Wind-Tunnel Free-Flight Investigation of a Supersonic ...mln/ltrs-pdfs/tp3258.pdfprogram bewteen the...
56
NASA Technical Paper 3258 February 1993 Wind-Tunnel Free-Flight Investigation of a Supersonic Persistence Fighter David E. Hahne, Thomas R. Wendel, and Joseph R. Boland
Transcript of Wind-Tunnel Free-Flight Investigation of a Supersonic ...mln/ltrs-pdfs/tp3258.pdfprogram bewteen the...
NASATechnicalPaper3258
February 1993
Wind-Tunnel Free-FlightInvestigation of a SupersonicPersistence Fighter
David E. Hahne,Thomas R. Wendel,and Joseph R. Boland
NASATechnicalPaper3258
1993
Wind-Tunnel Free-FlightInvestigation of a SupersonicPersistence Fighter
David E. HahneLangley Research CenterHampton, Virginia
Thomas R. Wendeland Joseph R. BolandMcDonnell Aircraft CompanySt. Louis, Missouri
The use of trademarks or names of manufacturers in this
report is for accurate reporting and does not constitute an
o�cial endorsement, either expressed or implied, of such
products or manufacturers by the National Aeronautics and
Space Administration.
Summary
Wind-tunnel free- ight tests were conducted inthe Langley 30- by 60-Foot Tunnel to examine thehigh-angle-of-attack stability and control characteris-tics and control law design of a supersonic persistence�ghter (SSPF) at 1g ight conditions. The SSPF in-corporated a 65� arrow wing, twin vertical tails, anda canard. The SSPF was also equipped with un-conventional controls including de ectable wingtips(tiperons) and pitch and yaw thrust vectoring inaddition to conventional control surfaces. Beforethe free- ight tests, a direct eigenstructure assign-ment technique was used to design control laws thatblended these unconventional and conventional con-trol surfaces. The combined controls were intendedto provide good ying characteristics well into thepoststall angle-of-attack region.
In general, the SSPF exhibited good ying char-acteristics up to an angle of attack of 80�. Flightsmade with reduced feedback gains indicated signif-icant robustness in the control law design. Use ofthrust vectoring, blended with conventional pitchand yaw control surfaces, provided good stabilityand control characteristics throughout the angle-of-attack range tested. The tiperons, coupled with con-ventional ailerons, provided adequate roll control upto an angle of attack of 70�. Overall, free- ight testsindicated that it was possible to blend e�ectivelyconventional and unconventional control surfaces toachieve good ying characteristics well into the post-stall angle-of-attack region.
Introduction
The desire for aircraft with sustained and e�cientsupersonic cruise performance has resulted in con-�gurations with high-�neness-ratio fuselages, highlyswept low-aspect-ratio wings, and highly integratedcontrol surfaces. Con�gurations that incorporatethese features, which are conducive to low cruise drag(ref. 1), generate strong vortical ows. These vortical ows and their breakdown can lead to nonlinear aero-dynamic characteristics and high levels of instability.Powerful control devices that maintain their e�ec-tiveness at high angles of attack must be used tostabilize and control such con�gurations adequately.The challenge is to develop a ight control systemthat blends these control devices in order to main-tain good ying characteristics well into the post-stall angle-of-attack region. One method for evaluat-ing both the stability and control characteristics of acon�guration and the e�ectiveness of a ight controlsystem is the use of the wind-tunnel free- ight testtechnique.
Wind-tunnel free- ight tests have been conductedat the Langley Research Center since the late 1930's(ref. 2). From its early beginnings to the present in-vestigation, the free- ight test technique has focusedon obtaining qualitative data about the dynamic sta-bility and control characteristics of aircraft at mod-erate to high angles of attack. Free- ight testing ofdynamically scaled models is an important test tech-nique for understanding the complex aerodynamicsand nonlinear ight mechanics associated with mod-ern �ghter aircraft at high angles of attack. Free- ight testing can also qualitatively evaluate potential ight control systems. With properly scaled massand inertial characteristics, a scale model of an air-craft can be thought of as a simulator that has allthe vehicle aerodynamics and interactions properlymodeled.
The con�guration used in the present investi-gation, called the supersonic persistence �ghter orSSPF (ref. 1), is shown in �gure 1. The SSPF con-�guration is the result of a series of previous wind-tunnel studies conducted as part of a cooperativeprogram between the NASA Langley Research Cen-ter (LaRC) and the McDonnell Aircraft Company(MCAIR). The purpose of these studies was to de-velop a low-speed design data base for supersoniccruise con�gurations (refs. 1, 3, 4, and 5). Past inves-tigations focused on the e�ect of component integra-tion on airframe stability characteristics and the de-velopment of advanced control devices such as thrustvectoring and the use of de ectable wingtips. Re-sults of previous investigations of the static and dy-namic aerodynamic characteristics of the SSPF arepresented in reference 6. The present report willfocus on the results of a recent free- ight investiga-tion of the SSPF. Limited results from reference 6are presented to aid in the analysis and discussion ofresults from the recent free- ight tests. A discussionof the control law development process that precededthe free- ight tests is also presented.
Symbols
All data were initially measured in the body-axis system shown in �gure 2. Longitudinal forceand moment data are presented in the stability-axissystem; lateral-directional force and moment dataare presented in the body-axis system.
b wingspan, ft
CD drag coe�cient, Drag
�qS
CL lift coe�cient, Lift
�qS
CL;max maximum lift coe�cient
Cl rolling-moment coe�cient, Rolling moment�qSb
Cm pitching-moment coe�cient,Pitching moment
�qS�c
Cn yawing-moment coe�cient,Yawing moment
�qSb
CY side-force coe�cient, Side force�qS
�c mean aerodynamic chord, excludingtrailing-edge extension, ft
f frequency of oscillation, cycles/sec
g acceleration due to gravity, 32.17 ft/sec2
Ix; Iy; Iz mass moments of inertia about X, Y ,
and Z body axes, slug-ft2
k reduced-frequency parameter, !b2V
Ny lateral acceleration, g units
p; q; r angular velocity about X, Y ,and Z body axes, rad/sec
�q free-stream dynamic pressure, lb/ft
S wing area, excluding trailing-edge
extension, ft2
s frequency domain independent vari-able, 1/sec
u; v; w linear velocity along X, Y , and Z bodyaxes, ft/sec
V free-stream velocity, ft/sec
X;Y;Z body axes
� angle of attack, deg
� angle of sideslip, deg
_� rate of change of sideslip, rad/sec
�Cl incremental rolling-moment coe�cient
�Cn incremental yawing-moment coe�cient
�CY incremental side-force coe�cient
�a aileron de ection, positive for left roll,deg
�c canard de ection, positive trailing-edge down, deg
�F trailing-edge extension ap de ection,positive trailing-edge down, deg
�f leading-edge ap de ection, positiveleading-edge down, deg
�pv pitch-vane de ection, positive trailing-edge down, deg
�r rudder de ection, positive trailing-edge left, deg
�wt wingtip de ection, positive for left roll,deg
�yv yaw-vane de ection, positive trailing-edge left, deg
�d dutch roll damping ratio
�sp short period damping ratio
�R roll-mode time constant, sec
! angular velocity, 2�f , rad/sec
!d dutch roll frequency, rad/sec
!sp short period frequency, rad/sec
Stability derivatives:
Clp =@Cl
@pb2V
Clr =@Cl
@ rb2V
Cl� =@Cl@�
Cl _�=
@Cl
@_�b2V
Cmq =@Cm
@q�c2V
Cm� = @Cm@�
Cnp =@Cn
@pb2V
Cnr =@Cn
@ rb2V
Cn� =@Cn@�
Cn _�= @Cn
@_�b2V
CYp =@CY
@pb2V
CYr =@CY
@ rb2V
CY� =@CY@�
CY _�=
@CY
@_�b2V
Cn�;dyn = Cn� cos��IzIxCl� sin�
Abbreviations:
alpha �ltered angle of attack
BetaF �ltered angle of sideslip
BetDot estimated rate of change of sideslip
BL butt line
c.g. center of gravity
DEA direct eigenstructure assignment
FCL ight control laws
LE leading edge
MCAIR McDonnell Aircraft Company
MS model station
Pejector ejector pressure
PIO pilot-induced oscillations
Pstab stability-axis roll rate
qbar tunnel free-stream dynamic pressure
2
Rstab stability-axis yaw rate
RTES real-time-engineering simulation
SQR square root
SSPF supersonic persistence �ghter
TEX trailing-edge extension
Model
Tests were made with a 0.14-scale model of theSSPF in the Langley 30- by 60-Foot Tunnel. A sketchof the SSPF and details of the wing, canard, verticaltails, and control surfaces can be found in �gure 3.Geometric, mass, and inertial characteristics of theSSPF free- ight model are presented in table I. Themodel had an arrow wing with a 65� swept leadingedge and an aspect ratio of 1.95. A close-coupled ca-nard was mounted just above the engine inlets. De- ectable surfaces on the wing included leading-edge aps, ailerons, and tiperons (de ectable wingtips).In addition to the canard, a ap at the end of thetrailing-edge extension (TEX) was used for pitch con-trol (�g. 3(b)). Twin vertical tails, incorporatingconventional rudders, were canted inboard 15� andmounted on the outboard edge of the trailing-edgeextension. The model was also equipped with thrustvectoring in both pitch and yaw axes. Angular de- ections of all moving surfaces were measured per-pendicular to their respective hinge lines ; the rangesof de ections are given in table II.
