High-Alpha Handling Qualities Flight Research on the NASA F/A-18 High Alpha Research Vehicle

,, .! , -=-, ( .

NASA Technical Memorandum 4773

High-Alpha Handling

Qualities Flight Research

on the NASA F/A- 18 High

Alpha Research Vehicle

Keith D. Wichman, Joseph W. Pahle, Catherine Bahm, John B. Davidson, Barton J. Bacon, Patrick C. Murphy, Aaron J. Ostroff, and Keith D. Hoffier November 1996 NASA Technical Memorandum 4773

High-Alpha Handling

Qualities Flight Research on

the NASA F/A- 18 High Alpha

Research Vehicle

Keith D. Wichman, Joseph W. Pahle, and Catherine Bahm NASA Dryden Flight Research Center Edwards, California John B. Davidson, Barton J. Bacon, Patrick C. Murphy, and Aaron J. Ostroff NASA Langley Research Center Hampton, Virginia Keith D. Hoffler ViGYAN, Inc.

Hampton, Virginia National Aeronautics and Space Administration Office of Management Scientific and Technical Information Program HIGH-ALPHA HANDLING QUALITIES FLIGHT RESEARCH ON THE NASA F/A-18 HIGH ALPHA RESEARCH VEHICLE Keith D. Wichman Joseph W. Pahle Catherine Bahm NASA Dryden Flight Research Center Edwards, CA John B. Davidson Barton J. Bacon Patrick C. Murphy Aaron J. Ostroff NASA Langley Research Center Hampton, VA Keith D. Hoffler ViGYAN, Inc.

Hampton, VA ABSTRACT A flight research study of high-angle-of-attack handling qualities has been conducted at the NASA Dryden Flight Research Center using the F/A-18 High Alpha Research Vehicle (HARV).

The objectives were to create a high-angle-of-attack handling qualities flight database, develop appropriate research evaluation maneuvers, and evaluate high-angle-of-attack handling qualities guidelines and criteria. Using linear and nonlinear simulations and flight research data, the predictions from each criterion were compared with the pilot ratings and comments. Proposed high-angle-of-attack nonlinear design guidelines and proposed handling qualities criteria and guidelines developed using piloted simulation were considered. Recently formulated time-domain Neal-Smith guidelines were also considered for application to high-angle-of-attack maneuvering.

Conventional envelope criteria were evaluated for possible extension to the high-angle-of-attack regime. Additionally, the maneuvers were studied as potential evaluation techniques, including a limited validation of the proposed standard evaluation maneuver set. This paper gives an overview of these research objectives through examples and summarizes result highlights. The maneuver development is described briefly, the criteria evaluation is emphasized with example results given, and a brief discussion of the database form and content is presented.

NOMENCLATURE

Acronyms

ACM air combat maneuvering actuated nose strakes for enhanced rolling ANSER BFM basic fighter maneuvers CAP control anticipation parameter CHR Cooper-Harper rating HARV High Alpha Research Vehicle HUD head-up display KEAS knots equivalent airspeed LOES lower-order equivalent system MATV Multi-Axis Thrust Vectoring MDA McDonnell Douglas Aerospace PIO pilot-induced oscillation RFCS research flight control system RMS root mean square standard evaluation maneuvers set STEMS TVCS thrust-vectoring control system Symbols D time-domain Neal-Smith acquisition time, sec Neal-Smith compensator gain, dB n/o_ ratio of load factor change for each angle-of-attack change, g/rad Neal-Smith compensator lag time constant, sec

rt g

Neal-Smith compensator lead time constant, sec Tt_ad time constant of the lag between flight path and pitch attitude, sec

to2

angle of attack, deg short-period damping ratio Neal-Smith compensator delay time constant, sec _pilot tO c Smith-Geddes critical frequency, rad/sec short-period frequency, rad/sec O)sp INTRODUCTION Interest in flying qualities in the high-angle-of-attack flight regime has traditionally been lim- ited to departure-resistance characteristics. With the advent of advanced control effectors such as multiaxis thrust vectoring and closed-loop forebody vortex control, this flight regime can now be exploited for enhanced fighter maneuverability and, therefore, tactical advantage. New definitions of desirable handling qualities at high angles of attack must complement this new maneuvering potential in order to effectively use these advanced controllers. Aircraft designers and testing agents must have flight-validated guidelines, evaluation maneuvers, and flight test techniques to define handling qualities at high angles of attack and to facilitate the meaningful evaluation of these aircraft.

Advanced flight research within the high-angle-of-attack flight regime has been conducted at the NASA Dryden Flight Research Center (NASA Dryden) as part of the High-Alpha Technology Program. For the purposes of this research, "high angles of attack" are angles ranging from 30 ° to 70 °. This research was conducted using the NASA F/A-18 High Alpha Research Vehicle (HARV). One facet of this project was to improve understanding of high-angle-of-attack handling qualities. The research objectives were to create a high-angle-of-attack handling qualities flight database, develop appropriate research evaluation maneuvers, and evaluate high-angle-of-attack handling qualities guidelines and criteria.

A few aircraft have been developed and tested using such advanced control effectors. These aircraft were oriented primarily towards operational utility of the control effectors rather than gen- eral research of the flight regime. The X-31A flight program featured multiaxis thrust vectoring and primarily investigated enhanced fighter maneuverability and the tactical utility of vectoring.

A limited study of high-angle-of-attack handling qualities was conducted at the end of the pro- gram using maneuvers similar to those used in the HARV research programJ The resulting data will provide an important overlap with HARV research to validate results for more than one air- frame. A significant limitation in the scope of the X-31A handling qualities was the lack of control system variations.

The F-16 Multi-Axis Thrust Vectoring (MATV) aircraft high-angles-of-attack handling quali- ties were evaluated. 2 Some evaluation maneuvers developed by the HARV program were used. Be- cause of the MATV program goal of rapid evaluation and demonstration of an operational-type thrust-vectoring system, the primary emphasis "was to evaluate the dynamics of the aircraft config- uration rather than to conduct specific research on high-angle-of-attack handling qualities. As with the X-31A program, few variations in aircraft dynamics were evaluated.

This paper gives an overview of these HARV handling qualities research objectives through examples and summarizes result highlights. The maneuver development is described briefly, the criteria evaluationis coveredat lengthwith exampleresultsgiven, anda brief discussionof the database form andcontentis presented.

DESCRIPTIONOF FACILITIES

The facilities requiredto conductthe handlingqualitiesresearch includedthe HARV aircraft, nonlinearsix-degrees-of-freedom simulations(piloted and batch), andvalidatedlinear models.

Thesefacilities aredescribed in the following sections.

High Alpha Research VehicleAircraft

The HARV aircraft,equippedwith a reconfigurable research flight control system(RFCS), wasanexcellentfacility forhigh-angle-of-attack investigation (fig. 1).Theaircraftfeaturesinclud- edamultiaxis thrust-vectoring controlsystem (TVCS)and,for thefinal phase of theprogram,fore-

body vortex control using the actuatednose strakesfor enhancedrolling (ANSER) system

integratedinto the stablehigh-angle-of-attack F/A-18airframe.TheRFCSprovidedthe capability

to examinemultiple controllaw designsandtheir variantswithoutcompromising the safetyof the

pilot or aircraft.This research computer wasdesigned to providea ClassB (missioncritical but not safetycritical) controlenvironment that,by definition,ensures safereversionbackto the standard F/A- 18configurationif a failure weredetected or if thepilot believeda dangerexisted.Theinher- ent spin resistance of the F/A-18 airframeandthe spinrecoverychuteaddedto the suitability of

EC9643479-5

Figure 1.HARV aircraft.

the testbedfor safehigh-angle-of-attack flight research. Additionally, themanyaerodynamic con-

trol surfacesandthe addedthrustvectoringandforebodystrakesprovidedsignificant variety in

controlpower usage.A morecomprehensive descriptionandextensivereferences of the HARV

research andfacility is givenby Bowers. 3

An undesirable resultof addingall of this hardware to the aircraftexists.Thedramaticchange

in weight and inertias introducedby the TVCS and related systemsresulted in a significant

changein the handlingqualitiesof the HARV underthe controlof the unmodifiedF/A-18 flight

controlsystem.The addedweight wasprimarily at the extremeaft andextremeforward aircraft

positions,essentiallyproducinga flying "dumbbell."In-flight refuelingusingthebasicflight con- trol systemwasanextremelydifficult taskwith very stronglongitudinalpilot-inducedoscillation

(PIO) tendencies. Figure2 showstime-historydataof onesuchrefuelingattempt.A potentialef-

fect of this tendencywasto causethe research controllaw designers to beginwith a handicapin

pitch dynamics. Eachlongitudinalresearch controllaw designexhibitedatleastsometendencyto

be sensitiveor oscillatoryin the pitch axis, with the commonresultof lessthan desirablepitch- trackinghandlingqualities.

SimulationFacilities

A broad range of ground-testand simulationfacilities were usedconcurrentlyduring the

HARV program.The primarypiloted simulationusedin the development of performance guide-

lines andhandlingqualitiesevaluations wasthe fixed-base, 40-ft domedifferential maneuvering

simulatorattheNASA LangleyResearch Center.Thedifferentialmaneuvering simulatorwasalso

2 HARV PIO during refueling attempt, Mach 0.5, 18k ft Pitch stick, In -2 Pitch rate, 0 _ , deg/lmc -10 i 1.5 Normal acceleration, g .5 Stabllator, deg -10 0 2 4 6 8 Time, sec 960541 Figure 2. Time history of in-flight refueling PIO.

used in the control law design process and flight maneuver development. 4 The piloted simulation at NASA Dryden was limited to forward visuals only, but could be linked with an all-software, hardware-in-the-loop, or iron-bird capability. The NASA Dryden simulation was used for flight planning, engineering, and software development, and was the primary site for software and hard- ware verification and validation testing. A configuration-controlled batch simulation, common to both sites, was used as a benchmark from which to compare other dissimilar simulations.

