An Investigation of the Use of Bandwidth Criteria for Rotorcraft Handling-Qualities Specifications

,.",.,.,....,.~~-~~'~- .... ~~LH_n"~·~·~'··~OT'Cn.H._.4 .. _.~~~~--~' ~'-.' .' .~,

-

OCT 2 3 1985

AD-A160 664

A

AN INVESTIGATION OF THE USE OF BANDWIDTH CRITERIA FOR ROTORCRAFT HANDLING-QUALITIES SPECIFICATIONS Chris L. Blanken, C~urtland C. Bivens, and ~atthew S. Whalley Aeromechanics Laboratory U.S. Army Research and Technology Laboratories (AVSCOM) NASA Ames Research Center Moffett Fleld, California 94035 U.S. military rotorcraft-handllng-qualities speCi- fication, MIL-H-8501A (Ref. 2), Is a 1961 revision

'/

The objective of this study was to investi- of a 1952 document. This specification contains gate bandwidth concepts for deriving rotorcraft many criteria that are inadequate f'or design guid- handling-qualities criteria from data obtained in ance or flight testing. New or up~ated design two simulator experiments conducted at the Aero- criteria sho~ld be developed and substantiated to mechanics Laboratory. The first experiment wab an provide data ror the new specification. Theue investigation of the effects of helicopter revised criteria must account for the numerous ve~tical-thrust-response characteristics or. hand- demands on the pilot of an advanced military ling qualities; the second experiment Investigated rotorcraft. These demands will vary, depending on the effects of helicopter yaw-control-response the mission and task and on the environment in characteristiCS. In both experiments, emphasis which they must be flown. For example, a pilot was on low-speed nap-of-the-Earth (NOE) tasks. Of", flyi ng an at tack hell copter nap"of -the-Earth (NOE) The results from the thrust-response simulation at night under adverse weather conditions will be indicate the open-loop vertical velocity to col- : subject to different demands, which Imply differ- lective h/6 bandwidth is greatly influenced.

ent aircraft design requirements, than a pilot c by vertical damping. For the task investigated, flying a cargo helicopter I.n clear day conditions.

Level I handling qualities may require an open To aid in providing dat~ for miss!on-oriented loop bandwidth greater than approximately -1 handllng-qualitles cr'lterla, the analysis and 0.5 rad/sec for vertical damping of -0.25 sec correlation aSSOCiated with a propos~d design and approximately 0.75 rad/sec for damping of criteria. called bandwidth, was applied to the -0.65 sec-I. These results Imply that for tI.e results of two helicopter simulations. ihese thrust response, criteria :ased on the open-loop piloted Simulations, conducted by the U.S. Army h/6 bandwidth are nol sufficient to ensure good A~.·omecha".lcs Laboratory at NASA Ames Reaearch handYing qualities. The results from the yaw- Center, invest 19ated the effects of hel! copter response simulation indicate that an open-loop thrust- and yaw-response characteristiCS on hand- bandwidth of at least 2.5 rad/~ec is required ~or ling qualities fOI" NOE flight tasks.

the decel~ration task, that a bandwidth of at least 3 rad/sec is requ.red for the NOE and hover The fol:owlng sections describe the bandwidth turn tasks, and that a ~undwidth of at least concepts as applied In open- and closed-loop b~tween 2.5 and 4 rad/!ec is required for the air- analyses; the piloted Simulations, including the to-air target acquisition task. Yaw-response conduct and variables or the thrust- and yaw- "losed-Ioop bandwidt.h analysis showed a high response simulations; the results of applying COl relation with the open-loop analysis and may the bandwidth analysis; and the conclusions and be useful in predicting the relative merits of a recommendations.

configuration before going to a piloted simulation.

Bandwidth Concepts Bandwlath is a ~ualltatlve measure of the Introduction Input-to-output response of a dynamic system.

There Is a major effort under w~i ~, revise Since it Is a me&sure of the system Input-to- and update the general speclrlcatlon for, anc:'.l1ng output response, multi-parameter changes within qualities of military rotorcral't.' The CUI-~~.v the system should be captured. This phenomenon makes bandwidth an attractive criterion. Band" width analySiS is conducted In the frequency domain and results in a fundamental measure of the Presented at the International Conference on ability of the system output to follow the system Rotorcraft Basic Research. American Helicopter Input. A higher system bandwidttl reflects a SOCiety, Research Triangle Park, North Carolina, raster and ~ore predictable alr~raft response to Feo. 19-21, 1985.

control i nput.s. The input and output ql.!snt i ties selected to define the system bandwidth are those _.