During free- ight testing, the model was equippedwith two multiport ejectors (�g. 4(a)) supplied withcompressed air to generate thrust. Secondary airfrom the model engine inlets was entrained withthe high-pressure air from the ejector. The primary(high-pressure) air and the secondary (inlet) air weremixed as they owed to the exhaust nozzles. Aphotograph of the thrust-vectoring vane arrangementis presented in �gure 4(b). Geometric details of thevanes are given in table I. The ejectors and thrust-vectoring vanes, used only during free- ight tests,were calibrated at wind-o� conditions. The modelwas unpowered, with ow through inlets, duringconventional static and dynamic force and momenttests.
Test and Apparatus
Static and Dynamic Tests
Static and dynamic force tests were conducted inthe Langley 30- by 60-Foot Tunnel at a free-streamdynamic pressure of 10 psf, which corresponded to aReynolds number of 1:89 � 106, based on the wingmean aerodynamic chord. Aerodynamic force and
moment data were measured with an internal six-component strain-gauge balance. Static data wereobtained over an angle-of-attack range of 0� to 65�
at angles of sideslip of 0� and �5�. These datawere obtained for a moment reference center of 0.36�c,which corresponds to the one used for free- ight tests.Flow angularity corrections were made for both angleof attack and angle of sideslip. Basic aerodynamicdata, including static stability derivatives and controle�ectiveness, were obtained during static force tests.Lateral-directional derivatives were calculated fromdata obtained at the � = �5� conditions.
A second investigation in the Langley 30- by60-Foot Tunnel was conducted to determine the aero-dynamic damping characteristics of the SSPF inthe roll and yaw axes. A small-amplitude forced-oscillation technique combined balance force and mo-ment outputs with the known angular position of themodel to calculate aerodynamic damping character-istics (see ref. 7 for a complete description of theforced-oscillation test technique). These tests wereconducted at a dynamic pressure of 10 psf with a mo-ment reference center of 0.38�c. Data were obtainedover an angle-of-attack range of 0� to 90�. Forced-oscillation tests were conducted at an amplitude of�5� and a frequency of 0.75 Hz. Use of this frequencyresulted in a reduced-frequency parameter k of 0.13.All captive force and moment tests were conductedwith ow-through inlets and the model unpowered.
Free- ight tests
Free- ight tests were conducted in the Langley30- by 60-Foot Tunnel to assess stability and con-trollability of the SSPF and the e�ects of control lawdesign on these characteristics. All ights were madewith a moment reference center of 0.36�c. During free- ight tests, the model was powered by compressed air(using the ejector system previously described) andwas own unrestrained in the open-throat test sec-tion of the tunnel (�gs. 5(a) and (b)). The conditionsrepresented 1g, wings-level ight; angle of attack wasvaried by trimming the model at di�erent dynamicpressures. Flights were conducted over an angle-of-attack range of 20� to 80�. The model was remotelycontrolled by three pilots: a roll and yaw pilot, apitch pilot, and a thrust pilot. Air lines and signalwires were contained in an umbilical line that ledfrom the top of the test section to the model. During ights the umbilical was kept slack by a safety ca-ble operator to minimize its e�ect on the model mo-tions. A sketch of the free- ight test setup is shownin �gure 5(c).
The model was equipped with a three-axis rategyro for measuring body-axis pitch, roll, and yaw
3
rates. A miniaturized �=� vane sensor was boommounted from the model nose. A three-axis ac-celerometer was also installed in the model to mea-sure body-axis accelerations. Only the lateral accel-eration signal was intended for use during these tests.Output from these sensors was used to augment thestability characteristics of the SSPF through the useof a y-by-wire control system. The primary compo-nent of this y-by-wire control system was a digitalcomputer programmed with the control laws. Thissystem was designed to allow in- ight variations ofkey control law parameters such as gains, �lter con-stants, and control surface command limits. Thecomputer combined pilot command signals with datasignals from the model sensors and computed the ap-propriate control surface commands. The controlsurfaces, moved using electropneumatic actuators,were capable of moving the model control surfaces atmore than 120�/sec. All inputs to the computer andall commands to the model were updated at 10-msecintervals.
Control Law Development
The control laws for the SSPF free- ight modelwere generated using a direct eigenstructure assign-ment (DEA) synthesis technique for angles of at-tack up to 65�. Because aerodynamic data did notexist above an angle of attack of 65�, the controllaws were linearly extrapolated for ights made be-yond this angle of attack. This multivariable, model-following technique uses a set of desired eigenvaluesand eigenvectors as its design goals. The desiredeigenspace is chosen so that the dynamic system itdescribes will meet a given set of ying-qualities cri-teria. A weighted least-squares-solution algorithm isthen used to obtain the control gains.
DEA is an output feedback formulation that doesnot o�er the guaranteed stability margins of linearquadratic techniques; however, it is not constrainedto full-state feedback so the control system designercan specify the variables used as feedbacks. TheDEA technique also has the potential to includemany higher order dynamics in the system modelwithout a signi�cant increase in complexity of thecontrol system. Further discussion and examples ofthe methodology and use of DEA can be found inreferences 8 through 13.
The ying-qualities design goals for the SSPFwere initially based on requirements speci�ed in ref-erence 14. However, these guidelines were devel-oped at low angles of attack; recent studies (refs. 15through 17) show that di�erent design goals are re-quired at higher angles of attack. A set of design
goals (short period, dutch roll, and roll-mode char-acteristics) was determined with the guidelines ofreferences 14 through 17. Although full-scale aircraftmodes could be used as given, the desired dynam-ics could not be determined until the speci�ed fre-quencies were increased to model scale. Guidelinesfor control system robustness of 6 dB gain marginand 45� phase margin were determined from refer-ence 18. These robustness guidelines were strictly fol-lowed throughout the control system design process.