Linear Models Validating that the linear models used extensively for evaluation of the handling qualities criteria represented adequate reflections of the observed HARV flight dynamics was important.

The models compared quite favorably with the nonlinear simulation in both the frequency and time domains (figs. 3 and 4). This nonlinear simulation has had a long history of accurately modeling the F/A-18 and HARV flight dynamics. These comparisons were made in both the time and frequency domain for relatively small input magnitudes. As input size increases, the number of system nonlinearities increases and the confidence in the models decreases. For most of the cases researched, this small input assumption is appropriate (fine and not-so-fine tracking). For gross- acquisition analysis, these linear assumptions, and therefore any linear analysis technique results, become suspect. Throughout the HARV program, the nonlinear simulation and the linear models proved representative of the actual aircraft system dynamics, given the assumptions just discussed.

Through this and additional experience with these models, great confidence in their validity has been developed.

Pitch stlck-to-pltch rate frequency response for nonlinear simulation and linear model for 30 ° (_

L

Magnitude, dB -50 Phase, deg i i i i J i i i i I -200 Coherence .5 0 a i i i i i i I i i i i i i i i I 10-1 10 0 101 Frequency, rad/nc 960542 Figure 3. Frequency-domain linear model validation example.

Nonlinear simulation and linear model for a 0.5 in. doublet at 30 ° O_ deg 30 Nonlinear

° J

Linear , i _ _ _ i Pitch rate, 0 deg i i i i i i i -5 Stabilator, .... i....... i ....... i_, • ...... i....... i ..... i ......

deg 0 -10 Pitch vectoring, 0 deg i i i i i i -10 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Time, aec 9sos43 Figure 4. Time-domain linear model validation example.

DESCRIPTION OF EXPERIMENT The research experiment conducted using the HARV aircraft resulted in a comprehensive yet limited handling qualities flight database for the high-angle-of-attack flight regime where none previously existed, and important insights can be derived from it. The study evaluated longitudinal and lateral-directional gross-acquisition and tracking handling qualities from 30 ° to 60 ° angle of attack (_), parametric gross acquisition, open-loop performance effects, basic fighter maneuver- ing, and air combat maneuvering (ACM). Numerous control law configurations and advanced con- trol effectors, including the forebody strakes and thrust vectoring, were used to vary available control power and the closed-loop dynamics. The handling qualities of these configuration varia- tions ranged from good tracking and gross-acquisition characteristics to very poor handling quali- ties, including divergent PIOs. The database was created primarily using two project pilots and several guest pilots: An extensive simulation database using fixed-base dome and some motion- based simulators was created for comparison with flight results. Head-up display (HUD) video and pilot comments were cataloged with the flight time-history data.

In spite of its significance, the HARV handling qualities flight database was limited and prob- lematic in many respects. These limitations resulted from the fact that the HARV project was not dedicated exclusively to handling qualities research but to high-or flight in general, embracing mul- tiple disciplines. This necessity to share the facility and the large amount of valuable flight time required for the handling qualities maneuvers dictated a "first look" approach to this research. This approach led to a broad but shallow database that was compiled over several years, often with large time gaps separating research points. These time gaps disrupted the continuity and consistency of boththemaneuverquality andpilot evaluation. Therefore, theresultingdatabase cannotbeconsid- eredstatisticallysignificant.

A descriptionof the variousresearch controllawsdeveloped andtestedby the HARV project

teamis givenby Pahle. 5Exceptwherenoted,thispaperwill only beconcerned with thedistinctions amongthecontrollawsin orderto present thebreadth of thedatabase andto illustrateexamples of the performance of the variouscandidate handlingqualitiescriteriaandguidelines.

The threemajor sectionsof this paperthatfollow address the handlingqualitiesresearch ob-

jectives of the HARV project.Theseobjectiveswereto createa high-_ handlingqualitiesflight

database, developappropriate research evaluationmaneuvers, andevaluatehigh-o_ handlingqual- ities guidelinesandcriteria.Thegreatest emphasis will begivento theguidelinesandcriteriaeval-

uation; however, brief overviewsof the databaseand maneuverdevelopmentobjectives are

included.The testconductwasseparated into the following distinctefforts: • Flight dataacquisition:Conducthandlingqualitiesflight maneuvers andacquireflight data,

HUD video, pilot ratings,andcommentsfor the control law variationsmadepossibleby

other research objectives. The resultingdatabase wasthetruth model.

• Handling qualitiescriteria predictions:Perform analysisof simulationdatausing various

handling qualifiescriteria to generatethe predictionsfor the handling qualities of each

control law configurationflown.

• Evaluation and modification: Comparethe criteria predictions with the flight results

evaluating the applicability and accuracyof eachcriteria and attempt to modify them

if necessary.

HIGH-ANGLE-OF-ATTACK EVALUATION MANEUVERS

In orderto evaluatehigh-orhandlingqualities,developingand flight-validating appropriate

maneuvers wasnecessary. Because of the unconventional maneuvering possiblein this flight re-

gimewith the advanced controllersmentioned, conventionalevaluationmaneuvers for low angles

of attackneededto be adaptedandnewmaneuvers neededto be developed. Ground-based simu-

lations wereusedextensivelyto developandrefine thesemaneuvers. 6,7 Somein-flight trial and

error wasrequiredto furtherrefinethe maneuvers. Mostnoteworthyof thesemaneuvers werethe

high-txtracking andacquisitionmaneuvers (fig. 5). Thesemaneuvers werevery effective at iso- lating the longitudinal from the lateral-directionalaxesandgross-acquisition from fine-tracking performance. These maneuvers were conducted at discrete angles of attack of 30 ° , 45 ° , and 60 ° .

The maneuvers were highly repeatable and gave good discrimination of both good and poor handling qualities.

Many other maneuvers were developed or refined as discussed in the next section of this paper.

Open-loop performance, parametric closed-loop (such as g, _, pitch attitude-angle, and bank- angle captures), targeted closed-loop (tracking and acquisition), basic fighter maneuvers (BFM), and ACM make up the general classes of maneuvers developed and refined by the HARV team for high-or evaluation. All of these maneuvers aided in the evaluation of various guidelines and criteria as discussed in the next section.

'_m:_ 20 ° Oc;maintain 200 knots 25,000 ft Pull altitude Features. Developed and refined" with ground-based simulation _ _ • Repeatability Roll into plane-'_ • Isolate tracking and aquisltion • Isolate longitudinal and latsral -directional dynamics of target_ _L o HARV objectives • Flight validation of the maneuvers as handling qualities evaluation tools 3000 ft • Use for comments and ratings to evaluate handling qualities criteria and control law configurations Successfully evaluated handling qualities end identified o HARV _ _ problem areas 30 acC(qlu(_:glt:ndi::/mgrp _:s" <_ 25, "a000ch :;45ude 960544 Figure 5. High-angle-of-attack tracking and acquisition maneuvers.

Maneuver Pyramid The handling qualities maneuvers were examined as a group as a potential evaluation tech- nique. These maneuvers were organized into a "maneuver pyramid" (fig. 6) where maneuvers use- ful to engineering design could be progressively linked with maneuvers having operational significance to a combat pilot. Maneuvers towards the bottom of the pyramid were relatively large in number, repeatable, easy to fly, well-suited for simulation evaluation, and gave results that the controls designer could easily translate to design parameters. Maneuvers near the top of the pyra- mid were comparatively small in number, costly to fly, less repeatable, less suited for simulation evaluation, and had less direct translation to design parameters than the other maneuvers. These maneuvers were, however, significantly more representative of the operational tasks that the con- trol law and aircraft system were designed to enhance than the maneuvers towards the bottom of the pyramid.

This maneuver pyramid can be described by the following progression. Open-loop perfor- mance maneuvers were followed by parametric closed-loop maneuvers such as cx or pitch-attitude captures. These parametric closed-loop maneuvers were followed by target acquisition and track- ing maneuvers. Finally, a small number of BFM such as break turns and "J turns" against targets were employed, followed by an even smaller number of ACM engagements. These engagements were one-against-one, unstructured, simulated combat engagements, unlike the other maneuvers, which were predefined and highly structured. The purpose of the ACM (and to some extent the BFM) was to uncover any characteristics or latent deficiencies that might only be detected by such free-form maneuvering. An attempt was made to identify a link throughout this maneuver progres- sion of "good" or "bad" performance and handling qualities and to relate the specific deficiency or 360* tulle (These maneuvers are Integrated into envelope expansion to optimize flight Ume) 960545 Figure 6. Maneuver pyramid for high-angle-of-attack handling qualities evaluation.

enhancement to specific elements of control system design. The effectiveness of this pyramid for uncovering control law deficiencies early in the development process and linking these with parameters the designer could readily use was evaluated during the program.

Using this pyramid, figure 7 shows pilot ratings from flight arranged as a function of maneuver and flight condition for the Version 27 control law. This configuration demonstrated pitch bob- bling and sensitivity problems during the target tracking and acquisition maneuvers. The implica- tion can be drawn from this figure and the associated pilot comments that a necessary but not sufficient relation of the low-level engineering maneuvers with high-level operational maneuvers exists. Poor performance or handling qualities indications during the engineering-type maneuvers typically carried through to deficiencies in the operationally significant maneuvers. However, poor performance or handling qualities uncovered during the operational-type maneuvers were not necessarily indicated clearly by the results of the low-level maneuvers.