This doc:ument has been approved

10 IS 01S

f~l ~ubl~c release and sale; its

/

dlstnbuhon is unlimited.

mos' appropriate to the task being evaluated; for time delay. This model represents a "comfortable" eXo"'7Ile, heading regulation involvAs rudder pedals pilot who is not providing any lead or lag as the input snd yaw angle as the output. The compensatior, for def iciencies in the vehi cle bandwidth hypothesis originated from the idea dynamics. The objective of this analysiS is to that the pilot's eval~ation of aircraft handling determine the mRximum obtainable bandwidth. The qualities is dominated by the response character- procedure for the pilot-in-the-loop anulysls Is ~s ist:os of the aircraft when it is operated in a follows: clos~d-loop tracking task. That is, the pilot's capability to make rapid and precise control 1) After obtaIning the characteristic trans- inputs to minimize errors and thereby improve fer function for the aircraft, for example, ~/6p' closed-loop tracking performance dominates his it is combined with the pilot model (Fig. 3). The evaluation. assumed form of the pilot's transfer funcCion was tS P(s) • K e- , where K is the pilot gaIn, l The classioal definition of closed-loop ~and is the r*action time del~y, and s is the ~aplace TS width is the frequency at which the Bode ampli- operator. For e- , the Pade approximation was tude is 3 dB less than the steady-state amplitude used with the initial value of l set to 0.3 sec, of the system (see Fig. 1). Note that for a representative of the human neuromuscular time K/s aircraft response characteristio, the delay.

bandwidth frequenoy (wnw) and the inverse of the system time-constant (lIT) are identical. 2) The pilot gain K~ was adjusted to achieve a gain of 0 dB at the specified crossover Open-~oop Bandwidth frequency. For the yaw response, the crossover frequency wa, 3 radlsec (Fig. 4). The system was Since "" '')en-loop crossover frequency is then checked for stability, that is, the phase equal to (" ~ r higr.-order systems appr~xi- margin was t300, and the gain margin was t4 dB.

matelyequal:. the classical closed-loQP band- width, the definition of bandwidth and crossover 3) The loop was closed around the system in frequency are equivalent. That is, the open-loop step 2, a~d the gain K~/K. and an integrator bandwidth is defined (from Ref. 3) as the cross- ~ over frequency for a simple, pure gain pilot with were combined (Fig. 5). The K~/K. gain was a ~5° phase margin or a 6 dB gain marg!n, which- ~ ever frequency is lower. For example, the fre- adjusted to achieve a maximum gain and phase quency for neutral stability, ~S' is observed on margin of 4 dB and 30°, respectively (Fig. 6).

the phase curve of a Bode plot. Note that typi- 4) The Quter-loop was finally closed around cally the output quantities selected to define the the system in step 3 to obtain the system shown in system bandwidth, as related to a vertical-axis Fig. 7. A Bode plot was obtained of this closed- and yaw-axis traCking task, are altitude and yaw loop sy~ .em. and the banClwidth was determined.

angle. For convenient application of the olosed- For the yaw ~esponse e~periment, the closed-loop loop bandwtdth analysis, the open-loop bandwidth bandwidth W~j defined as the frequency at which criterion is ~pplied to both translational and there was either a 3 dB amplitude ratio change or angular rate control responses. Since these out- a 90° phase change, whichever was less (Fig. B).

put responses are one integration away from alti- tude and yaw angle, the margins are measured using 90 of phase angle as a reference (Fig. 2). The 0 Piloted Simulation Studies frequency for which a 45 phase margin exists is defln~d as w.. The frequency correspondln!!

The bandwidth concepts were applied to two oWpHASE' helicopter simulator experiments deSigned and to an amplitude ratio that Is 6 dP less than the performed by the U.S. Army Aeromechanics ~abora amplitude ratio at neutral stability is denoted as tory at Ames Research Center. These piloted, W The bandwidth Wau Is the lessor of ground-based simulations were conducted on the BWGAIN "" Ames V~rtical Motior Simulator (VMS) (Fig. 9).

the two frequencies, w..w and w..w • The simulator cab was conri~ured to include a o PHASE 0 GAIN typical helicopter instrument panel and control- Closed-~OOp Bandwidth lers. The visual display consisted of a computer- generated image (CGI) scene p"esented on four A closed-loop handwldth analysis using a windows, furnishing the pilot with a 28° by 120° simplified pilot model was also investigated using field of view above the instrument panel; in addi- the techniques defined In Ref. 4. The Intent of tion, there was a 29° by 40° right-hand chin- this analysis was to take Into account the closed- window scene.

loop nat~e of the track1ng task conducted by a human pilot represented by a gain and an effective

Thrust-Response Slm~latlon ways: Increases In goy·?rnor time constant can significantly degrade the handling-qualities The thrust-response simulation studies we;'e rating. but increasee In rotor inertia have performed on the VMS using a ten-degree-of- only a minor and desirable effect on handling freedom, nonlinear, full-force mathematical mode'.