The model of desired dynamics for the SSPF,to be used in the control law synthesis, was devel-oped primarily with the VECTOR program (ref. 19).VECTOR allows the designer to determine the air-craft control e�ector requirements and stability aug-mentation capability for a con�guration from basicaircraft geometry and aerodynamic data (�g. 6(a)).Inputs to VECTOR include geometry, weight and in-ertias, aerodynamic characteristics (such as CL andCD), and control surface rate and de ection limits.A simple engine model that represented the ejec-tors used in the free- ight tests was used for theVECTOR inputs. After the data entries have beenmade, VECTOR allows the designer to vary thesedata inputs parametrically to study their e�ect on ying characteristics. VECTOR results for the SSPFwere used to determine the desired eigenstructuremodels necessary for the DEA control law synthesis.
A MCAIR in-house program entitled SCHEDULEwas also used extensively during the development ofthe SSPF free- ight control laws. Using SCHEDULE,the designer can quickly set up control surface sched-ules for trim and assess the e�ect that these scheduleshave on lift, drag, and stability and control charac-teristics (�g. 6(b)). SCHEDULE can also be usedto evaluate the impact of trim de ections on aero-dynamic control requirements and center-of-gravity(c.g.) movement.
A parametric study was conducted to determinethe best c.g. location for the SSPF. The �nal choiceof c.g. location (0.36�c) represented a trade-o� be-tween stability, control power, and trim require-ments. SCHEDULE was then used to develop a ca-nard schedule, based on angle of attack, to improvethe basic nonlinear longitudinal stability character-istics of the airframe. Figure 7 illustrates the sta-bility improvements obtained from the SCHEDULEanalysis. The canard schedule eliminated a pitch-upproblem and resulted in a uniform stability level atangles of attack less than 30�. SCHEDULE was alsoused to ensure that the �nal canard schedule did notadversely a�ect the lateral-directional stability char-acteristics of the SSPF.
4
With the given model of desired dynamics de-veloped from the VECTOR results, the control lawsynthesis was performed using the DEA technique.The block diagrams for the resulting control laws,including additional switches used during free- ighttests for evaluation of the control system, are pre-sented in �gures 8(a) through (e). The longitudinalaxis was an angle-of-attack command system. Pro-portional angle-of-attack and pitch-rate feedbackswere used to stabilize the airframe and provide thedesired longitudinal ying qualities. The lateral axisused a roll-rate command system; the directional axisused a sideslip-angle command system. Cross feedswere included between the roll and yaw commandsto ensure roll and yaw coordination without the useof pilot yaw command inputs. For the 1g, wings-level ight conditions encountered during free- ighttests, these cross feeds allowed simultaneous lateraland directional axis control. Roll rate, yaw rate, an-gle of sideslip, and estimated time-rate-of-change ofsideslip angle were used as feedbacks in the lateral-directional axes to augment stability and to improve ying qualities. The control law design blended allavailable control devices, including pitch and yawthrust vectoring, as a function of angle of attack. Thebaseline gain schedules, determined from the controllaw synthesis and initial free- ight tests, and the �-nal gain schedules are shown in �gure 9. These twogain schedules are listed in appendixes A and B.
The SSPF control laws were evaluated with bothlinear and nonlinear methods. The linear evaluationmethods included equivalent systems and stabilitymargin analysis. The results of the linear analysisare shown in �gure 10. Although the desired yingqualities were not achieved at every angle of attack,in general the desired characteristics were obtained.The stability margin analysis showed that gain mar-gins above 15 dB and phase margins beyond 60� wereachieved for the entire angle-of-attack range. A fullnonlinear six-degree-of-freedom batch simulation wasalso conducted for the SSPF. A �nal step in the anal-ysis of the SSPF free- ight control laws was a real-time piloted simulation with the MCAIR real-time-engineering simulation (RTES) package (�g. 11),which consists of a Silicon Graphics IRIS work sta-tion and a Digital Equipment VAX computer. TheIRIS provided the pilot-vehicle interface and the re-quired graphics; the six-degree-of-freedom equationswere computed on the VAX. This simulation pro-vided a pilot-in-the-loop validation of the control lawdesign. Because the aircraft could be own from anyperspective, including outside the aircraft, the free- ight test environment could be simulated. All theevaluations indicated that the stability margins and
control response of the SSPF met the desired ying-qualities and robustness guidelines. A detailed ac-count of the SSPF control system design process canbe found in reference 20.
Captive Test Summary
The purpose of the present investigation was touse the free- ight test technique to evaluate the SSPF ying characteristics and ight control system. Asummary of the existing data base on the SSPF ispresented to facilitate analysis of the free- ight re-sults. Both static and dynamic force and momentdata are presented with the leading-edge ap de- ected (�f = 30�) to correspond with the con�gu-ration used during free- ight tests. Additional dataand further discussion of these results are found inreference 6.
Static Force Tests
Results from static force tests are presented in�gures 12 through 15(a) and 15(b). The pitching-moment data of �gure 12 show that the SSPF wasslightly unstable for most canard de ections with aslight pitch-up near � = 25�. Although this phe-nomenon is common for highly swept wings, the on-set angle of attack was lower than expected becauseof the placement of the vertical tails (ref. 4). Withthe canard de ected �40�, the con�guration was sta-ble for low angles of attack and exhibited a severepitch-up near � = 15�. This change in the pitching-moment characteristics, common for large negativecanard de ections (see refs. 21 and 22), is believedto result from the canard wake interacting with thewing ow �eld. For angles of attack of about 30�
to 35�, a sharp, stable break was evident in pitchingmoment for all canard de ections.
The canard was scheduled with angle of attack toimprove the longitudinal stability characteristics ofthe SSPF, and linear pitching-moment characteris-tics (�g. 8) were achieved up to an angle of attack of30�. Beyond an angle of attack of 30�, canard e�ec-tiveness decreased rapidly, and other means of sta-bility augmentation were required. The pitch controlprovided by the TEX aps, shown in �gure 13, wasadequate for stability augmentation and trim up toCL;max. At poststall angles of attack, all trim capa-bility and stability augmentation in the longitudinalaxis were provided by thrust vectoring. The controle�ectiveness associated with thrust vectoring will bediscussed later.
The static lateral-directional characteristics ex-hibited by the SSPF are presented in �gure 14 andare generally representative of modern �ghter air-craft (refs. 21 and 22). Lateral stability increased
5
sharply with increasing angle of attack for low an-gles, as expected for a highly swept wing con�gura-tion. The expected decrease in lateral stability nearmaximum lift, also common with highly swept wings,was reduced by the in uence of the vertical tails (seeref. 4). The placement of the vertical tails causedsymmetric bursting of the strong wing leading-edgevortices, thereby delaying the decrease in static lat-eral stability. Directional stability decreased withangle of attack, and the con�guration became un-stable at angles of attack above 20� to 25�. For� > 45�, the forebody vortex system dominated the ow �eld and resulted in a stabilizing increment toCn�
(ref. 2). At angles of attack between 15� and
30�, nonzero canard de ections decreased directionalstability. In the poststall angle-of-attack region, onlythe larger canard de ections a�ected Cn�
. With the
canard highly loaded (�c = 20�) directional stabil-ity was decreased, whereas with the canard unloaded(�c = �40
�) directional stability was improved.