Standard Evaluation Maneuvers Set The standard evaluation maneuvers set (STEMS) of 20 maneuvers was developed by McDon- nell Douglas Aerospace (MDA) (St. Louis, Missouri) under contract to the U.S. Air Force. 7 These maneuvers have been candidates for inclusion in the revision of MIL-STD-1797 as handling qual- ities demonstration maneuvers. The STEMS was intended to provide a "mission-oriented" evalu- ation of aircraft handling qualities as opposed to a parametric criteria-based specification. The HARV facility was used in support of the MIL-STD-1797 revision to validate 17 of the maneuvers in the set (table 1) in terms of flyability and effectiveness at evaluating flying qualities, particularly at high angles of attack. The tests were conducted using two HARV project pilots and two flight Pushover from 60 ° C_, capture 10 ° (_ captures from (X captures g captures with Level 3 mach 0.6, 25k ft from 20 ° (_, 25k ft 90 ° heading change Level 2 Level 1 38 ° 45 ° 60 ° 15 ° 30 ° 38 ° 450 60 ° 1.5g 2.0g 3.0g4.0g a Tracking maneuvers and BFM Level 3 Level 2 Level 1 30 ° (x 30 ° a gross Break Split-S Fiat J turn tracking acquisition turn scissors 960546 Figure 7. Longitudinal closed-loop pilot ratings for Version 27 control law.

control systems, including various thrust-vectoring and forebody strakes combinations. The pilots evaluated each maneuver based on definition, target setup, difficulty, repeatability, and similarity to an operational task and were asked to offer improvement recommendations. Results favorably indicated that such a maneuver set could be effectively used to demonstrate handling qualities, uncover operational deficiencies, and represent some elements of operational tasks, s Table 1. STEMS maneuver sequence flown on the HARV aircraft.

STEM 6 Maximum pitch pull STEM 16 1-g stabilized pushover STEM 14 Minimum speed for 80 ° pitch STEM 17 J turn STEM 15 Minimum time 180 ° heading change STEM 9 Pitch rate reserve STEM 13 High-or roll and capture STEM 12 High-o_ roll reversal STEM 5 Rolling defense STEM 7 Noseup pitch angle capture (medium) Table1.Concluded.

STEM 7

Noseuppitch anglecapture(low)

STEM 10

High-orlongitudinalgrossacquisition

STEM 3

High-o_ lateralgrossacquisition

STEM 1

Trackingduring high-orsweep

STEM 2

High-txlongitudinalandlateraltracking

STEM 4 Dual attack

STEM 8

Crossingtargetacquisitionandtracking

STEM 11 Sharkenhausen

Stepping-Target SimulationTask

A longitudinal PIO was encountered during fine tracking during the flight test of the NASA- 1A control law. The PIO was not accurately predicted by ground-based simulation prior to the 1994 flight test of the control law. After the 1994 flight test evaluation, a piloted simulation test tech- nique was developed to reproduce the PIO tendencies seen in flight. The technique was then used to test iterations of the control law redesign, ANSER, in an attempt to eliminate the PIO. A second flight test phase was made with the ANSER control law in 1995 and 1996. The PIO tendencies were nearly eliminated in flight with the modified control laws. Hoffler discusses the original and modified tracking task, or "stepping-target" task, simulation techniques in detail. 9 The goal of the stepping-target tracking task technique was to increase pilot gain during simu- lation evaluations. The original tracking tasks were changed in two ways: pilot requirements were changed from tracking criteria and Cooper-Harper ratings to the use of a pseudogun model to "kill" the target, and the target trajectory was modified from a steady turn to a steady turn with discrete steps superimposed on the flight path (the "stepping target"). The prerecorded targets used for ear- lier evaluations were utilized with the stepping target superimposed on their trajectory. The target made discrete steps along the normal axis of the tracking airplane. From the pilot' s point of view, the target made random discrete steps along the vertical axis of the HUD. The steps occurred at bounded random time intervals and had bounded random amplitudes. The piloting task was changed from requiring time on target to requiring the pilot to attempt to achieve a "gun kill" against the target. Kills required 75 hits and the pilot was given 100 rounds; hence, 75 percent of the rounds available were required to kill the target. The stepping-target tracking task was shown to be an effective simulation technique for reducing PIO potential during control system design. 9

HIGH-ANGLE-OF-ATTACK CRITERIAEVALUATED

Anothermajorobjectiveof theHARV handlingqualitiesresearch wasthe evaluationof tech-

niquesfor predictingandspecifyinghigh-o_ handlingqualities.Usinglinearandnonlinearsimula-

tions andflight research data,the resultsfrom eachanalysistechniquewerecomparedwith the

pilot ratingsandcomments from the flight evaluations.

Thesedesignguidelinesandcriteriafell underthe generalsubheadings of open-loopperfor-

manceguidelines,lineartechniques, andtechniques compensating for systemnonlinearities. Con-

sideredin this study wereproposedhigh-_ nonlineardesignguidelinesand proposedhandling

qualitiescriteria and guidelinesdevelopedusing piloted simulation.Recentlyformulated time-

domainNeal-Smithguidelineswerealsoconsidered for applicationto high-ormaneuvering. Con-

ventionalenvelope criteriasuchasbandwidth, Neal-Smith,Smith-Geddes, andlower-orderequiv-

alentsystems-based guidelinesfoundin MIL-STD-17971° wereevaluated for possibleextension

to the high-t_regime.The following sectionsbriefly introducethesecriteria andguidelinesand

presentexampleresults.Table2 showsa very brief summaryof the generalhandlingqualitiesof

the controllaw variationsthatarereferenced throughout the paper.

Table2.HARV controllaw versionsummary.

NASA-0 NASA DrydenmodifiedMDA design(thrustvectoringonly).

Apparently good pitch tracking; poor acquisition predictability.Lateral axis

Version26

inertialcouplingandwind-axisbank-angle overshoot deficiencies.

Very bad pitch tracking (bobble, PIO tendency)but good pitch acquisition

Version27

characteristics. Goodlateral-directionalqualities.

Good pitch tracking and acquisition qualities. Same good lateral handling

Version28

qualitiesasVersion27.

NASA-1 NASA Langleydesign(thrustvectoringand/orforebodystrakes)

Vectoringonly,risk reductionfor ANSERcontrollaws.Severe pitch PIO during

NASA- 1A

trackingandrelateddifficultieswith acquisition. Lateralhandlingwasdifficult to

assess because of pitch PIO but generallygood.

Vectoringandstrakes. Pitchcontrollawredesigned andincludedthreedial-a-gain

ANSER

pitch sensitivityoptions.Majority of pitch PIO from NASA-1A eliminatedbut

somesensitivitywhile trackingremained.Goodlateralhandlingqualities.

Figure8 showsa summaryof the30° otlongitudinaltrackinghandlingqualitiesfor thecontrol

law versionsevaluated. This summarywill serveasa guideto the examplecriteriacomparison re-

suits that follow. Most noteworthyis the NASA-0 control law version series(Versions26, 27,

and28), where the trackingcharacteristics went from Level 1-2 to Level 2-3 and back to the

NASA-(} control law NASA-1 and ANSER control law Level 3 Level 2 Level 1 V 27 V 28 NASA-I'A ANSER V 26 General pilot commands during longitudinal tracking task: Version 26: "Pitch lightly damped; not PIO prone. Compensation: two hands."

Version 27: "Excessive pitch bobble during tracking; this affects the lateral axis as well. Tried to compensate by being very gentle with the longitudinal stick. A very difficult task."

Version 28: "Slight pitch bobble tandency....very predictable; pleased with it."

NASA-1A: "Strong PIO tendency....was shocked at how bad it was. I thought that I was maybe doing it wrong...

was not apparent that it was becoming convargent....forced to back almost completely out of the loop."

ANSER: "Slight longitudinal bobble....predictability was good, though. Compensation technique was to back off a little bit."

96054 Figure 8. Summary of 30 ° ot longitudinal tracking characteristics for tested control laws.

Level 1-2 boundaries, respectively. The pilot comments were considered more significant than the pilot ratings. The comments indicate a definite problem with pitch sensitivity introduced with Version 27 and subsequent improvement in Version 28 to a level somewhat better than that of both previous versions. This known and discernible version history was a valuable resource for evalu- ating the ability of the various longitudinal criteria to predict the same trends and levels. (It must be noted that there were very few pitch-tracking points flown using Version 26. As will be shown, poor handling qualities were typically predicted by the various criteria for this version, especially for high g levels. Little flight data exist to adequately validate or invalidate these predictions for Version 26. The data for this version are retained in the following examples for comparison of the predictions among the various criteria.) The other longitudinal control law variations flown were the NASA-1A and ANSER control laws. These variations provided a severe pitch sensitivity example with PIO susceptibility and the subsequent successful redesign, respectively.

Open-Loop Performance Guidelines A set of nonlinear design guidelines for poststall flight did not exist when the HARV thrust- vectoring and control law design began. A set of guidelines was developed at the outset of this pro- gram and iterated upon during the design process. 4,6 The initial set of open-loop performance guidelines is listed below: • Maximum pitch rate from 1-g and loaded conditions.

• Maximum pitch acceleration from 1-g and loaded conditions.

• Maximumroll ratethrough90° bank-angle changefrom 1-gandloadedconditions.

• Time through90° bank-angle change.

• Couplingcriteria(for example, maximumo_ andangle-of-sideslip excursions during rolls).

With theseguidelines,pitch andyaw acceleration requirements for the thrust-vectoringsystem

could bedetermined, andgoalsweresetfor thecontrollawdesigns. Detailsabouttheseguidelines

aregivenby Hoffler. 6

Overall results using these guidelines indicated that these guidelines worked well for the HARV and appear to be a good start for any poststall-capable airplane. Two significant lessons were learned about the use of these guidelines through ground-based simulation and flight test of the HARV airplane. In the longitudinal axis, the pitch-rate requirements for maneuvers initiated from low angles of attack were too high for the HARV to achieve while maintaining good handling qualities. In the lateral--directional axis, another guideline, the bank-angle overshoot criterion, was developed to make gross-acquisition maneuvers predictable (fig. 9). Figure 9 shows the region meeting the criterion (shaded area) as a function of ct for simulation results of three lateral- directional modes of the ANSER control law. This example displays the differences various con- trol effector combinations exhibited with respect to this design guideline. The bank-angle over- shoot criterion is also described by Hoffler. 6 This criterion essentially requires a consistent amount of bank-angle overshoot for angles of attack greater than 25 ° and dictates that the overshoot be "small." Observations of pilot performance led to the criterion. It was determined that when a tradeoff between bank-angle overshoot and rate was required, bank-angle overshoot should take precedence. This priority results from the fact that "getting there fast" is only part of the desired outcome. Stopping on the target rather than badly overshooting it or oscillating around it can often make up the difference in time to acquire the target.