qualities. These two parameters must. there- termed ARMCOP (Ref. 5). Aural cueing ot the rotor fore. be treated Independently in handling speed (rpm) fluctuations and blade slap, a visual qualities requirements.

display of rotor speed, and an overs peed and Jnderspeed warning light were provided to the It was this conclusion that prompted Interest In pilot. The evaluation task (Fig. 10) consisted the bandwidth hypothesis with the hope that these of two phases: 1) a 40-knot dolphin, or hedge- two opposing etfects would collapse into one so hopping, phase, and 2) a qUlck-stop and bot- that a unique parameter (bandwidth) could be used up/bob~down phase. Three Army and three NASA test to charactel'l ze good handl1 ng quall ties.

pilots served as evaluation PllotS' The pilots used the Cooper-Harper Rating Scale to assess the Yaw-Response Simulation effects of height (or flight path) control and rotor speed control on handling qualities. Each The yaw-control simulation studies were con- phas~ of th& evaluation task was rated separately.

ducteg using a small-perturbation helicopter model which has the full nonlinear set of The Cooper-Harper (C-H) pUot I'ating scale Is kinematic terms In seven-degree-of-freedom (DOF) structured Into three distinct groups or levels.

equations of motion. (A rotor speed DOF was Level 1 corresporlds to C-H pUot ratings of " 2, Included.) The evaluation task (Fig. 11) con- and 3; Level 2 c~rresponds to C-H pilot ratings of sleted of NOE flight and deceleration, low- and 4, 5, and 6; Level 3 to C-H pilot ratings of 7. 8, high-maneuvering turns, and an al~-to-alr target- and 9. Levell handling qualities are clearly Bcqulsl tlon arid tracking task. Four test pilots

adequate for the mission task; Level a handling

served as evaluation pilots. The pilots used the qualitlds are adequate to accomplish the mission C-H rating scale to assess the effects of direc- task, but some Increase In pilot workload or tional rate damping N • directional control r degradation In miSSion effectiveness exists; and sensitivity N , and weathercock stability N ' v Level 3 handling qualities are such that the air- p craft can be controlled safely, but pilot workload Each phase of the evaluation task 14as rated sepa- Is excessive or miSSion effectiveness is Inade- rately. Details of these configurations along quate or both.

with pilot ~atlng results and conclUSions are given In Ref. 9. Three pOints were noted In the The primary variables In this study were evaluations: those which atfect the power-system response time. They were the engine-governor response time 1) Higher values of directional gust senSi- and the rotor Inertia. Other variables were heli- tivity N require greater values of yaw damping v copter vertical damping, collective control sensi- N • to achieve satisfactory handling qualities for r tivity, excess power available. and the require- NOE flight. deceleration, and hover turns.

ment that the pilot maintain rotor speed within specified limits. The variations In englne- 2) Performance measures. when used alone. can governor response, rotor inertia. and vertical give misleading Information regarding aircraft damping provided the basis for the bandwidth handling qualities.

analysis. For the purposes of this paper the 3) N Is not a sufficient parameter for engine-governor response character~stlcs are cate- r gor\zed as slow, Intermediate, and fast. Like- fully defining acceptable handling qualities for wise. the values ot rotor Inertia are categorlz~d the air-to-air target-acquisition taskl control- as light. medium. and heavy. The two values ot response criteria are also needed. especially for vertical damping Zw were Investigated: SCAS configurations.

-J.25 sec- and -0.65 sec-I.

The bandwidth analysis was applied to the various Details of these confl~uratlons along with yaw-response configurations to determine if band- C-H pilot rating results and conclusions tor all width would capture these multlp~rameter effects the variables may be found In Ref. 7. The follow- 1n a slngle parameter.

Ing Is a conclusion from Ref. 7: Results of Applying Bandwldth Concepts to Increasing rotor Inertia and engine-governor Simulation Studies time constant will 1ecrease power-system natu- The results of applying the bandwidth ral frequ lCY, but for the simulated tasks.

analysis to the thrust-response and a yaw-response will affect handling qualities In d1fferent

simulation results are discussed. The correlation value of w W is based on the medium rotor of ban~wldth with the pllot ratings Is presented inertia. T~e larger collecti"e control inputs together with recommendations for millir1um values

(16 I . to.5 and tl.0) Indicate an h/6

of system bandwidth to meet handl1ng-qllal1ties banSwidth greater than about 0.4 rad/seg as neces- requirements for the tasks investigate/I.