Lateral-directional control characteristics areshown in �gures 15(a) and (b). The ailerons, thoughquite e�ective at low angles of attack, rapidly becameine�ective past � = 20� as ow over the wing becameparallel to the aileron hingeline (ref. 2). The tiper-ons, designed to take advantage of this spanwise ow,remained e�ective at higher angles of attack and pro-duced little or no yawing moment. Combined, theailerons and tiperons provided good body-axis rollcontrol throughout the angle-of-attack range tested(�g. 15(a)). Figure 15(b) shows that the twin rud-ders were e�ective for providing yaw control up toan angle of attack of 40�. Beyond that point, rud-der e�ectiveness rapidly decreases; therefore, thrustvectoring in yaw would be necessary for directionalcontrol.
Forced-Oscillation Tests
Results of the forced-oscillation tests are sum-marized in �gures 16 and 17. Negative values forroll damping (C
lp+ C
l_�sin�) and yaw damping
(Cnr � Cn_�cos�) are stable. The SSPF exhibited
stable, although low, roll damping characteristicsthroughout the angle-of-attack range tested. Canardde ection had minimal e�ect on roll damping. Theonly exception was �c = �40�, where roll dampingdecreased to zero near � = 30�. Canard de ectionhad a more pronounced e�ect on yaw damping, par-ticularly at angles of attack near 15� and 40� (�g. 17).Near � = 15�, yaw damping decreased as canard de- ection was changed from �40� to 20�. The oppo-site e�ect, with a larger change in magnitude, wasseen near � = 40�. These results, when combined
with those for static directional stability, indicatethat the canard is interacting with the forebody ow-�eld characteristics. Above about an angle of attackof 50�, the SSPF exhibited unstable yaw damping;changes in canard de ection had no e�ect.
Forced-oscillation tests were not conducted in thepitch axis; however, pitch-damping characteristicswere estimated with the strip theory method de-scribed in reference 23. The pitch-damping esti-mation shown in �gure 18 indicates that the SSPFshould exhibit stable pitch-damping characteristicsthroughout the angle-of-attack range of interest.
Thrust Calibration Tests
Static wind-o� thrust tests were conducted to de-termine the e�ectiveness of the pitch and yaw thrust-vectoring vanes. The results from these tests are pre-sented in �gure 19 for two thrust levels. Changesin pitching moment with pitch-vane de ection werelinear for de ections between �10�. For de ectionsgreater than +10� (nose down), the pitch vaneswere slightly less e�ective; for de ections less than�10�, the pitch vanes were slightly more e�ective.This trend was independent of thrust level. Yaw-ing moment exhibited linear behavior for yaw-vanede ections between �20�, with a slight decrease ine�ectiveness for greater de ections. The turning ef-�ciency for both the pitch and yaw vanes, de�nedas the ratio of thrust de ection to vane de ection,averaged about 55 percent. Variation was less than3 percent as thrust levels and de ection angles varied.
Figure 20 illustrates the pitch and yaw controlavailable with thrust vectoring. The crosshatchedarea represents the pitch and yaw control envelope asa function of angle of attack. These coe�cients arebased on dynamic pressures and thrust levels calcu-lated from the aerodynamic data base for the SSPFfor trimmed, 1g ight (see �g. 21). As angle of attackincreases, thrust levels increase and dynamic pressuredecreases; the result is increased pitch and yaw con-trol e�ectiveness. Even at an angle of attack of 20�,thrust-vectoring control far exceeds the e�ectivenessof the conventional aerodynamic controls.
Free-Flight Test Results
To aid in the analysis of the free- ight results,the trim stability derivatives for the SSPF were cal-culated for wings-level, 1g ight conditions based onthe aerodynamic data presented earlier. These calcu-lations were made with the ight control laws (FCL)inactive and for the baseline and the minimum FCLgains determined during free- ight tests. The re-sults from this analysis are presented in �gures 21
6
through 25. The yaw divergence parameter Cn�;dyn
was calculated from these results and is also included.The results in �gures 22 through 25 were based onthe predicted values of dynamic pressure and thrustlevels at trimmed ight conditions determined fromthe data base.
Baseline Longitudinal Characteristics
Initial free- ight investigations of the baseline ight characteristics of the SSPF were made at an-gles of attack between 21� and 50�. The baselinecon�guration consisted of the SSPF geometry shownin �gure 3, with the leading-edge aps de ected 30�
and the gain schedules listed in appendix A. Initial ights indicated that the baseline control system forthe SSPF provided su�cient stability augmentationfor good ying characteristics in pitch throughout theangle-of-attack range tested. Good ying character-istics can be de�ned as follows: (1) good response,(2) quick damping of disturbances, and (3) low pi-lot work load. Pitch control, which was blendedbetween the TEX aps and pitch vanes dependingon angle-of-attack, was good at all angles of attacktested without any noticeable loss in control powerat the higher angles of attack. Reduced pitch-vanegains compensated for the higher thrust levels asso-ciated with high-� ight conditions. This gain re-duction successfully prevented potential overcontrolof the model in pitch.
Longitudinal Gain Variations
The longitudinal characteristics of the SSPF werefurther evaluated by reducing the level of angle-of-attack and pitch-rate feedback to the TEX aps andpitch vanes. At angles of attack below 30�, good y-ing characteristics could be maintained even whenthe � feedback was reduced by 50 percent. Themodel could be own, however, at angles of attackbelow 30� without � feedback, even though the con-�guration was statically unstable (�g. 22(a)). Thehigh levels of augmented pitch damping (�g. 22(b))were su�cient to maintain control without � feed-back. Pilot workload, however, was signi�cantlyincreased without � feedback, and the model washighly susceptible to pilot-induced oscillations (PIO)which resulted in unacceptable ying characteris-tics in pitch. Above an angle of attack of 30� theSSPF was statically stable in pitch, and ights madeat higher angles of attack indicated that � feed-back, as expected, was not necessary for good yingcharacteristics.
Variations of pitch-rate feedback indicated thatthis gain could be reduced 50 percent and still main-tain good pitch-damping characteristics throughout
the angle-of-attack range tested. Although the con-�guration could not be own without pitch-rate feed-back, ights below an angle of attack of 30� werepossible with pitch-rate feedback reduced by 75 per-cent. With this reduced gain, however, the pitch-damping characteristics of the SSPF were marginal.Pitch oscillations were slow to damp out, and con-stant pilot attention was necessary to avoid a pitchdeparture. Flights were also made with both angle-of-attack and pitch-rate feedbacks reduced. These ights indicated that angle-of-attack and pitch-ratefeedback gains could both be reduced to 50 percentof nominal values and still maintain good ying char-acteristics with light pilot workload.
Baseline Lateral-Directional
Characteristics
Initial free- ight evaluations of the lateral-directional characteristics of the SSPF were made atangles of attack between 21� and 50�. During theseinitial ights the model accelerometer malfunctionedand resulted in a lateral acceleration signal that wasunreliable. Because of this malfunction, the washed-out yaw rate and lateral acceleration portions of the_� estimator were eliminated from the control lawsby setting Y k10 = 0 (see �g. 8(e)). This modi�ed_� estimator was used for all ights discussed in thisreport.