9O Thrust vectoring only l-g roll 25,000 ft Strakes only Strakes and vectoring combined 6O Area meeting Wind-axis bank-angle overshoot, deg 3O F/A-18

I 1 I D I 1 J

10 20 30 40 50 60 70 o_, deg _os_ Figure 9. Open-loop performance design guideline: wind-axis bank-angle overshoot from reversal after 90 ° of wind-axis bank-angle change.

LinearAnalysis Techniques

Most conventional handling qualities criteria are formulated with assumptions of linearity.

These assumptions are typically appropriate for conventional aircraft and maneuvers. Several lin- ear criteria were evaluated to determine if these linear assumptions could also be appropriate for maneuvering in the high-a regime.

Proposed High-Angle-of-Attack Criteria and Guidelines High-or handling qualities design guidelines, developed by MDA, were formulated in terms of modal parameters as well as frequency-domain Bode response envelopes.] l These guidelines spec- ify handling qualities boundaries for 30 °, 45 °, and 60 ° oc; longitudinal and lateral axes; and tracking and gross acquisition. A limited evaluation of these guidelines was performed using the HARV da- tabase. Because these guidelines are restricted by the International Traffic in Arms Regulations, all figures have been sanitized by removing the scales.

McDonnell Douglas Aerospace-Proposed High-Angle-of-Attack Handling Qualities Guidelines: Modal Guidelines The modal formulation of these criteria require lower-order equivalent system (LOES) transfer function representations of the full-order dynamics. Longitudinal LOES matches could only be de- termined for the pitch stick-to--a transfer function at high angles of attack. The pitch rate-to-pilot stick and normal acceleration-to-pilot stick transfer functions contain terms dependent upon the load factor for each ot change, nlet which becomes essentially 0 g/rad at high or. Figure 10 shows the LOES formulation for the pilot stick-to--a transfer function matched the full-order system quite 4O linear model i _ High-order -- Low-order Magnitude, 20 _-- equivalentsystem dB -20 , , , , , , I , , , , , , , , Ph_glm, -100 -20O -300 10 -1 10 0 101 Frequency, red/sec 960549 Figure 10. LOES as a function of high-order linear model pilot stick-to--_ frequency response comparison for Version 28 at 30 ° a.

favorably.This transferfunctionwassufficientto providethenecessary parameters (short-period frequencyanddamping)for this criterion.

Figure 11showsa sample comparison resultof the predictedhandlingqualitiesfor threeflight

control law versions.The short-periodfrequency of the criteriaboundaries is scaledto 200knots equivalentairspeed (KEAS)and 100 KEAS for acquisition and tracking, respectively, before plot- ting. Predictions for longitudinal tracking and acquisition at 30 ° _ are plotted with the correspond- ing Cooper-Harper rating (CHR) from flight test in parentheses. Ratings of 1-3 correspond to Level 1 handling qualities; CHRs of 4-6 correspond to Level 2; and CHRs of 7-9 correspond to Level 3. No Version 26 gross acquisition performed in flight existed, so this point is not shown.

The gross-acquisition predictions indicate very similar characteristics with, at best, a median Level 2 rating. The flight results also reflect similar acquisition characteristics between the versions, but the overall level is nearer to Level 1 handling qualities than the predictions.

The tracking criterion reflects the trends observed in flight fairly well; Version 27 was signif- icantly worse than the other two versions, and Version 28 exhibited the best qualities of the three.

The Version 26 prediction was Level 2 as opposed to the CHR of 3 received in flight. All three versions received comments of varying levels of pitch-bobble sensitivity in tracking and acquisi- tion. The guidelines predict similar pilot comments based on the location of the analysis values.

The lateral-directional axis of the ANSER control law was designed using the modal represen- tation of these guidelines. ]2 Figure 12 shows a comparison of the closed-loop linear design as a function of the MDA modal flying qualities criteria for 1-g design conditions at several angles of Longitudinal acquisition Longitudinal tracking r KEAS =;200 KEAS = 100

• / L.,,I/

Short-period Short-period damping damping ....

i .... ;; i,, ,;]

Version 27 (6) -- 4, ; o---Version 28 (2, 4) "_Vereion 26 (3)

/

Flight CHRe In parentheses Flight CHI_S in pllrentt_l ires } I I I Short-period frequency Short-period frequency 960550 Figure 11. MDA longitudinal modal tracking criteria for 30 ° ct with Versions 26, 27, and 28 linear predictions and actual pilot ratings.

.... _ .... i .... I .... T .... [ ' ' ' 30 (_ guideline e5 ° (L /_ -- --- 45 (_ guideline / _ ..... 60 a guideline / \ • ANSER / Level 1 region \ /-- 45 ° (x Stability-axis / _ _ Leveilreglon roll-rata sensitivity,

,oo../

(deg/sec)/Ibf

/ S

e20 ° (X /,//C Level I region ] //.. •45 _a I /./: "" •6o° a i , , , , I , , , , I , , , , I i • . * i .... i ....

Roll-mode time constant, sac 960551 Figure 12. MDA lateral modal tracking criteria for 30 °, 45 °, and 60" o_ with ANSER linear design conditions.

attack. This figure shows the high-o_ criterion for lateral tracking in terms of closed-loop roll mode time constant as a function of stability-axis roll-rate sensitivity for the ANSER design. The region depicted by the solid line is the Cooper-Harper Level 1 region for 30 ° oc, the region depicted by the dashed line represents the Cooper-Harper Level 1 region for 45 ° or, and the region depicted by the dotted line represents the Cooper-Harper Level 1 region for 60 ° o_ In general, the closed-loop de- sign points follow the trend of Level 1 flying qualities as a function of ot according to MDA criteria.

The lateral-directional tracking tasks were flown by four pilots at 30 ° ct, by five pilots at 45 ° or, and by two pilots at 60 ° or. At 30 ° or, the ratings were mostly Level 1. At 45 ° 0t, the ratings were mostly borderline Level 1-2. The average CHR at 30 ° and 45 ° et was 3.5 and 4.4, respectively. At these angles of attack, desired criteria were met by all but one pilot, who rated the dynamics sig- nificantly worse than the other pilots did. At 60 ° or, the average CHR was 7. Flying qualities were degraded at 60 ° o_ which, based on the pilot comments and flight data analysis, was most likely because of control power limitations. More overall comments on the applicability of the MDA modal criteria are given at the end of the MDA Bode criteria section.

McDonnell Douglas Aerospace-Proposed High-Angle-of-Attack Handling Qualities Criteria: Bode Guidelines Level 1 Bode guidelines have been developed for both tracking and acquisition at 30 ° c_ and for poststall flight for the full-order pilot stick-to-_ transfer function. The discussion in this paper is for o_ command systems, although the reference also includes guidelines for rate command systems.Z

As in the modal guidelines,theLevel 1 boundaries for theBodeformulationarea functionof

flight condition. Guidelinesaredifferent for trackingandacquisition,and the 30°-o_ guidelines mustbescaledwith airspeed. In figure 13,thethick solid linesshowthe 30°-o_ trackingguidelines

for two Mach numbers corresponding to longitudinaldesigndatashownby the solid anddashed

lines.(Thesecriteriarequirethe boundaries to be scaledin frequencyby the ratio of velocity in KEAS to a standard of 140 KEAS. In order to overplot these two cases, the frequency of the data was scaled by the inverse of this ratio rather than scaling the boundaries.) These designs were made at 35°-a flight conditions. The 1-g case shows that the magnitude is greater than the upper bound- ary at high frequencies. The Mach 0.6 magnitude at a 4.5-g load factor remains within the Level 1 boundary for all frequencies. (The Mach 0.4 magnitude at a 2°g load factor results fell between these two cases.) The phase angle remains within the Level 1 boundaries for all cases.

Several pilots flew the 30°-or tracking maneuver, and varying degrees of sensitivity were expe- rienced depending upon the pilot and technique used. 13 Also, the high-speed flight tracking cases that were flown at approximately 20 ° o_ were less sensitive than the low-speed cases.

Although both the MDA modal and Bode criteria often predicted handling qualities indicative of those observed in flight, this was not always the case. The Bode criteria, although useful for en- gineering design, did not give good resolution in the discrimination of the observed tracking qual- ities differences for the NASA-0 control laws (Versions 26, 27, and 28) (fig. 14). Identifying definitive trends predicted among the versions is difficult. The Bode criteria tended to give more pessimistic predictions for 2-g and 3-g normal acceleration cases with this control law.

A few application observations can be made.

.

The MDA simulation study indicated that the desired dynamics are dependent on the task (gross acquisition or fine tracking). The boundaries for Level 1 tracking are different than those for acquisition at the same o_. Because the Level 1 regions for the two tasks do not overlap, the control system designer must choose the best compromise between the Level 1 regions or tailor the control system (short-period frequency and damping) for the separate tasks to be flown at the same flight condition.

2. Applying the modal form of the criteria requires the design to be adequately represented by a LOES model that may be difficult to obtain from flight data for the high-ct regime.

.

Applying these criteria requires a separate guideline for each tx and for each task. Future research is needed to determine if critical "state" parameters of the handling qualities problem in the poststall region can be identified, thereby reducing the number of guidelines required.

4. These criteria were developed from fixed-based simulation data. Validation (or invalidation) of these guidelines through in-flight testing is still lacking.

.

For the modal formulation, an attempt to use linear representations of short-period frequency and damping for predicting gross-acquisition characteristics does not seem appropriate. These maneuvers are often flown with large pilot inputs and flight condition changes that result in significant nonlinearities.