sary for ~evel 1 pilot ratings. It is felt that the larger collective control Inputs may be more Thrust-Response SimulaU(~ representative for correlstion based on the magni- tude of the actual pilots' control Inputs requl.-ed To assist in charactel'lzing the various con- for this task. Note that the open-loop bandWidths figurations evaluated by the p~lots, frp.quency- for the configurations with 16cl· t1.0, are very response data were collected. This analysis con- Similar, and use of the bandwidth as a discrimina- sisted of 1) applying a single frequency, sin'~oi tor for levels of handling qualities seems mar- dal collective control Input, 6 ' to the simulated c ginal. Frequency-response data for collective aircraft: 2) recording this input and selected input amplitudes greater than tl.0 in. are neces- nnicle output states, such as vertical velocity sary before final analysis and recommendations can

h: 3) repeating this procedure with several fre-

be made.

q·;encies, between 0.1 and 10.0 rad/sec: 4) measur- .ng the amplitude ratiO and phase-angle shift Figure 14 also shows the effect of vertical associated with each frequency: and 5) plotting damping Zw' This figure shows the open-loop these data onto a Bode plot. The above procedure bandwidth for two different values of Zw versus was repeated for three different collective input the averaged C-H pilot ratings from the bob-up amplitudes, 16cl • to.2, to.5, and tl.0 In. to tasu. Note that for a medium rotor inertia, the account for possible nonlinear eff~cts. USing data point~ may again be bandeu, but into two these frequency-response data obtained from the di"tinot groups associated with the two values of various thrust-response configurations, an c,pen- vel"tlcal damping Zw· -0.25 sec- and loop bandwidth analysis was conducted. The band- -0.65 sec-I, The data points within these bands width analysis Is discussed with regard to the Include the variation in the engine-governor effects of 1) the variations In engine-governor response from fast to slow and v~riation In the response and rotor Inertia, 2) the variation In amplitude of the collective control input. Note the amplitude of the collective control Input, and that the vertical damping appears to have a 3) the variation in vertical damping.

greater' effect on the bandwidth requirement than does the engine-governor response or collective Figure 12 shows a typical Bode plot for a Input amplitude. Also note that there are con- fast- and a slow-responding engine-governor with figurations on the ~evel 2-3 borderline medium rotor Inertia and 16cl· iO.2. The open- -1 (Zw • -0.65 sec ) with 1 higher bandwidth loop bandwidth analYSis was applied to these and than some of t~r ~evel 1 configurations with similar Bode plots for the thrust-response con- Zw • -0.25 sec These results Imply that for figurations evaluated.

:he thr'ust response, criteria based on open-loop Figul'e 13 shows the open-loop bandwidth w h/6 bandwidth are not !~rflclent to ensure good BW c for the three engine-governor responses and three handling qualities. In particular, some reqUire- rotor inertias versus the averaged C-H pilot rat- ments based on Zw and (perhaps) engine-governor ings from the bob-up task. Note that the value of cha~acterlstlcs are required.

w was derived based on a collective in~ut BW In summary, the results of the thrust- amplitude of to.2 in. and Zw· 0.25 sec-, The response Simulation indicate that the open-loop data points may be banded to capture the variation bandwidth for ~evel I handling qualities showld of the engine-governor response ranging from fast be greater than about 0.5 rad/sec for to slow and the variation of the rotor inertia -1 Zw • -0.25 sec and 0.75 rad/sec for from light to heavy. The figure Indicates .nat an -1 Zw • -0.65 sec The effects of vertical h/6 bandwidth greater than about 0.5 rad/sec Is c damping em the open-loop :>andwidth should be necessary to ensure ~evel 1 handling qualities.

further investigated and analyzed.

Review of the pilot collective control The closed-loop analysis for the thruet- Input magnitudes during the initiation of the bob- response Simulation Is not discussed In this up maneuver showed ranges from 0.5 to 2 In., paper because of the need for additional depending on the pilot'S aggreSSiveness In frequency-response data and analysis In order to performing the maneuver. The open data symbols interpret further the existing results.

(Zw • -0.25 sec-I) In Fig. 14 show the open-loop bandwidth for the three different collective contr~l input magnitudes versus the averaged C-H pilot ratings from the bob-up task where the

Yaw-Response Simulation overshooting and under'shooting the target," "unable to hold the pipper steady on target," To cha~acte~lze the configurations evaluated and "excessive amount of time to decrease the by the pilot in the yaw-response simulation, an ,tracklng error." The pUot oomments for idealized headlng-rate-to-pedal control-input configuratlons with bandwldths between 2.5 and II rad/sec include "easy to generate a rlilpi"

~~:~:~:~ ~~~~~~~~ B!~!P~l:~: ::~~~~~ai~:~mf~~is

yaw rate, acqulre, and track target." An open-loop and closed-loop analyses, using the exceptlon to this occurs for the configuration matrix of the experimental variables that were with an open-loop bandwidth of II rad/sec and an evaluated. An idealized form of this transfer average C-H pilot rating of about 5.5. For this function may be assumed with good confidence since oase, even though the bandwidth was between 2.5 the mathematical helicopter mode1 used for these <Ind II rad/sec, it received a degraded pUot rBting studies was a small-perturbation model utilizing because the control-response parameter was set at stability derivatives, which are functions of c'ne-half of the nominal value, that Is, it was set velocity.

alt 3.5 dee.sec- In. -1 of pedal compared with 1 1 7'.5 deg.sec- in.- of pedal nominally.

