Aeronautics and Astronautics