30 ° (x longitudinal tracking for (I command using G-to-pitch stick transfer function .... i • - 1 Level 1 boundary Frequency S_sled to14o KEAS m _ 1 g, Mach 0.26, .--- Level 2 35 ° (1, 25k ft 4.5 g, Msch 0.6, 35 ° G, 25k ft Magnitude , i i ..... i ........ L "_ Phase Lays, 2 _ -\ xx / scaled to 140 KEAS _ _ _ "] Frequency !

, * i i i i i i I i i , i , i , i I_t Frequency 960552 Figure 13. Bode tracking guidelines for angle-of-attack command with frequency response data from ANSER design at 35 ° ct with Mach 0.26 (1 g) and Mach 0.6 (4.5 g).

30 ° G longitudinal tracking for 0_ command using G-to-pitch stick transfer function ._-=.W-,mm_ ..... ' ..... ' ......... _ " " _ Level 1 boundary Lev__ Version 26, 2 g _ Level 2 Version 27, 2 g ___.._";.,_._ Version 28, 2 g MagnRude __<...

---...._-_.

Frequency scaled to 140 KEAS "

L.v.,2

!

Frequency scaled to 140 KEAS ' -.-_.

Frequency _60553 Figure 14. Bode tracking guidelines for 30 ° ot with frequency response data from Versions 26, 27, and 28.

Low-Angle-of-Attack MIL-STD-1797 Criteria Extended to High Angle of Attack Conventional low-o_ handling qualities criteria were evaluated for possible extension to the high-a flight regime. These linear techniques were tested for their ability to adequately predict fly- ing qualities trends and levels observed in flight. The traditional boundaries were used by default but were not considered to be necessarily appropriate for the high angles of attack. The adjustment of these boundaries for high-o_ application was recognized as a potential outcome depending on the comparison with flight results.

Lower-Order Equivalent Systems-Based Criteria Several conventional criteria require LOES transfer function representations of the full-order dynamics. Longitudinal LOES-based criteria that depend upon the time constant of the lag between flight path and pitch attitude (T02) and the n/o_ were found to be particularly problematic at high ct.

These criteria include the control anticipation parameter (CAP); short period frequency, tOsp as a function of n/t_; and f.OspTo2 as a function of short period damping, _sp" The CAP is defined as: CAP = _sp/(n/tt) Because the lift-curve slope and, therefore, nRx become essentially 0 g/rad at high angles of attack, these criteria become meaningless. Figure 15 shows the general migration of predicted handling qualities with t_ using the CAP criterion as an example. As ct increases, the value of the CAP ap- proaches infinity. These conventional handling qualities criteria are based on the assumption that 1/To2 is approximately equivalent to the lift-curve parameter, which is appropriate for unaugment- ed aircraft. This assumption seems inappropriate when addressing highly augmented vehicles be- cause these parameters are not designed to be similar. Such criteria were abandoned for further consideration in this study.

MIL-STD short-period dynamic requirements: Catetjory A 103 ::!:.!!!i::ii!!i!!::i::_!i_i!i::!:.!!i::i::!::Higher_:,!! i!::::! !ii!!!::i i!i!!i 101 .:::: , , , , , , , CAP, _ i._ _!_: !i!ii! _!i ...._: !':i! !_: '!': ':!?: !i !i!!': .... ......

:iii ..... :''_''i ....... l_'r ..........................

10 0 3 ..... ' ........................

::::: ::: : i :Liiviil::i : :::: :::::::::::::::::::::::: 10-1 i !i: !

.... i .... i

.....iii iii iiiiii!iiii .... I

10 -2 " " ; ; ; ;;i ; ; ......

10 -1 10 0 101 Short-period damping 960554 Figure 15. Migration of CAP criterion predictions with increased o_.

Neal-Smith Criterion The Neal-Smith criterion, which is also included in MIL-STD-1797, was examined in detail for application to high angles of attack, l° This criterion was developed using a database produced by the Calspan Corporation (Buffalo, New York). j4 The criterion assumes a longitudinal tracking task with loop closure around a compensator of predefined form, an airframe, and flight control system dynamics. The compensator model is of the form kpe_Xpaots (Tlead s + 1) (Tlag s + 1 ) and closes the pitch attitude-to-pilot stick transfer function.

Figure 16 shows the Neal-Smith predictions for the NASA-0 three-version history previously described for tracking at 30 ° ct using a bandwidth assumption of 3.0 rad/sec and compensator time delay of 0.3 sec, The bandwidth value was reduced from the recommended 3.5 rad/sec for this class of aircraft as a result of engineering judgment and simulation results implying that the task frequen- cy requirements tend to decrease with increasing et. Overall HARV experience has indicated that these values were appropriate for correlation with flight results. Figure 16 also shows the variation with normal acceleration (1 g, 2 g, and 3 g) and the corresponding CHRs from flight test. (The flight conditions for these cases were between 2 g and 3 g.)

Predicted in the figure is a general improvement in the handling qualities as normal accelera- tion increases for each control law version. As mentioned earlier, the Version 27 flight test exhib- ited objectionable pitch bobble and sensitivity problems. From the pilot ratings and comments, 15 w )K Version 26 (flight handling qualities rating = 3) + Version 27 (flight handling qualities rating = 6; "excessive pitch bobble") O Version 28 (flight handling qualities rating : 3, 4) Resonance pea_ Level 2 dB

/

I 1\ I I

-20 20 40 60 80 Lead compensation, deg 960555 Figure 16. 30 ° tx Neal-Smith results for Versions 26, 27, and 28 at 1 g, 2 g, and 3 g.

Version28 appeared to givenotablybetterhandlingqualitiesthaneitherof the otherversions.The

predictionsindicatethe general trendsobserved in flight for thelow-g cases. The 3-g case predicts Version 26 to be inferior to Version 27, as was the case with the MDA modal criteria. In all cases, Version 28 was predicted and observed to be superior to the other versions. Additionally, all ver- sions were predicted to have Level 2 flying qualities, with the Version 28 results near the Level 1-2 boundary. Based on the flight results, one would expect the Version 27 level to be close to Level 3 and the Version 26 and 28 predictions to be close to the Level 1 region.

An important enhancement to the Neal-Smith criterion was the "carpet plot" representation (fig. 17). This representation was included to detect handling qualities "cliffs" where a severe deg- radation in flying qualities can occur for a configuration with changes in pilot gain or task band- width requirements. ]5 Figure 17 shows a family of Neal-Smith predictions for the Version 27 and Version 28 configurations but with variations in the bandwidth requirement and another task per- formance metric called "droop." Thus, robustness to gain and phase variations can be depicted by such plots. The asterisk on the figure represents the nominal prediction (bandwidth = 3.0 rad/sec; droop = -3 dB). The bandwidth was varied by 0.5 rad/sec from 2.5 to 3.5 rad/sec. The droop ranged from 2.5 to 3.5 dB in increments of 0.5 riB. (The Version 28 carpet plot collapses to a line at 2 dB for the low-bandwidth cases.) Steep vertical slope and rapid traversing across the parameter plane indicate poor robustness, as can be seen in the Version 27 carpet plot. Here, the prediction is in the median Level 2 region, but a change of only 0.5 rad/sec in bandwidth moves it fax into Level 3 flying qualities. Comparatively, the Version 28 predictions do not move as far with each increment, and their movement tends to be horizontal, indicating good handling qualities robustness.

This "third-dimension" view of the Neal-Smith criterion greatly enhanced its applicability and usefulness. It is important to have insight not only into the predicted nominal result, but also into NesI-Smlth carpet plot 12-- )1( Nominal result (bandwidth = 3 rad/$ec, time delay = 0.3 sec) 10 -- _ Level3 /_ \ Veral 8 -- O_ Resonance peak, dB 6 -- /'

/4 / f

Levsl 2 / \ / /

4 -- / / /Increasing 2 \ // Level 1

I/ i / bandw, th

0 I J I\ I I

- 20 0 20 40 60 80 Lead compensation, (:leg 960556 Figure 17. Neal-Smith "carpet plots" results for Versions 27 and 28 at 2 g and 30 ° ct.

how static that answer is in the presence of task and pilot variations. The carpet-plot analysis dra- matically reflected the actual results obtained in flight. Version 28 exhibited reliably good tracking handling qualities. Version 27 could produce good handling qualities as the pilot backed out of the loop (low bandwidth), but this latter result was very sensitive to changes in pilot gain or task requirements.

Smith-Geddes Criteria The Smith-Geddes criteria were designed to predict longitudinal and lateral-directional PIO susceptibility. _6 As in the Neal-Smith criterion, the longitudinal criterion examines the pitch attitude-to-pitch stick transfer function in the frequency domain but with no assumption of a compensator outer-loop closure. The slope in the 1-6 rad/sec band is examined to determine the critical frequency, co c = 6.0 + 0.24 * (slope in dB/octave) The phase angle of the transfer function at this critical frequency is related to average CHRs (fig. 18). Figure 18 shows the results for a normal acceleration of 2 g for the NASA-0 control laws (Versions 26, 27, and 28). The same comparative trends between the versions are evident; howev- er, the overall predicted levels worsen (by approximately 1-2 CHRs) when compared to the Neal- Smith predictions and the actual flight results. Versions 26 and 27 are predicted to be very near the Level 2-3 boundary, and Version 28 is predicted to be an improvement but only to the middle of the Level 2 region. The flight-observed CHRs (also shown in the figure) for these versions were generally more favorable. No serious effort has yet been put forth to adjust the criteria for the high-oc flight regime. The simulator studies used to produce the high-tx modal and Bode criteria 10 -- Smlth-Geddes overlay plot 9 -- ' "-. _ _Predlcated o N .... CHR All data: 2 g, 30 (X 8 .... "-,.X 7 r_ion 27 Level 3 Cooper-Harper 6 _ _ -- "_'r'\."-oPoNI_I V;rsi_°n26 Level2 rating _ J_'i _Ve I Flight CHRs rbion 28

4- I \IN- •

'. "J, \ ' "_, _ Level 2 \ / ___Fllght CHRa " _ ' _ " _ ' _.