The open-loop system block diagram lncludlng the assumed form of the ~/~ transfer function In summary, based on the configurations and whe,'e Y • ~/6 , is shown lR Flg. 15. The c t~asks evaluated in the yaw-response simulat;1on, closed-loop sys~em bloCk diagram Is shown In '~he open-loop bandwidth for the best C-H pilot Fig. 7. The experimental matrix showing the ,"atings show the following: primary configurations that were evaluated 15 show~ in Fig. 16. A linear analYSIS computer Open-Loop Bandwidth 10 .!M!i program was used to obtain the open-loop Bode NOE >3 rad/sec plots and to perform the closed-loop analYSIS and Deceleration >2.5 rad/sec subs~quent Bode plots.

Low hover turn >3.0 rad/sec Air-to-air 2.5 rad/seo S Figure 17 shows an example of the open- and ~BW target ~quisltlon SII rad/sec closed-loop Bode plot, the corresponding band- widths, and the averaged C·'H pilot ratl.1gs for the The closed-loop bandwidth analYSiS, whioh was tasks evaluated. Figure 18 shows the open-loop preViously described, was applied to the yaw- heading rate bandwidths Wsw for the experimental response configurations and correlated with the matrix of variables evaluated versus the averaged C-H pilot ratings. Flgure 19 shows the closed- C-H pilot ratings for the NOE task, the decelera- loop bandwidth versus C-H pilot ratings ror the tl on task, the low-haver-turn tasl(, and the al r- NOE task, the decelerat!on task, the low-hover to-air target-acqUisition taSk. The hlgh-hover- task, and the air-to-air target-acquisition task.

turn was omitted here because of the Similarity of There appears to be a moderate oorrelation between those data and the low-haver-turn data.

the C-H pilot ratings and the olosed-loop band- For the NOE task, open-loop bandwidths width. The NOE and deceleration task indicate greater than about 3.0 rad/sec result in conslder- that closed-loop bandwidths greater than about 3.6 rad/sec are necessary for Level 1 handling able improvement in the pilot ratings. Also, for qualities. Although Levell ratlngs were not the NOE task there is a relatlvely high linear correlation between Increased open-loop bandwidth attained for the low-hover-turn and air-to-air l and lm~roved C-H pilot ratings. The deceleration target-acquisi tlOtl tasks, the best rat 1 ngf for and hover turn tasks have some linear correlatlon, these tasks correspond to closed-loop bandwidths but the overall trend appears nonllnear and task- greater than or equal to 3.0 rad/seo.

dependent. The alr-to-alr task has a nonlinear A linear correlatlon analysis was performed trend of bandwidth versus C-H pUot ratings. Thls on the open- versus closed-loop bandwidth data for trend implles that the open-loop bandwidths Wsw eaoh configuratlon. Figure 20 shows that the correspondlng to the best pilot ratings for the correlation was extremely h1gh, thus ind1cattng task are 2.5 rad/sec $ ~BW $ 4.0 rad/sec. Pllot that either analysis may be used with a linearized ccmments support thls trend.

simulation model.

For the conflgurations with bandwldth greater Finally, an investlgatlon was made lnt9 the than 4.0 rad/sec, pilot comments include the use of a slmple p1lot model as a predictive tool rollowing: "the aircraft was slightly slugglsh," for ya\l-control handllng-qualltles research.

nthe tendency for overShoot," and "yaw alrcraft The pilot ga1n result1ng from the closed-loop fl control took moderate pl10t compensatlon. For bandw1dth analys1s was correlated w1th the the configuratlons wlth bandwldths less than C-H pilot ratinv,9 for the NOE task (see 2.5 rad/sec, pl10ts commented about "contlnuously Fig. 21). The correlation Indicates that if an

alroraft system and pilot model produoe a pilot 1) The tl dnds of the effects of variations in gain greater than 20 In. rad- then the handling directional rat& J~~p!ng and weatheroock stability qualities will be relatively good for that task.

on handling qualities can be predloted by an open- To oonflrm the val1dlty of this approaoh, a oon" loop (~/6p) bandwidth analysis.