The original � feedback gain provided highly sta-ble values of Cn�
at all angles of attack. This high-
gain � feedback system, combined with the highgain on the modi�ed _� feedback, caused some prob-lems during initial checkout ights. Therefore, thegains on these two feedbacks were reduced to 10 per-cent of the original values. All references to thebaseline control system include this reduction in the� and _� feedback gains. These initial ights alsoindicated that, for angles of attack below 25�, �and _� feedbacks to the yaw control surfaces over-stabilized the con�guration. It is believed that theloss of the lateral acceleration component in the _�estimator was partially responsible for this behavior.As angle of attack increased above 25�, � and _� feed-back did not adversely a�ect ying characteristics.However, these feedbacks did not appear necessaryat high angles of attack, even though the con�gu-ration became directionally unstable (�g. 23(a)). Apossible explanation for this result is indicated bythe directional divergence parameter Cn�;dyn
shown
in �gure 24. Positive values of this parameter indi-cate that a con�guration has a tendency to resist ayaw divergence (refs. 4 and 24). The stable values ofCl�
exhibited by the SSPF (�g. 23(b)) are su�cient
to maintain positive values of Cn�;dynat all angles
7
of attack tested. Because of this tendency to resist ayaw divergence, the SSPF appears more stable direc-tionally than the static data of �gure 23(a) indicate.
Because � and _� feedback to the yaw control surfacesdid not enhance the static and dynamic directionalstability characteristics of the SSPF, all remaining ights were made without these feedback paths.
Control characteristics of the SSPF were goodthroughout the angle-of-attack range tested. Goodcoordination between the roll and yaw axes wasmaintained by the cross feeds discussed previously.The aileron and tiperon surfaces provided good body-axis roll control at all angles of attack (see staticdata in �g. 15(a)). The yaw thrust vectoring, asexpected, provided ample yawing moment for bothdirectional control and directional stability augmen-tation. Damping characteristics in roll and yaw werealso good at all angles of attack, as expected giventhe levels of augmented stability (�gs. 25(a) and (b)).However, above an angle of attack of 40�, the con�g-uration was slightly overdamped in roll, as indicatedby the data of �gure 25(b).
Lateral-Directional Gain Variations
Lateral-directional gain variations included : (1) �
and _� feedback reduction to the ailerons and tiper-ons, (2) roll- and yaw-rate feedback to the aileronsand tiperons, and (3) roll- and yaw-rate feedback tothe rudders and yaw vanes. As the angle of attackwas increased from 20� to 34�, � and _� feedback tothe ailerons and tiperons could gradually be reducedto zero without a�ecting the ying characteristics ofthe model. For all angles of attack beyond 34�, thesetwo feedback paths could be eliminated.
A blend of roll- and yaw-rate feedback to theailerons and tiperons augmented the roll-dampingcharacteristics of the SSPF. Although _� feedbackalso provided some enhancement to the roll-dampingcharacteristics, the blended rate signal was the pri-mary feedback used. Flights were possible through-out the angle-of-attack range without roll- and yaw-rate feedback to the ailerons and tiperons. Flightsbelow an angle of attack of 42� indicated that modi-�cations to the airframe roll-damping characteristicswere necessary for smooth, controlled ights. There-fore, the blended roll- and yaw-rate feedback was in-creased to the baseline values as the angle of attackdecreased from 42� to 21�.
To augment yaw damping, roll and yaw rates wereblended for feedback to the rudders and yaw vanes.Flights were made throughout most of the angle-of-attack range with roll- and yaw-rate feedback to the
rudders and yaw vanes reduced by 50 percent. At an-gles of attack below 30� and above 50�, the model was yable with a 50-percent reduction in yaw-dampingaugmentation, although values closer to the baselinegains provided better ying characteristics and lowerpilot workload. Flights made with reduced lateral-directional feedbacks indicated that the minimumgain levels determined with isolated gain reductionstudies could be combined and still result in a sta-ble, easily own con�guration.
Final Gain Schedules
Two ights were made at angles of attack be-tween 21� and 80� using all the modi�ed gain sched-ules determined in the gain variation studies. Thesemodi�ed gains are shown in �gure 9 and listed inappendix B. Results from these ights indicatedno unfavorable interactions between the longitudi-nal and lateral-directional axes as a result of combin-ing the individual gain reductions. The SSPF, usingthe modi�ed gain schedules, exhibited good stabil-ity characteristics throughout most of the angle-of-attack range tested, with adequate control authorityto maneuver the model. Above an angle of attackof 70�, a signi�cant loss in body-axis roll control re-quired the pilot to rely on yaw control to maneu-ver the model in the lateral-directional axes. Finally,pitch damping above an angle of attack of 75� wasalso degraded, and the model exhibited small lon-gitudinal oscillations similar to those typically seenwhen damping is low.
Concluding Remarks
Wind-tunnel free- ight tests have been conductedto examine the high-angle-of-attack stability andcontrol characteristics and control law design of asupersonic persistence �ghter (SSPF) at 1g ightconditions. In general, the SSPF exhibited good y-ing characteristics at angles of attack between 20�
and 80�. A loss of roll control above an angle ofattack of 70� and degraded pitch damping abovean angle of attack of 75� were the only signi�cantproblem areas noted in the ight envelope. Flightsmade with reduced feedback gains indicated signi�-cant robustness in the control law design, which wasa primary goal of the control law synthesis process.Cross feeds of pilot inputs between the lateral anddirectional axes provided good roll coordination atall angles of attack tested. Use of thrust vectoringblended with conventional pitch and yaw control sur-faces provided good stability and control character-istics well into the poststall angle-of-attack region.De ectable wingtips (tiperons) coupled with conven-tional ailerons provided adequate roll control up toan angle of attack of 70�.
8
Free- ight tests indicated that conventional and
unconventional control surfaces could be blended
to provide good ying characteristics well into the
poststall angle-of-attack region for a con�guration
with highly nonlinear aerodynamic characteristics .
These results show that the direct eigenstructure
assignment technique for control law synthesis can
yield very good designs. However, caution must be
used in applying this technique as stability margins
are not guaranteed.
NASA Langley Research Center
Hampton, VA 23681-0001
October 19, 1992
9
References
1. Klein, John R.; Walck, Kenneth J.; and Hahne, David E.:
Airframe Component E�ects on the Aerodynamic Sta-bility and Control Characteristics of a Supersonic Cruise
Fighter Aircraftat HighAngles of Attack. AIAA-84-2110,
Aug. 1984.
2. Campbell, John P.: Free and Semi-Free Model Flight-
TestingTechniques Used in Low-Speed Studiesof Dynamic
Stability and Control . NASA TM X-50785, 1963.
3. Wood, Richard M.; Miller, David S.; Hahne, David E.;
Niedling, Larry G.; and Klein, John R.: Status Reviewof a Supersonically-Biased Fighter Wing-Design Study.
AIAA-83-1857, July 1983.
4. Hahne, David E.: Low-Speed Aerodynamic Stability and
Control Characteristics of an Advanced Fighter Aircraft
at High Angles of Attack . NASA TP-2617, 1986.
5. Marks, Bret A.; and Hahne, David E.: Innovative Con-
trol Concepts and Component Integration for a Generic
Supercruise Fighter. Aerodynamics of Combat Aircraft
Controls and of Ground E�ects, AGARD-CP-465, Apr.
1990, pp. 10-1{10-14.
6. Hahne, David E.: Low-Speed Static and Dynamic Force
Tests of a Generic Supersonic Cruise Fighter Con�gura-
tion. NASA TM-4138, 1989.
7. Chambers, Joseph R.; and Grafton, Sue B.: Static and
Dynamic Longitudinal Stability Derivatives of a Pow-
ered 1/9-ScaleModel of a Tilt-Wing V/STOL Transport .
NASA TN D-3591, 1966.