I L I I 1 L i t

-220 -200 -180 -160 -140 -120 -100 -80 Phase angle at bandwidth, deg 96O557 Figure 18. Smith-Geddes criterion predictions for Versions 26, 27, and 28 at 2 g and 30 ° ct.

presented earlierpointedto a decrease in bandwidthrequirement asthe (_ increased. This decrease would tend to force the predictions to have better correlation with the flight results. More research and analysis is needed to place these boundaries more definitively.

Bandwidth Criterion Another conventional longitudinal tracking criterion included in MIL-STD-1797 is the band- width as a function of time delay criterion, l0 Pitch attitude-to-pilot stick transfer function frequen- cy responses are used to determine the bandwidth frequency. The effective-time delay parameter is computed from the phase angle at twice the 180 ° phase crossover frequency and is plotted as a function of the bandwidth frequency for the conventional envelope (fig. 19).

Figure 19 shows the bandwidth criterion predictions for the NASA-0 control law versions (Versions 26, 27, and 28) for 1 g, 2 g, and 3 g. The same general improvement of handling qualities predictions with increased normal acceleration exists as in the Neal-Smith case. Additionally, the same trend among the versions exists, with Version 28 consistently predicted to be more favorable than the other two versions. The overall handling qualities predictions show a close grouping near the poor side of the Level 2 region with the most favorable prediction (Version 28, 3 g) being in the middle of this region. Like in the Smith-Geddes example, these results do not reflect the overall flying qualities levels exhibited by these versions in flight. Additionally, the flight test indicated much stronger discrimination between the sensitivity of Version 27 and the redesigned Version 28 configurations than is indicated by the bandwidth criterion predictions. Perhaps, as in the Smith- Geddes case, this discrepancy points to a potential requirement to adapt the boundaries of the level regions for the high-oc flight regime. Reducing the required bandwidth frequency would shift the level regions to the left as indicated in the figure, improving the overall predicted levels. More re- search and analysis is needed to adequately determine what adaptation, if any, is required. Still, the criterion reflected the trends among versions observed during flight test.

.25 -- Version 26 + Version 27 O Version 28 .20 -- Level 3 .15 -- Effective time delay, 8ec ,/ Level 2 .10 -- f _-Potential hlgh._ boundary / .05 -- 1_2_3g I [ Level 1_'%%%%

I

5 10 15 Bandwidth, red/nc _0558 Figure 19. Bandwidth criterion predictions for Versions 26, 27, and 28 at 2 g and 30 ° ot with pos- sible adjusted high-o_ boundary.

NonlinearityCompensating Techniques

Handling qualities criteria that include various levels of systemnonlinearities were also

evaluated. Thesecriteriaarediscussed in thefollowing sections.

NonlinearDescribingFunctionsandLimit-CycleAnalysis

Because of a severe episode of PIO encountered earlyin theflight testof the NASA-1A control

law anda resolvethatthis shouldnot happenagainwith the ANSERdesign,work wasinitiatedin

assessing the stateof the artin predictingPIO susceptibilityat high-_ flight. To makethis assess-

ment,the"predictions"(madeafterflight test)for theNASA-1A controllaw wereusedin conjunc-

tion with the predictionsfor the ANSER controllaw andthe subsequent flight results.A detailed summaryof this research is given by Bacon. IvIn brief, two strategies wereemployedto predict PIO susceptibility:

• Various linear PIO criteria were applied, incorporating describingfunctions to handle

nonlineareffects(for example, actuatorratesaturation).

• Multivariable limit-cycle approach wascoupledwith variouspilot models,both linear and

nonlinear,to uncoverpotentialfull-blown PIO tendencies. ]8

The linear procedures considered includethe Neal-Smithcriterion, 14the Smith-Geddes crite-

ria, 19 and the bandwidth criterion 2° described earlier in this paper. The rate saturation element of the stabilator actuator is replaced by a describing function that is a gain between zero and unity. 21 A lo- cus of handling quality parameters, different for each method, is then generated as a function of commanded actuator rate from the changing pitch-attitude frequency response. This nontraditional use of describing functions to ascertain limit-cycle or PIO potential of the pilot/vehicle system led to diagrams such as that shown in figure 20 depicting the effect on the Neal-Smith linear result for two flight conditions with the inclusion of these describing functions.

(_ Linear result • .51 < describing function variations < 1

ii! iiii iiiii iiii iii il ' iiiI

..... i .......... t*veJSl Resonance 10 peek, dB 8 ....... .............. .................

-2 - 20 0 20 40 60 80 Lead compeneltlon, deg 960s59 Figure 20. Quasilinear extension of Neal-Smith criterion with NASA-1A predictions for two flight conditions.

Themultivariablelimit-cycle approach usedallowsnonlinearbehaviorto beconsidered in both

the vehicleandpilot. Theapproach combinesstructured singular-value methods, multivariablede-

scribing function methods,and pilot modeling to uncoverpotential PIO tendencies. 18Unlike

Anderson'sapproach, a relay-switchingmodelof the pilot is considered in additionto linear rep- resentations of the pilot. In the approach, structured singularvalueslimit the search space,provid- ing a necessary condition for solution, and also producea likely pilot candidatefor a set of uncertainpilot dynamicsconsidered.

Althoughdetailedresultsaregivenby Bacon, 17 somegeneralresultsaresummarized here.All

linear methodsand quasilinearextensionsgave consistentresults in predicting PIOs. For

NASA- IA, thequasilinearextensions thataccounted for ratesaturation werenecessary to predict PIOsbecause thelinearanalyses alone(withoutusingthecarpet-plot representation previouslydis- cussed) revealed no PIO tendencies. Thequasilinear extensions, however,failed topredictthelev- el of commanded actuatorraterequiredto initiate a PIO episode. The actualcommanded actuator

ratewassmallerthanpredicted. FortheANSERcontrollaw, the quasilinear extensions reflected

the trend in observedsensitivitybut failed to predictthe nuisanceoscillations(bobbles)experi- encedin flight. To complicatematters,not all of the pilots experienced the nuisanceoscillations

usingthe ANSERcontrollaw. It became apparent thatthe collectivetestpilots' strategyusedin

uncovering PIO susceptibilitywasnot covered adequately by the linearor quasilinear extensions.

The second approach, whichmoreaccurately considers theeffectof multiplenonlinearities aswell asthe uncertaindynamics of thepilot, wasusedto address the limitationsof the first.

Proposed Time-DomainNeal-SmithCriterion

Theconventionalfrequency-domain Neal-Smithcriterionpreviously discussed has been refor-

mulated in the time-domain by Calspan Corporation under contract to the U.S. Air Force. 22 The same compensator is retained in the form kpe xPa°'s (Tlead s + 1) (Tlag s + 1) This compensator model is embedded into the nonlinear simulation of the subject aircraft, clos- ing the loop around the full-order nonlinear control system, actuator models, and aerodynamics.

The required task is to respond to a step error of 5 ° pitch attitude within a required acquisition time, D. An optimization loop is wrapped around this configuration to find the gain, kp and com- pensation, Zlead and Tlag such that the root mean square (RMS) error between the response and the target pitch attitude is minimized after D. The resulting RMS error is plotted as a function of the phase angle at the bandwidth frequency on the time-domain Neal-Smith parameter plane where handling qualities regions are drawn.

The example time history (fig. 21) shows the optimized result for the HARV Version 28 configuration at approximately 27 ° ct using an acquisition time of 1.5 sec. The resulting RMS error was approximately 0.2, and the phase angle at the bandwidth frequency was 64.2 ° . Figure 22 shows this result, along with those using a D of 1.3 and 1.7 sec, in order to estimate the robustness of this result to variations in bandwidth. The tracking handling qualities for this configuration in flight were observed to be near the Level 1-2 boundary as the prediction indicates. Additionally, this Target pitch attitude angle Pitch attitude 30 angle, deg 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Time, sec _0560 Figure 21. Time-domain Neal-Smith constrained pitch-attitude response with Version 28 at 27 ° o_ optimized time history.

5.0 , f 4.5 4.0 3.5 Level 3 3.0 RMS (pitch attitude angle error) 2.5 ............................ i ........................... ............................

after D, Level 2 deg 2.0 1.5 1.0

/ 1 o:13

0.5 ............ / ...... : .................. _ ____\ ____-_ D = 1.5 / : o=,.,---_-_ .........

-50 Compensation phase angle, deg gsossl Figure 22. Time-domain Neal-Smith robustness predictions for Version 28 with variations in required acquisition time.

configurationexhibitedgoodrobustness to variationsin taskandpilot gain as is predictedby the very little migrationof thepredictionastheacquisitiontime is changed.

The potentialbenefitsof thiscriteriontothehigh-ctflight regimewererecognized. Theinherent

inclusionof system nonlinearities andfull-orderdynamics is attractive, especiallyathigh anglesof attack.Additionally, thetimerequiredonconditionis approximately 5 sec,whereas in orderto gen-

eratea linearizedfrequency response from thenonlinearsimulationusinga frequencysweepinput

andFastFourierTransformanalysis, morethan50 secarerequiredto producemeaningfulresults.

For high-orflight, altitudeis lost at a dramaticrate,resultingin large changes in flight condition fromthe beginning to theendof themaneuver. Thus,theshortmaneuver time requiredfor this anal-

ysistechniqueis animportantbenefit.Onedrawbackof thecriterion is thevery long optimization

time requiredfor eachcasebecause the simulationneeds to betrimmedandoperated for 5 sec,and

theoptimizationcodeselectsa newvariableset,for eachiteration.

Somemodificationswereintroduced for adaptation to the HARV high_ problem.Most note-

worthy wastheadditionof anintegrator pathin thepilot modelnecessitated by the_ command con- trol usedathighanglesof attack.Additionally,thenominal D, which is the time-domain equivalent to bandwidth in the frequency domain, was increased from that of the low-a formulation. This in- crease was done based on engineering judgment and experience with other criteria adaptation to high angles of attack. Final recommendations forD in this criterion for high ct require further study.