figuration not evaluated during the simulation but known to be a bad oonflguratlon from the Initial 2) For the oonflguratlons evaluated the fol- Simulation oheokout phase was analyzed. Closed- lowing open-loop bandwidths provide the best hand- 11ng qual1tIes: gr·eater than 2.5 rad/seo for the loop bandWidth analysis of this oonflguratlon produoed a pilot gain of 13.86 In. rad- Com- deoeleratlon task, greater than 3 rad/seo for the NOE and hover tasks, and between 2.5 and 4 rad/seo paring this pilot gain with the results presented In Fig. 21 shows this oonflguratlon to yield pre- for the air-to-air target-aoqulsltlon task.

dloted handling qualities In the Level 2 region 3) Yaw-response olosed-loop bandwidth analy- (C-H pilot rating of 6). "hls analYSis providE'S a sis results showed a high oorrelation with thos~ preliminary oonflrmation of the predlotlve cap- of the open-loop analysis and may be useful In ability of the olosed-loop bandwidth analysis.

prediotlng the handling quallt"es of a parttoular configuration for, a speoiflc flight task.

Conclusions and Reoommendatlons The ooncepts of bandwidth, open- and olosed- Referenoes loop, were evaluated for deriving handllng- Qualities orlterla from data obtained during two 1 Key , D. 1,., "The Status of Military Helloopter helloopter simulator experiments. The first Handling-Qualities Criteria," Paper No. 11, Con- experiment, a thrust-response Simulation, was an ferenoe Proceedings CP-333, the AGARD FMP Speoial- Investigation of handllng-quallt:es effects of Ists Meeting on Criteria for Handling Qual1t1es of englne-governo~ response time, rotor Inertia, Military Alroraft, Ft. Worth, Texas, Apr. 1982.

vertloal damplng, oollectlve control sensitivity, exoess power, and rotor speed control. Applloa-. 2"Helloopter flying and Ground Handling Qualit1es, tion of a bandwidth analysiS to these simulation General ReQuirements for," MIL-H-8501A, Sept. 7, data Indicates the following. 1961.

1) Tr,.e variations In engine-governor response 3 Hoh , R. H., Myers, T. T., Ashkenas. I. L., times and rotor Inertia and their effects on hand- Ringland, R. F., and Craig, S. J •• , "Development ling qua'.ltles oan be oaptured by an open-loJp of Handl1ng Quality Criteria for Alroraft with bandwidth criterion. But the bandwidth criterion Independent Control of Six Degrees of Freedom," must be aocompanl ed by requl rements for the all'- AFWAL-TR-81-3027, Systems Technology, Ino., oraft vertical damping. For a bob-up task, with Hawthorne, Calif., Apr. 1981.

vertical damping of -0.25 sec- , the open- loop (h/6 ) bandwidth to ensure Level 1 handling 4Lebaoqz, J. V. and Aiken, E. W., "A Flight o qualities must be greater than aprroxlmately Investigation of Control, Display and Guldanoe 0.5 rad/sec. With a vertical damping of Requirements for Deoeleratlng Desoendlng VTOL -0.65 sec- , the open-loop bandwidth to ensure Instrument Transitions Using the X-22A Variable Levell handling qualities must be greater than Stability Airoraft," Cal span Report No. AK-5336- approximately 0.75 rad/sec. The effects of verti- F-1, Buffalo, New York, Sept. 1975.

cal damping on the open··loop h/6 bandwidth c needs further Investigation and analysis.

5Talbot, P. D., Tlnlln~, B. E., Deoker, W. A., and Chen, R. T. N., "A Mathematloal Model of a Single 2) Sinoe the frequency-response data Main Rotor HeJ.1oopter for Piloted Simulation," (h/6 ) were gathered for oollectlva control Inputs NASA TM-81203, 1982.

equa£ to or less than %1.0 In., additional fre- quency-response data with larger oollective con- 6cooper, G. E. and Harp~r, R. P., "The Use of trol Inputs will alsJ have to be investigated Pilot Rating In the Evaluation of Aircraft Hand- before a final bandwidth criterion recommendation ling Qualities," NASA TN 0-5153, 1969.

can be made.

7Blanken, C. L. and CorliSS, L. D., "Thrust The second experiment, a yaw-response Simu- Dynamlos and Helicopter Handl1ng Qualities," Army lation, was an Investigation of the handllng- Solence Conference Proceedings, Vol. 1, p. 39, qualities effects of directional rate damping, U.S. Ml1ltary Academy, West POint, New York, directional control Bensltlvlty, a~d weatheroook June 1984.

stability. Application of a bandwidth analysis to these simulator data Indicates the following.

8 AiI (en, E. W., "A Mathematioal Representation of 9Divens, C. C., "Direotional Handling Qualities an Advanced Helicopt~r for Piloted Simulato~ Requirements for Nap-of-the-Eerth Tasks," to ba Investigations of Control System and Display presented at the AHS Forum, Ft. Worth, Texas, Variations," NASA TM-8'1203, 1980.