8. Schmidt, David K.; and Davidson, John B.: Flight Con-trol Law Synthesis for an Elastic Vehicle by Eigenspace
Assignment. AIAA-85-1898,Aug. 1985.
9. Wendel, Thomas R.: Flight Control Synthesis To Meet
Flying Qualities Speci�cations: An Evaluation of Multi-variable SynthesisTechniques. AIAA-87-2880,Sept. 1987.
10. Moore, B. C.: On the Flexibility O�ered by State
Feedback in Multivariable Systems Beyond Closed Loop
Eigenvalue Assignment. IEEE Trans. Autom. Control ,vol. AC-21, no. 5, Oct. 1976, pp. 689{692.
11. Cunningham, Thomas B.: Eigenspace Selection Proce-
dures for Closed Loop Response Shaping With Modal
Control. Proceedings of the 19th IEEE Conference on
Decision & Control Including the Symposium on Adap-
tive Processes,Volume 1, IEEE CatalogNo. 80CH1563-6,
IEEE Control Systems Soc., 1980, pp. 178{186.
12. Andry, A. N., Jr.; Shapiro, E. Y.; and Chung, J. C.:Eigenstructure Assignment for Linear Systems. IEEE
Trans. Aerosp.& Electron.Syst., vol. AES-19, no. 5, Sept.
1983, pp. 711{729.
13. Sobel, Kenneth M.; and Shapiro, Eliezer Y.: Applicationof Eigenstructure Assignment to Flight Control Design:
Some Extensions. J. Guid., Control, & Dyn., vol. 10,no. 1, Jan.{Feb. 1987, pp. 73{81.
14. Military Standard|Flying Qualities of Piloted Aircraft .MIL-STD-1797A, Jan. 30, 1990. (Supersedes MIL-STD-
1797(USAF), Mar. 31, 1987.)
15. Krekeler, Gregory C., Jr.; Wilson, David J.; and Riley,
David R.: High Angle of Attack Flying QualitiesCriteria.
AIAA-90-0219, Jan. 1990.
16. Wilson, David J.; and Riley, David R.: Flying Quali-
ties for Stall/Post-Stall Angles of Attack. High-Angle-
of-Attack Technology, Volume I, Joseph R. Chambers,
William P. Gilbert, and Luat T. Nguyen, eds., NASA
CP-3149, Part 2, 1992, pp. 569{585.
17. Wilson, David J.; and Riley, David R.: Flying QualitiesCriteria Development Through Manned Simulation for
45� Angle of Attack . NASA CR-4311, July 1990.
18. Military Speci�cation|Flight Control System|General
Speci�cation For . MIL-F-87242(USAF),U.S. Air Force,
Mar. 31, 1986.
19. Smereczniak, P.: Exhaust Nozzles for Aerocontrol. Vol-
ume I|Summary of Nozzle Static Test and Thrust Vec-
toring Utilization Results. WRDC-TR-89-3068, Vol. I,
U.S. Air Force, July 14, 1989. (Available from DTIC as
AD B137 805.)
20. Wendel, Thomas R.; Boland, Joseph R.; and Hahne,David E.: High Angle of Attack Control Law Develop-
ment for a Free-FlightWind Tunnel Model Using Direct
Eigenstructure Assignment. A Collection of TechnicalPapers, Volume1|AIAAGuidance,Navigationand Con-
trol Conference, Aug. 1991, pp. 255{266. (Available as
AIAA-91-2627-CP.)
21. Murri, Daniel G.; Nguyen, Luat T.; and Grafton, Sue B.:
Wind-Tunnel Free-Flight Investigation of a Model of a
Forward-Swept-WingFighter Con�guration.
22. Murri, Daniel G.; Grafton, Sue B.; and Ho�er, Keith D.:Wind-Tunnel Investigationand Free-Flight Evaluation of
a Model of the F-15 STOL and Maneuver Technology
Demonstrator . NASA TP-3003, August 1990.
23. Thomas, R.W.: Analysisof Aircraft Stabilityand Control
Design Methods, Volume I . AFWAL-TR-84-3038,Vol. I,U.S. Air Force, May 1984.
24. Moul, Martin T.; and Paulson, John W.: Dynamic
Lateral Behavior of High-PerformanceAircraft . NACA
RM L58E16, 1958.
18
Table I. Mass, Inertial, and Geometric Characteristics of Model
Overall fuselage length, ft . . . . . . . . . . . . . . . . . . . . . . 9.38Weight, lb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109.3
Ix, slugs-ft2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. 51
Iy, slugs-ft2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.5 7
Iz, slugs-ft2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20.26
Wing:Airfoil section . . . . . . . . . . . . . . . . . . . . . . NACA 64A004Span, ft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.14
Area, ft2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.56Mean aerodynamic chord, ft . . . . . . . . . . . . . . . . . . . . 3.22Aspect ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.95Leading-edge sweep, deg . . . . . . . . . . . . . . . . . . . . . . . 65
Aileron area (one side), ft2 . . . . . . . . . . . . . . . . . . . . . 0.30
Tiperon area (one side), ft2 . . . . . . . . . . . . . . . . . . . . 0.29
Vertical tails:Airfoil section (root) . . . . . . . . . . . . . . . . . . . NACA 65A005Airfoil section (tip) . . . . . . . . . . . . . . . . . . . . NACA 65A003
Area (each), ft2 . . . . . . . . . . . . . . . . . . . . . . . . . . 1.56Span, ft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.29Root chord, ft . . . . . . . . . . . . . . . . . . . . . . . . . . 1.96Tip chord, ft . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.45Aspect ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.07Leading-edge sweep, deg . . . . . . . . . . . . . . . . . . . . . . 62.8
Rudder area (each), ft2 . . . . . . . . . . . . . . . . . . . . . . 0.29
Trailing-edge extension:Length, ft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.6Width, ft . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.65
Flap area (each), ft2 . . . . . . . . . . . . . . . . . . . . . . . . 0.45
Canard:Airfoil section (root) . . . . . . . . . . . . . . . . Biconvex (t=c = 0:05)Airfoil section (tip) . . . . . . . . . . . . . . . . . Biconvex (t=c = 0:03)
Area, ft2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.36Span, ft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.84Tip chord, ft . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.30Aspect ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.48Leading-edge sweep, deg . . . . . . . . . . . . . . . . . . . . . . . 50
Thrust-vectoring vanes:Pitch-vane area (each), ft2 . . . . . . . . . . . . . . . . . . . . . 0.21Pitch-vane chord, ft . . . . . . . . . . . . . . . . . . . . . . . . 0.25
Yaw-vane area (each), ft2 . . . . . . . . . . . . . . . . . . . . . 0.05Yaw-vane chord, ft . . . . . . . . . . . . . . . . . . . . . . . . 0.25
19
Table II. De ection Range of Moving Surfaces
Surface De ection range, deg
LE aps 0 to 30
Ailerons �15
Tiperons �30
TEX aps �30
Rudders �30
Canard �40 to 20
Pitch vanes �25
Yaw vanes �25
20
-
-
Figure 1. Supersonic persistence �ghter free- ight model.
21
r
Z
q
Yp
X
u
w
Vv
α
β
Figure 2. De�nition of body-axis system.
22
65°°
.38c
5.14 ft
15°°
112.5869.00
38.3217.30
MS 0.0
(a) Three-view sketch of model.