It should be noted that this work has not been finalized and the criterion cannot be considered vali- dated for high-a application as of yet. The results, however, have been promising. Additionally, no obvious reason exists why this technique could not be applied to the lateral axis as well.

HIGH-ANGLE-OF-ATTACK HANDLING QUALITIES DATABASE The archival of the HARV high-ct handling qualities raw information into a usable database was an objective of the HARV handling qualities research. Several components make up the HARV database: • Linear models of all control law versions.

• Nonlinear six-degrees-of-freedom simulations.

• Piloted simulation handling qualities database for baseline flight comparison.

• Pilot ratings and comments (transcribed and digitized audio) from flight.

• Pilot ratings and comments from dome simulations.

• HUD or glareshield video for each flight maneuver.

• Flight data in engineering units.

Figure 23 shows a sample HARV handling qualities comment card transcribed from the digi- tized audio. This database is currently in the process of being archived in a form useful for future analysts. This database should help engineers to further develop and evaluate criteria, plan flight programs, and add to the database where it is lacking. An important by-product of this handling qualities research has been the production of a MATLAB ® (The MathWorks, Inc., Natick, Massa- chusetts) handling qualities toolbox of analysis routines. All of the examples given in this paper were produced using these tools developed by the NASA team.

HARV HQ Comment Card HARV flight 319 card 12A 45 Much Longitudinal/Later al Traclung T'V Flight: 319 $0 -- Ak_a Card #: 12A CLAW: 151.1 Mode: "IV Gain Set: Medium Pilot: Schneider Target A/C: 850/Smolka Much No: 0.45

t/t '

Altitude: 25000.

lo Target AOA: varies sl ....

Maneuver: .45M Tracking , I I Actual AOA: 11-46 (30 avg.)

Reticle/Depression: 12.5/80mil Desired: within 5rail 50% time and within 25mil remainder Fd.e=lh319 12a; Signal Suffix=lno,P.e]: Date=l_txel Adequate: within 5rail 10% time and within 25mil remainder Date: 07Sept95 Maneuver,... Burner,... ok, on him.., a little longitudinal bobble, 32 alpha.., nose.., right wing.., left wing. Kind of a hellico- [Controller] Reverse. [Pilot] nose.., reverse, good.., see the pitch bobble? [Controller] Reverse. [Pilot] here comes the reverse._ [Controller] Reverse .... and recover, altitude. [Pilot] OK, knock it off. [Target] Knocking it off. [Pilot] Very good.

(snip to post-maneuver comn_nts) Attitude Control: Undesirable Motion: OK, comments; undesirable motions; the only one that I really did not like was the longitudinal pitch bobble that we've seen before on other tracking tasks. And it's still there, probably due to the major dumbbell effect that we have with the an'plane.

Predictability: Predictability was pretty good, except in the long axis. You could sometimes over-control pretty easily there and get the pipper off from where you wanted it.

Initial Response: Initial response was good in both axes.

Aggressiveness Effects H/Q: Aggressiveness had a major effect on the longitudinal pitch bobble. If you really got on it it wasn't.., wasn't gonna settle down. And you had to compensate with.., being a little more open hand, little bit less aggressive with the controls in that axis. And... trying to... separate out the lateral control movements from the longitudinal. That's a little bit aggravating too when you go from wingtip to wingtip.

Compensation Techniques: Roll Performance: Roll performance, I thought was quite good.

RPC: Feel System: Forces: And ... forces were light, satisfactory.

Control Motion: Control motions; small.

Harmony: Harmony was good.

Nonilnearlly: It was linear.

Cooper Harper Rating: Longitudinal HQR: [4] Longitudinal first. Controllable? Yes. Adequate performance? Yes. Satisfactory without improvement? No.

Desired performance required moderate compensations. I'd give it a 4. I think ! had desired performance there, but I definitely did not like the longitudinal bobble, and you can ask Keith to think about that for a minute and see if he agrees.

Laterai/Dir. HQR: [3] And for the lateral axis... Controllable?Yes. Adequate performance? Uh, yes. Satisfactory without improvement?

I would say "yes". And I would say.., minimal compensation in the lateral axis. HQR 3.

Confidence Rating: [!] And... confidence rating: I'd give it a 1 on that one.

PIO Rating: Other Comments: Times: TargetS: 11:11:18 (right turn} l_neuv_r: 11:11:21 Acquisitic_: 11:11:28 (42°AOA) Knock It Off: ii:12:i0 Comments Beg/n: 11:12:31 _ts End: 11:14:29 Figure 23. Sample HARV database pilot comment transcription.

RECOMMENDATIONS

Basedon this research, the following recommendations for furtherinvestigationcanbe made:

• A dedicatedhigh-orhandlingqualitiesinvestigation using a true variable-stabilityaircraft

wouldgreatlyimprovetheunderstanding in this area.This investigation would allow greater

variationin dynamics, especially in the lateral--directional axes, to morecompletelyevaluate

handlingqualities criteria, including the placementof level boundariesfor conventional

criteriafor this regimethantheHARV handlingqualitiesresearch programallowed.

• More detailed evaluation of the current data using lateral--directionalcriteria should

beconducted.

• Direct comparison with X-31A handlingqualitiesresultsshouldbeperformedandreported.

• The high-tx pyramid maneuvers developedusing the HARV aircraft shouldbe seriously

considered for adoptioninto anyfuture high-orflight research program.The maneuvers can

also improve efficiency of flight test by doubling as envelopeexpansionand research

maneuvers simultaneously.

• Conventional, low-o_ handlingqualitiescriteria(Neal-Smith, Smith-Geddes, andbandwidth)

shouldalsobe seriouslyconsidered for applicationto high-o_ design.Boundaryredefinition

assuggested in this papershouldbeconsidered.

• Nonlineartechniques showedgreatpromiseandshouldbe exploredthoroughly,especially

for applicationto high anglesof attackwherenonlinearities areprevalent.

• Pilotedsimulationhandlingqualitiesevaluation shouldbe exploitedbut with caution.Often

handlingqualitiesdeficiencies (especiallyPIO) arenot readilyuncovered from simulation.

The "steppingtarget"simulationtaskcanbe very effectiveat exposingdeficiencies before

flight thatmight go unnoticedotherwise.

SUMMARY

A flight research studyof high-angle-of-attack handlingqualitieswasconducted at theNASA

Dryden Flight Research Centerin cooperation with the NASA Langley Research Centeron the

F/A-18 High Alpha Research Vehicle (HARV). The HARV aircraft,equippedwith a reconfig-

urableresearch flight controlsystem,multiaxisthrust-vectoring controlsystem,andforebodyvor- tex controlusing the actuated nosestrakes for enhanced roiling systemintegratedinto the stable high-angle-of-attack F/A-18airframe,wasanexcellentfacility for this investigation.Theresearch objectiveswereto createa high-angle-of-attack handlingqualitiesflight database, developappro-

priate researchevaluation maneuvers,and evaluate high-angle-of-attackhandling qualities

guidelinesandcriteria.

Maneuvers weredeveloped andflight-validatedfor theevaluationof high-angle-of-attack han-

dling qualities.Ground-based simulationswereusedextensivelyto developandrefinethesema-

neuvers.Somein-flight trial anderror wasrequiredto refinethem.Thesemaneuvers werevery

effectiveat discriminatinghandlingqualitiesfrom 30° through60° angleof attack.

Anothermajorthrustof theHARV handlingqualitiesresearch wastheevaluationof techniques

for predictingandspecifyinghigh-angle-of-attack handlingqualities.Using linear andnonlinear

simulationsandflight research data,theresultsfrom eachanalysistechniquewerecomparedwith

thepilot ratingsandcomments from theflight evaluations. Considered in this studywereproposed

high-angle-of-attack nonlineardesignguidelinesdevelopedby NASA Langley Research Center

andproposedhandlingqualitiescriteria andguidelinesdevelopedusing piloted simulation.Re-

cently formulated time-domainNeal-Smithguidelineswere also consideredfor applicationto

high-angle-of-attack maneuvering. Conventional envelope criteriasuchasbandwidth,Neal-Smith,

Smith-Geddes, and lower-orderequivalentsystems-based guidelinesfound in MIL-STD-1797

wereevaluatedfor possible extensiontothehigh-angle-of-attack regime.Additionally, themaneu-

vers themselves were examinedas potentialevaluationtechniquesand guidelinesincluding a

limited validationof the proposed standard evaluationmaneuvers set.

Resultsto datesuggest thatthetraditionallineartechniques canbeeffectively extended to this

flight regimein spiteof theincrease in nonlinearityassociated with high-angle-of-attack flight. As- sumptions of linearity appear appropriate, providedtheevaluation is limited to relativelysmallper-

turbations abouta quasisteady-state condition.Thus, handlingduring coarse-andfine-tracking

maneuvers could generallybe predictedusing linear techniques suchas the Neal-Smith,Smith-

Geddes,andbandwidthcriteria.Maneuvers requiringlargeinputsor resultingin greatchangeof

flight condition were sharplyaffectedby nonlinearitysuchas control surfacerate andposition

saturation.Nonlinear extensionsto the linear techniques,such as describing functions and

multivariablemethods, aswell asalternativeguidelines wereevaluated andfoundtobe promising.

It wasalsoshownthatfor this flight regime,criteriaor guidelines requiringlower-orderequiv-

alent systemsevaluationwere particularlyproblematic.Includedin this categorywere military

standardhandling qualitiescriteria andproposedhigh-angle-of-attack modal handling qualities

criteria. Anothersetof proposed guidelinesthatwereframedin termsof transferfunctionbound-

ariesin the frequencydomainwasalsoevaluated. However,anattempttopredictgross-acquisition

characteristicswhere nonlinearity likely aboundsbecauseof large inputs and flight condition changes wasnot considered appropriate usinga linear frequency response representation.