May 15-11. 1985.

Shah , 10 S. C., Walker, R. A., and Gregory, C. Z., "MATRIX-X User's Guide, V4.0," Integrated Syatems, Ino., Palo Alto, CaUf'., 1985.

PURE GAIN PILOT AIRFRAME

o

I ·

tru

OPEN LOOP -: ~ .

AIRFI1AME.O/6

-12,/

1~ldB

-18 wsw 3 dB PURE GAIN

____ ...,...;:~I ;._~.i~_ - -POLOT CLOSURE

idB -24 ./ GAIN CLOSED lOOP RESPONSE - O/IJ c log w. rid/sec 41. dill 0 -20 -60 -

~~-~f--m-~~roop

~4I~M~·~4=50~~~~==z=~~==~

-100~ .1 ,10

~-~~-----------.-------~~-~

wNS FREQUENCV DOMAON w. rad/sec Fig. 1 Classical definition of bandwidth from Fig. 2 Definition of olosed-loop bandwidth, IUBW (from Ref. 3). (wsw· lesser of w or w ).

B BWpHASE W GAIN AlC CHARACTEROSTIC POILOT TRANSFER

:;.- :;:: o;)""For--

MOOED. IFUNCTOON r r':I-;C:-CRA&I---Jf-'--;

-+.·-~H f.lm f-

\ D1IC TAB 0

U:lannounced 0

J .J 5 t it I cat i 0 n ......... _ .. _ .. _ .•• _._.

\ ~

(PAIDE APPROXOMATOO" B _------t

y ......... _ ... _.

Llistribution' Fig. 3 Pilot model and aircraft transfer function.

Availability Coc'es 1 -~~

t

o '--------....;~-~ W, rad/sec

GAIN MARGIN, 4 dB w, ,ad/lic Flg. 4 Crossover frequency. Flg. 5 Intermediate step in closed-loop anR':~ls combines K~/K~ gain and integrator.

AIC CHARACTERISTIC TRANSFER PILOT FUNCTOON INTEGRATOR GAIN , t/i

~

i

O<~ Fig. 6 Morgins for K~/K~ gain adjustment.

AIC CHARACTIEROSTIC TRANSIFER PO LOY FUNCTOON INTEGRATOR MODEL 1 '"

S

Fig. 7 Complete closed-loop model.

~, 3--------

~Q' ::;t:

.~ O~----~--~~----~--------~w

-3

o~--~~-----------~~--------W

-180 I~ THlfiI.ASE: wBANDW',"JTH· wBO BECAUSE "'60 < wBP Fig. 8 Definition of olosed-loop bandwidth (Ref. 4).

Fig. 9 Vertioal Motion Simulator (VMS).

40 knot QUICK STOP

DOLPHIN soaup AND DOWN

_~ __________ ~A .~

~ /-'\ /'-\ ,/'\ /-\ t""

,,- .... ;/\ \ .. ---) A .......... ,,1\ \ ........ ,/\ ..........• , 1-1oooft"""1 I I I Fig. 10 Thrust-response simulation task.

• NAP·OF·EARTH (NOEl FLIGHT WIND • NOE DECELERATION TO A HOVER / DIRECTION • IN GROUND EFFECT HOVER • OUT OF GROUND EFFECT HOVER VEL 18-26 knots • AIR TARGET ACQUISITION AND 'rRACKIIIIG GUST u. v c 6 knots RMS w -4 knots RMS RANGE: '000 ft SPEED: 60 knots HEiGHT: 100 ft AGL Fig. 11 Yaw-respons~ simulation task.

MAGNITUDE :10

11.1i 15

~10

I

':!d

-40 ~ 6 0 -80

I

!

-6 -80 -10 -100 MEDIUM ROTOR INERTIA IIlloll • to.2, 'w. -0.25110- -15 -120 -140 -20 .1 1.0 10.0 FREQUENCV, rld/IIC Fig. 12 Typical thrust response, hIll Bode plot.

c aOB·UP TASI( VERTICAL DAMPING 'w.IIC-' 1I0cil • to.2In.

o 0 V -0.25

ROTOA INERTIA MEDOUM ROTOR INERTIA • • .. -u.S5 o LIGHT COLLECTIVE F.i\lGINE GOVERNOR RESPONSE

o MEDIU',I AMPLITUDE. 1I0cil

!= - FAST, • -INTERMEDIATE. S - SLOW A HEA\. V

o to.2

C to.5 9 V t1.0 C.!I LEVEL i= LEVEL <I': /I:

S

~ it 2 2 1: ,U II: I.