0 10 20 30-.4
-.2
0
.2
BLWing twist distribution
Twist,deg
LE upLE down
Trailing-edgebreak
6.16
9.24
7.82
7.60
6.72
8.31
MS 112.60MS 69.00MS 31.08
Theoreticalside of fuselage
4.82
35°
65°
6.63
4.69LE flap
Tiperon
Aileron
TEX flap
TEX
(b) Wing and TEX geometry; NACA 64A004 airfoil.
Figure 3. Details of baseline model geometry. Linear dimensions in inches except as noted.
23
7.73
MS 38.36
Canardspindle
13.98
3.45
MS 89.20
14.14
22.08
3.60
62.8°
23.52
15.48
5.40
2.1
5.8
50°
(c) Canard and vertical tail geometry.
Figure 3. Concluded.
24
-
-
(a) Ejector unit used to generate thrust.
Figure 4. Ejector and thrust-vectoring vanes.
25
-
-
(b) Thrust-vectoring vanes.
Figure 4. Concluded.
26
-
-
(a) Side view of model in ight.
Figure 5. Free- ight test technique.
27
-
-
(b) Three-quarter rear view of model in ight.
Figure 5. Continued.
28
Tunnel ground boardRoll
and yawpilot
Safetycableoperator
Tunneloperator
Pitchpilot
Thrustoperator
Flight controlcomputer
Safetycable
Power andcontrol cable
(c) Free- ight test setup.
Figure 5. Concluded.
29
���
����
��
Input Output
Aerodynamicdata
Enginedata
Deflection andrate limits
Aircraft response
Augmentation capability
ωsp,rad/sec
0 20 40 60α, deg
ProgramVECTOR
(a) Inputs and outputs of program VECTOR.
SCHEDULE setup
CreateModifySaveRetrieve
Computations
AerodynamicsTrimStability
Display results
StabilityControlLiftDragSchedules
Mainmenu
Savedata
Mass andinertiadata
Aerodata
(b) Flow chart of program SCHEDULE.
Figure 6. Computer programs used in development of SSPF free- ight control laws.
30
2.0
1.6
1.2
.8
.4
0
-.4
CL
.16 .08 0 -.08 -.16Cm
Basic aerodynamic characteristics
2.0
1.6
1.2
.8
.4
0
-.4
CL
.16 .08 0 -.08 -.16Cm
Augmented aerodynamic characteristics
10
5
0
-5
-10
-15
-200 15 30 45 60 75
α, deg
Canard schedule
cδ
Figure 7. E�ect of canard schedule on longitudinal stability.
31
Pitchstick
Pitchtrim
++ Pk1
α,deg filter1 Sw4p 1
.05s + 1Canardschedule
16.0α, ref Pk2
q ,deg/sec filter1 Pk3
Pk4
A-
+ ++
δ ,cdeg
-+ Sw1p
Sw2p
Sw3p
p
r
--
(a) Longitudinal feedback.
Figure 8. Block diagrams for free- ight model ight control system.
32
Pk5
A
Pk6
1 ..05s + 1
Pk7
Sw5p δF,deg
Sw6p δpv,deg
++
30
-30
25
-25
-30 Pk8
1 ..05s + 1-30 Pk8
++
(b) Longitudinal control blending.
Figure 8. Continued.
33
Rolltrim
Rollstick Rk1
Rk2
C
Sw4rδ ,a/wtdeg
D
Pstab,deg/sec Rk3
Rstab,deg/sec Rk4
BetDot,deg/sec Rk5
BetaF,deg Rk6
+
+ Rk34 Sw7r
Sw9r
Sw8r
15
-15
++
Rk56
--
--
-+
-
(c) Lateral feedback and control.
Figure 8. Continued.
34
Yawtrim
Yawstick Yk1
Yk2
Sw5rδ ,rdeg
Pstab,deg/sec Yk3
Rstab,deg/sec Yk4
BetDot,deg/sec Yk5
BetaF,deg Yk6
+
+ Yk34
Sw6rδ ,yvdeg
C
Yk7
Yk8
Sw1r
Sw3r
Sw2r
30
-30
25
-25D
+
+
--
++
-
-
-
(d) Directional feedback and control.
Figure 8. Continued.
35
p,deg/sec Filter1
r,deg/sec Filter 1
BetDot,deg/sec
Pstab,deg/sec
β,deg Filter 1 BetaF,
deg
cos α
sin α
+
+
p
r
Rstab,deg/sec
cos α
sin α
-+
-
+ 50
-50
sin α
cos α
p
r p
r
Filter 1
2500 .s**2 + 70.7s + 2500
r
Yk10
0.1s.05s+1
Filter 1Ny,
g units
++
+-
(e) Lateral-directional feedback signals.
Figure 8. Concluded.
36
20
16
12
8
4
0
0
-.2
-.4
-.6
-.8
-1.0
.010
.008
.006
.004
.002
0
0
-.4
-.8
-1.2
-1.6
10 20 30 40 50 60 70α, deg
Pk 1,deg/unit stick
Pk 2,deg/unit stick
Pk 3,deg/deg/sec
Pk 4,deg/
(deg/sec)2
Baseline gains
Modified gains
Figure 9. Free- ight feedback gain schedules.
37
1.2
.8
.4
0
-.4
1.2
.8
.4
0
-.4
1.2
.8
.4
0
-.410 20 30 40 50 60 70
α, deg
Pk 5,deg/deg
Pk 6,deg/unit stick
Pk 7,deg/deg
Pk 8,deg
0
-.2
-.4
-.6
-.8
-1.0
Baseline gains
Modified gains
Figure 9. Continued.
38
0
-4
-8
-12
-16
0
5
10
15
20
.25
.20
.15
.10
.05
010 20 30 40 50 60 70
α, deg
.20
.15
.10
.05
0
Baseline gains
Modified gainsPk 1,
deg/unit stick
Pk 2,deg/unit stick
Pk 3,deg/deg/sec
Pk 4,deg/deg/sec
Figure 9. Continued.
39
1.2
.8
.4
0
-.4
20
18
16
14
12
1010 20 30 40 50 60 70
α, deg
Rk 5,deg/deg/sec
Rk 6,deg/unit stick
Yk 1,deg/unit stick
Yk 2,deg/unit stick
Baseline gains
Modified gains
0
-10
-20
-30
-40
-50
.5
0
-.5
-1.0
-1.5
-2.0
Figure 9. Continued.
40
0
-.1
-.2
-.3
.8
.6
.4
.2
0
-.210 20 30 40 50 60 70
α, deg
Yk 3,deg/deg/sec
Yk 4,deg/deg/sec
Yk 5,deg/deg/sec
Yk 6,deg/deg
0
-.1
-.2
-.3
-.4
-.5
.3
.2
.1
0
-.1
Baseline gains
Modified gains
Figure 9. Continued.
41
10 20 30 40 50 60 70α, deg
Yk 7,deg/deg
Yk 8,deg/deg
Yk 10,deg/deg
50
40
30
20
1.2
.8
.4
0
1.2
.8
.4
0
-.4
Baseline gains
Modified gains
Figure 9. Concluded.
42
6
4
2
0 20 40 60α, deg
ωd,
rad/sec
6
4
2
0
1.5
1.0
.5
0 20 40 60α, deg
ζsp
ζd
Desired
Achieved
6
4
2
0
ωsp,
rad/sec
1.5
1.0
.5
τR,
sec
Figure 10. Linear evaluation of control system.
43
-
-
(a) Pilot at work station.
-
-
(b) Close-up of work station.
Figure 11. Real-time-engineering simulation work station.
44
-
-
Figure 12. E�ect of canard de ection on longitudinal characteristics. �f = 30�; �F = 0�.