The research experimentconductedusingthe HARV aircraftresultedin a comprehensive yet

limited handlingqualitiesflight database for the high-angle-of-attack flight regime where none

previouslyexisted,andimportantinsightscanbederivedfrom it. The studyevaluated longitudinal

and lateral-directionalgross-acquisition and tracking handlingqualities from 30° through 60°

angleof attack.Numerouscontrollawconfigurations andadvanced controleffectors,includingthe

forebody strakesand thrust vectoring,wereusedto vary both availablecontrol power and the

closed-loopdynamics.The handlingqualitiesof theseconfigurationvariationsrangedfrom excel-

lent trackingand gross-acquisition characteristics to very poor handlingqualities,including sus-

tained pilot-inducedoscillations.An extensivesimulationdatabase using fixed-basedomeand

somemotion-based simulationwas createdfor comparisonwith flight results.Head-updisplay

video and pilot commentswere catalogedwith the flight time-history data. An importantby-

productof this handlingqualitiesresearch hasbeentheproductionof a MATLAB ®handlingqual-

ities toolboxof analysisroutines.All of theexamples givenin thispaperwereproduced usingthese toolsdevelopedby theNASA team.

REFERENCES

IStoliker,P.C.andBosworth,JohnT., Evaluation of High-Angle-of-Attack Flying Qualities for

the X-31A Using Standard Evaluation Maneuvers, NASA TM-104322, Sept. 1996.

2Kidman, David S., Vickers, Jesse E., Olson, Brian P., and Gerzanics, Michael A., Evaluation of the F-16 MultiAxis Thrust Vectoring Aircraft, AFFFC-TR-95-12, Sept. 1995.

3Bowers, Albion H., Pahle, Joseph W., Flick, Bradley C., and Rood, Richard L., "An Overview of the NASA F-18 High Alpha Research Vehicle," High-Angle-of-Attack Technology Conference, Hampton, Virginia, Sept. 1996.

4Hoffler, Keith D., Brown, Philip W., Phillips, Michael R., Rivers, Robert A., Davidson, John B., Lallman, Frederick J., Murphy, Patrick C., and Ostroff, Aaron J., "Evaluation Maneuver and Guideline Development for High-Alpha Control Law Design Using Piloted Simulation," AIAA-94-3512-CP, Aug. 1994, pp. 435-.448.

5Pahle, Joseph W., Foster, John V., Bundick, W.T., and Wichman, Keith D., "An Overview of Controls and Flying Qualities Technology on the F/A-18 High Alpha Research Vehicle," High- Angle-of-Attack Technology Conference, Hampton, Virginia, Sept. 1996.

6Hoffler, Keith D., Messina, Michael D., and Wilson, Ronald J., "Flight Test and Piloted Simulation Assessment of Evaluation Maneuvers and Nonlinear Guidelines Developed for High- Alpha Control Law Design," High-Angle-of-Attack Technology Conference, Hampton, Virginia, Sept. 1996.

7Wilson, David J., Riley, David R., and Citurs, Kevin D., "Aircraft Maneuvers for the Evaluation of Flying Qualities and Agility," WL-TR-93-3081-3084, vol. 1--4, Aug. 1993.

8Klyde, David H., Citurs, Kevin D., and Mitchell, David G., "In-Flight Evaluation of the Standard Evaluation Maneuver Set (STEMS) with the NASA F/A-18 HARV," High-Angle- of-Attack Technology Conference, Hampton, Virginia, Sept. 1996.

9Hoffler, Keith D. and Carzoo, Susan W., "A Piloted Simulation Test Technique Developed to Reveal Pilot Induced Oscillation (PIt) Tendencies During Tracking Tasks," High-Angle-of-Attack Technology Conference, Hampton, Virginia, Sept. 1996.

I°U.S. Department of Defense, Flying Qualities of Piloted Vehicles, MIL-STD-1797, Mar. 1987.

l lWilson, David J. and Citurs, Kevin D., High Angle of Attack Flying Qualities Design Guidelines, NASA CR-4681, Feb. 1996.

_2Murphy, Patrick C., Davidson, John B., Lallman, Frederick J., Morelli, Eugene A., Messina, Michael D., Connelly, Patrick J., and Thomson, Michael P., "An Evaluation of Design Methodology and High Alpha Design Criteria for the ANSER Lateral-Directional Control Law," High-Angle-of-Attack Technology Conference, Hampton, Virginia, Sept. 1996.

13Ostroff, Aaron J. and Wichman, Keith D., "Flight Evaluation of a Variable-Gain Output Feedback Longitudinal Control Law for the NASA F/A-18 High Alpha Research Vehicle (HARV)," High-Angle-of-Attack Technology Conference, Hampton, Virginia, Sept. 1996.

14Neal, T. Peter and Smith, Rogers E., "An In-Flight Investigation to Develop Control System Design Criteria for Fighter Airplanes," AFFDL-TR-70-74, vol. 1, Dec. 1970.

15Bailey, Randall E. and Smith, Rogers E., "Analysis of Augmented Aircraft Flying Qualities through Application of the Neal-Smith Criterion," AIAA-81-1776, Aug. 1981.

16Smith, Ralph, "The Smith-Geddes Criteria," SAE Aerospace, Control & Guidance Systems Symposium, Reno, Nevada, Mar. 1993.

17Bacon, Barton J., "PIO Susceptibility of the F-18 HARV: Analysis/Prediction vs. Flight," High-Angle-of-Attack Technology Conference, Hampton, Virginia, Sept. 1996.

18Anderson, Mark R. and Page, Anthony B., "Multivariable Analysis of Pilot-in-the-Loop Oscillation," AIAA-95-3203-CE Aug. 1995, pp. 278-287.

19Smith, Ralph H. and Geddes, Norman D., "Handling Quality Requirements for Advanced Aircraft Design: Longitudinal Mode," AFFDL-TR-78-154, Aug. 1979.

2°Mitchell, David G. and Hoh, Roger H., "Development of a Unified Method to Predict Tendencies for Pilot-Induced Oscillations," WL-TR-95-3049, June 1995.

21Adams, Richard J., Buffington, James M., and Banda, Siva S., "Active Vortex Flow Control for VISTA F- 16 Envelope Expansion;' AIAA-94-3681, Aug. 1994, pp. 1259-1269.

22Bailey, Randall E. and Bidlack, Timothy J., Quantitative PIO Criterion Development - Time-Domain Neal-Smith Criterion, Calspan Final Report no. 8184-12, Aug. 1995.

REPORT DOCUMENTATION PAGE Form Approved OMB No. 0704-0188 Public reporting DurOen for this coltectpon of information is esttmatecl to average 1 hour per response, inclucling 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 aspacl of this collection of information including suggestions for reOucing this burclen, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management and Bu0gat, Paperwork Reduction Project (0704-0188), Washington, DC 20503 1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORTTYPE AND DATES COVERED November 1996 Technical Memorandum 4. TITLE AND SUBTITLE 5. FUNDING NUMBERS High-Alpha Handling Qualities Flight Research on the NASA F/A-18 High Alpha Research Vehicle WU 505-68-30 S. AUTHOR(S) Keith D. Wichman, Joseph W. Pahle, Catherine Bahm, John B. Davidson, Barton J. Bacon, Patrick C. Murphy, Aaron J. Ostroff, Keith D. Hoffler 7. PERFORMING ORGANIZATION NAME(S) ANDADDRESS(ES) 8. PERFORMING ORGANIZATION REPORT NUMBER NASA Dryden Flight Research Center P.O. Box 273 H-2138 Edwards, California 93523-0273 9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSORING/MONITORING AGENCY REPORT NUMBER National Aeronautics and Space Administration NASA TM-4773 Washington, DC 20546-0001 11. SUPPLEMENTARY NOTES Presented at the NASA Langley High-Angle-of-Attack Conference, Langley Research Center, Sept. 17-19, 1996.

Keith Wichman, NASA Dryden Flight Research Center, Edwards, California. Joseph Pahle, NASA Dryden Flight Research Center, Edwards, California 12e. DISTRIBUTION/AVAILABILITY STA]rEMENT 12b. DISTRIBUTION CODE Unclassified--Unlimited Subject Category 08 13. ABSTRACT (Maximum 200 words) A flight research study of high-angle-of-attack handling qualities has been conducted at the NASA Dryden Flight Research Center using the F/A-18 High Alpha Research Vehicle (HARV). The objectives were to create a high-angle-of- attack handling qualities flight database, develop appropriate research evaluation maneuvers, and evaluate high-angle-of- attack handling qualities guidelines and criteria. Using linear and nonlinear simulations and flight research data, the predictions from each criterion were compared with the pilot ratings and comments. Proposed high-angle-of-attack nonlinear design guidelines and proposed handling qualities criteria and guidelines developed using piloted simulation were considered. Recently formulated time-domain Neal-Smith guidelines were also considered for application to high- angle-of-attack maneuvering. Conventional envelope criteria were evaluated for possible extension to the high-angle-of- attack regime. Additionally, the maneuvers were studied as potential evaluation techniques, including a limited validation of the proposed standard evaluation maneuver set. This paper gives an overview of these research objectives through examples and summarizes result highlights. The maneuver development is described briefly, the criteria evaluation is emphasized with example results given, and a brief discussion of the database form and content is presented.

14. SUBJECT TERMS lS. NUMBER OF PAGES Database, Criteria guidelines, Handling qualities, High-angle-of-attack, Maneuvers 16. PRICE CODE AO36 17. SECURITY CLASSIFICATION 18. SECURITY CLASSIFICATION 19. SECURITY CLASSIRCATION 20. LIMITATION OF ABSTRACT OF REPORT OF THIS PAGE OF ABSTRACT Unclassified Unclassified Unclassified Unlimited NSN 7540-01-280-5500 Available from the NASA Center for AeroSpace Information, 800 Elkndge Landing Road, Standard Form 298 (Ray. 2-89) PreQcribed by ANSI Std Z3LF18 Linthicum Heights, MD 21090; (301)621-0390 298-102