.(

~

!AI ~ °.1 2 ~ A ~ ~ ~ .fJ .1 2 ~ A ~ ~ ~ .8 OPEN·LOOP BANDWODTH, waw, rad/pc OPEN· LOOP BANDWODTH, wBW' rad/IIC Fig. 13 Correlation of pilot rating with thrust- Fig. 14 Correlation of pilot rating with thrust- response open-loop bandwidth; effects of rotor response open-loop bandwidth; effects of input inertia and engine-governor response.

amplitude, vertlcal damplng and englne-governor response.

Il Yp

Yo }--J

--1 ·1

.1\~q(,RAFT PILOT Yp. ;~o· Nil Kp .'

p opeN LOOP 2tw· N, SYSTeM K,S Yo • 1IJ2. UoNyCOItjlo S2 + 2t<.lS+ w Fig. 15 Yaw-response block diagram for open-loop analyeie.

PILOT RATINGS 0.76S NOE 6.76 ~ .

2 DeCEL 4.76

6 p 8 + 8 + 0.353& LOW HOVER 6.75 Am·TO·AIR TARGET ACQUISITION TASk 7.00 -6 -0.6 -1.0 -4.0 -6.0 -12 0.6 0.76 1.00 1.66

~ldB

-18 wBW· 1.26 rlld/llc -24 , 41, deg -20 -60 -100 2& t:, CLOSED LOOP 1& &

~dB

-6

J •. -

-16

wew " 2.40 rld/llc

~--""~~'f-

-25 1&0 2 2

~

4>,dlg

N = !!,.d/1IC 0

N"~

lip In.

V ft/~IC -30 -90 ~ PRIMARY CONFIGURATIONS -150 , R -YAWSCAS .1 R1 • RATE COMMAND HEADING HOL.D FREQUENCY, rld/IIC Fig. 16 Yaw-response variables matrix. Fig. 17 Ty!>ical yaw-response open- and closed- loop Bode pInt.

NOE TASK DECELERATION TASK ~ 9 • NO VAW·TO·COLLECTIVE o VAW·TO·COLLECTIVE LEVEL LEVEL COUPLING ~ 8 COUPLING 3 3 II:

§ 7

o--T---------------

it 6

ffi 6

o

~ 4

~ 3 CCRRELATION • -ClI.616 II: LINEAR w 2 m--O.240 CORRELATICN b· 4.8Cl1 CJ 1 011..._ ....... _...", __ 1_._..----1.

.....

1.01/1; Itll.l~"~1R TURN TASK AnR·TC·ADR TARGET ACaUISIT80N TASK LEVEL LEVEL 3 3

-e- - -- - - - ------ ----~-

2' LnNEAR 2 CORRELATnON '" -0.678 CORRELATION· -ClI.389 CORRIEO.ATnON moo -ClI.258 nil - -0.180 Ii '" 6.22 11». 5.45 '/.

2 3 4 8 10 2 3 4 5 6 8 10 OPEN·LOOP BANDWIDTH, rllCll/HC OPEN·LOOP BANDWDDTIHI, rllCll/sec Fig. 18 Correlation of pIlot rating with yaw-response open-loop bandwidth.

DECELERATION TASK NOE TASK LEVEL LEVEL :! 3

--------~----------

2 2 o ___________ ,1._4 CORRELATION· -0.009 m· -0.817 b· 6.63 LOW HOVER TURN TASK AIR·TO·AIR lARGEr ACQUISITION TASK ~ 9 LEVEL LEVEL

~ 8

3 3 a: "I o

_ .. _-----------------

§

--------0----------

o Ii: o ('.) o a: o III 2 .0 o 0 ___________ ~A_~- __ 0(1) 00 &: <t

-------------------

% CORRELATDON - -0.670 CORRELATION· -0.666 Ii: III 2 m- -0.626 m. -0.741 b· 7.12 b - 8.38

§

, , , , d~--~--~--~----~--~--~--~~~ 1.0 1.4 1.8 2.2 2.6 3.0 3.4 3.8 4.2 1.0 1.4 1.8 2.2 2.11 3.0 3.4 3.8 4.2 CLOSED-lOOP BANDWIDTH, ,ed/. CLOSED· LOOP BA'AIDWIDTH, red/.

Fig. 19 Correlation of pilot rating with yaw-response closed-loop bandwidth.

•

LEVEL 3 o

--~r--------------

lib 2 a

------------~-~--

CORRELATION. -0.799 12 14 11 " 20 22 24 CLOSEO·LOOP PILOT GAIN. Kw CORAELA TION • 0.815 Fig. 21 Correlat10n or pilot rating with yaw- response cl~sed-loop p1lot ~aln.

, _~._...&.t- __ , __ ='

o 2.0 2.6 3.0 3.& 4.0 CLOSEO LOOP BANDWIDTH. raJ/ac Fig. 20 Correlation or yaw-respunse opon- and ~losed-loop bandwiuth.