45
-
-
Figure 13. E�ect of TEX ap de ection on longitudinal characteristics. �f = 30�; �c = �10�.
46
-
-
Figure 14. E�ect of canard de ection on lateral-directional characteristics. �f = 30�; �F = 0�.
47
-
-
(a) Combined aileron and tiperon control e�ectiveness. �f = 30�; �c = �10�; �F = 0�.
Figure 15. Lateral-directional control e�ectiveness.
48
-
-
(b) Rudder control e�ectiveness. �f = 30�; �c = �10�; �F = 0�.
Figure 15. Concluded.
49
4
2
0
-2
1.2
.8
.4
0
-.4
-.8
.4
0
-.4
-.8
-1.2
-1.60 10 20 30 40 50 60 70 80 90
α, deg
-10-20-40
δc , deg
CYp + CY β
sin α•
Cnp + Cn β
sin α•
Clp + Cl β
sin α•
Figure 16. E�ect of canard de ection on roll damping. �f = 30�; �F = 0�.
50
1.6
1.2
.8
.4
0
-.40 10 20 30 40 50 60 70 80 90
2
0
-2
-4
.8
.4
0
-.4
-.8
-1.2
-40-20-1020
α, deg
δc , deg
CYr – CY β
cos α•
Cnr – Cn β
cos α•
Clr – Cl β
cos α•
Figure 17. E�ect of canard de ection on yaw damping. �f = 30�; �F = 0�.
51
-3
-2
-1
0
-10 0 10 20 30 40 50 60 70α, deg
Cmq
Stable
Unstable
Figure 18. Estimated pitch-damping characteristics.
52
-100
-50
0
50
100
-40 -20 0 20 40
Thrust, lb
δyv , deg
38 89
Yawing moment,ft-lb
(a) Yaw vectoring.
-150
-100
-50
0
50
100
150
-40 -20 0 20 40
Thrust, lb
38 89
Pitching moment,ft-lb
δyv , deg
(b) Pitch vectoring.
Figure 19. Wind-o� thrust-vectoring e�ectiveness.
53
-.6
-.4
-.2
0
.2
.4
.6
∆Cn
10 20 30 40 50 60α, deg
70
(a) Yaw vectoring.
-1.5
-1.0
-.5
0
.5
1.0
1.5
∆Cm
10 20 30 40 50 60 70α, deg
(b) Pitch vectoring.
Figure 20. Thrust-vectoring e�ectiveness for estimated trimmed ight conditions.
54
-40
-30
-20
-10
0
10
10 20 30 40 50 60 70α, deg
δc
δF
δpv
Trim deflection,deg
(a) Trim control settings.
0
2
4
6
8
30
40
50
60
70
80
90
100
10 20 30 40 50 60 70α, deg
Dynamic pressure Thrust
Thrust,lb
Dynamicpressure,
lb/ft2
(b) Trim dynamic pressure and thrust.
Figure 21. Predicted trim controls, dynamic pressure, and thrust of free- ight model.
55
-.02
-.01
0
.01
10 20 30 40 50 60 70
Cmα
Unstable
Stable
α, deg
Off Baseline Modified
FCL
(a) Static pitch stability.
-40
-30
-20
-10
0
10 20 30 40 50 60
Unstable
Stable
70
Cmq
α, deg
Off Baseline Modified
FCL
(b) Dynamic pitch stability.
Figure 22. Longitudinal stability characteristics of free- ight model for trimmed ight.
56
-.006
-.004
-.002
0
.002
.004
10 20 30 40 50
Stable
Unstable
60 70
Cn β
α, deg
Off Baseline Modified
FCL
(a) Directional stability.
-.006
-.004
-.002
0
.002
10 20 30 40 50 60
Unstable
Stable
70
Cl β
α, deg
Off Baseline Modified
FCL
(b) Lateral stability.
Figure 23. Static lateral-directional stability characteristics of free- ight model for trimmed ight.
57
-.01
0
.01
.02
.03
.04
10 20 30 40 50 60 70
Stable
Unstable
Cn β,dyn
α, deg
Off Baseline Modified
FCL
Figure 24. Yaw divergence resistance of the free- ight model for trimmed ight.
58
-5
-4
-3
-2
-1
0
1
10 20 30 40
Unstable
Stable
50 60 70
Cnr
α, deg
Off Baseline Modified
FCL
(a) Yaw-damping characteristics.
-1.2
-0.8
-0.4
0
10 20 30 40 50 60 70
Unstable
Stable
Cl p
α, deg
Off Baseline Modified
FCL
(b) Roll-damping characteristics.
Figure 25. Lateral-directional damping characteristics of free- ight model for trimmed ight.
59
10 11 12 13 14 15 16 17 18
Appendix A
Baseline Control Law Gain Schedules
Appendix B
Modi�ed Control Law Gain Schedules
1
REPORT DOCUMENTATION PAGEForm Approved
OMB No. 0704-0188
Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources,gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of thiscollection of information, including suggestions for reducing this burden, toWashington Headquarters Services, Directorate for Information Operations and Reports, 1215 Je�ersonDavis Highway, Suite 1204, Arlington, VA 22202-4302, and to the O�ce of Management and Budget, Paperwork Reduction Project (0704-0188), Washington, DC 20503.
1. AGENCY USE ONLY(Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED
February 1993 Technical Paper
4. TITLE AND SUBTITLE
Wind-Tunnel Free-Flight Investigation of a Supersonic PersistenceFighter
6. AUTHOR(S)
David E. Hahne, Thomas R. Wendel, and Joseph R. Boland
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)
NASA Langley Research CenterHampton, VA 23681-0001
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)
National Aeronautics and Space AdministrationWashington, DC 20546-0001
5. FUNDING NUMBERS
WU 505-59-30-07
8. PERFORMING ORGANIZATION
REPORT NUMBER
L-17040
10. SPONSORING/MONITORING
AGENCY REPORT NUMBER
NASA TP-3258
11. SUPPLEMENTARY NOTES
Hahne: Langley Research Center, Hampton, VA; Wendel and Boland: McDonnell Aircraft Company,St. Louis, MO.
12a. DISTRIBUTION/AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE
Unclassi�ed{Unlimited
Subject Categories 02 and 08
13. ABSTRACT (Maximum 200 words)
Wind-tunnel free- ight tests have been conducted in the Langley 30- by 60-Foot Tunnel to examine the high-angle-of-attack stability and control characteristics and control law design of a supersonic persistence �ghter(SSPF) at 1g ight conditions. In addition to conventional control surfaces, the SSPF incorporated de ectablewingtips (tiperons) and pitch and yaw thrust vectoring. A direct eigenstructure assignment technique wasused to design control laws to provide good ying characteristics well into the poststall angle-of-attack region.Free- ight tests indicated that it was possible to blend e�ectively conventional and unconventional controlsurfaces to achieve good ying characteristics well into the poststall angle-of-attack region.
14. SUBJECT TERMS 15. NUMBER OF PAGES
Control law design; Advanced �ghter; Free ight; Dynamic stability 60
16. PRICE CODE
A0417. SECURITY CLASSIFICATION 18. SECURITY CLASSIFICATION 19. SECURITY CLASSIFICATION 20. LIMITATION
OF REPORT OF THIS PAGE OF ABSTRACT OF ABSTRACT
Unclassi�ed Unclassi�ed
NSN 7540-01-280-5500 Standard Form 298(Rev. 2-89)Prescribed by ANSI Std. Z39-18298-102
NASA-Langley, 1993
