https://ntrs.nasa.gov/search.jsp?R=19700029309 2018-12-14T09:16:48+00:00Z c FLIGHT INVESTIGATION OF THE ROLL REQUIREMENTS FOR TRANSPORT AIRPLANES IN CRUISING FLIGHT by Ezcclid C. Holleman Flight Research Cmter Edwards, Cal$ 93523 NATIONAL AERONAUTICS AND SPACE ADMINISTRATION WASHINGTON, D. C. SEPTEMBER 1970 TECH LIBRARY KAFB, NM 1I111111 Ul1l1l11 lllll IIIII lllll lllll H1I 1111 0332839 I- I 1. Report No. 2. Government Accession No. 3. Recipient's Catalog No. NASA TN D-5957 .~ 4. Title and Subtitle 5. Report Date FLIGHT INVESTIGATION OF THE ROLL REQUIREMENTS FOR TRANS- September 1970 PORT AIRPLANES IN CRUISING FLIGHT 6. Performing Organization Code 7. Author(s) 8. Performing Organization Report No. H-616 Euclid C. Holleman .. 10. Work Unit No. 9. Performing Organization Name and Address 126-62 -01 -04-24 NASA Flight Research Center P. O..Box 273 11. Contract or Grant No. Edwards, California 93523 13. Type of Report and Period Covered 12. Sponsoring Agency Name and Address Technical Note -. WNaatsiohninagl tAone,r oDn.a uCt.i cs 2a0n5d4 6S pace Administration 14. Sponsoring Agency Code ~ -. ~. . ... 15. Supplementary Notes ~- 16. Abstract An airborne simulator P rovided a wide range of maximum roll control power (0.05 to 3.5 rad/sec ) and time constants (0.1 to 10 sec) for pilot evaluation and rating. Roll criteria were developed and compared favorably with previously reported criteria. Maximum roll angular acceleration, maximum roll rate, roll time constant, time to bank, and bank-angle change in a given time all appear to be effective roll-criteria parameters. Steady-state roll rates 01 about 20 deg/sec and roll time constants of 1.8 seconds or less were required for satisfactory pilot ratings. With experienced test pilots, valid evaluation of single-degree-of-freedom roll response can be obtained with a fixed-base simulator. -- - - .. - - . .. ~. .- 17. Key Words (Suggested by Author(s) ) 18. Distribution Statement Roll handling qualities Transport airplanes Unclassified - Unlimited I I I 19. Security Classif. (of this report) 20: Security Classif. (of this page) 21. NO. of Pages 22. Price' Unclassified Unclassified 106 $3.00 'For sale by the Clearinghouse for Federal Scientific and Technical Information Springfield, Virginia 221 51 FLIGHT INVESTIGATION OF THE ROLL REQUIREMENTS FOR TRANSPORT AIRPLANES IN CRUISING FLIGHT Euclid C. Holleman Flight Research Center SUMMARY An in-flight evaluation of roll handling for transport aircraft in cruise was con- ducted utilizing a general- purpose airborne simulator to provide single -degree -of - freedom roll dynamics. Maximum roll control power to 3.5 rad/sec2 and roll time constants from 0. 1 second to 10 seconds were evaluated and rated by five pilots in smooth-air conditions. Pilot evaluation and ratings were the important results from the study and provided the basis for the roll criteria that were developed and compared with other criteria. Pilot response to a well-designed questionnaire was effective in developing the roll criteria. Maximum roll-control angular acceleration, maximum available roll rate, roll time constant, and bank-angle change in a given time all appear to be ef€ective roll- criteria parameters. A steady-state roll rate of 15 to 20 degrees per second and roll time constants of 1.8 seconds or less were required for acceptable and satisfactory pilot ratings. Optimum pilot ratings were given for a roll capability of about 40 degrees per second with a time constant of 0.3 to 0.4 second. A wide range of roll response per unit of wheel control travel was rated satisfactory. Transport cruise rolling could be accomplished with very low levels of roll damping with increased pilot attention and compensation. The roll criteria developed from this program were in general agreement with previously proposed roll criteria. INTRODUCTION The roll requirements for fighter and other highly maneuverable types of airplanes have been studied in some depth with airplanes and with moving and fixed-base simulators. Creer et al. (ref. 1) used a "roll chair" to provide rolling motion for a piloted simulation of up-and-away flight. The result was the definition of satisfactory, unsatisfactory, and unacceptable regions of roll -control power and damping for fighter airplanes in up-and-away flight. Only single degree -of -freedom roll was considered, but the results have been verified to some extent in flight with variable -stability airplanes. Other investigations extended these results by considering the effects of other variables on roll response. In reference 2, for example, a fixed-base simulator was used to consider the influence of aerodynamic coupling, control coupling, and airplane damping and stability on roll handling. From these results, pilot ratings may be esti- mated for a wide range of airplane types and missions. Pilot variability was also con- sider ed. Theoretical studies have also contributed to the understanding of roll requirements. The study reported in reference 3 investigated the implication of roll-rate capability during attack and avoidance maneuvers and concluded that relatively low roll rates were required for maneuvering all types of airplanes. For a 2g turn, for example , only a 2-percent reduction in collision range was obtained with roll rates greater than 20 degrees per second, indicating that most airplane missions can be accomplished with relatively low rates. A summary of roll handling-qualities research is presented in reference 4. Sum- marized are flight, simulator, and analytical considerations of the acceptability of airplane roll characteristics, including pilot gain required for the roll -control task, Roll control and response for transports were also considered, but there were little actual flight data to support the analysis and conclusions. New airplane designs stimulate reviews of handling-qualities criteria and predic - tions of design acceptability. Reference 5 used a piloted simulator to predict the roll handling qualities of supersonic -transport configurations in cruising flight. Bisgood (ref. 6) and Leyman and Nuttall (ref. 7) reviewed handling-qualities research in an attempt to determine rrpplicability for future designs. Roll -control criteria were con- sidered and proposed as a result of these studies. Proposed revisions to the Military Specifications for piloted airplanes also provided the impetus for updating handling- qualities criteria. Recently, a comprehensive review and updating of Military Speci- fications were completed by the Cornel1 Aeronautical Laboratory (refs. 8 and 9). This investigation included the roll requirements for many classes of airplanes, but again showed little flight data on which to base roll requirements for transport and other large aircraft in up-and-away flight. Several investigators have studied the roll control required for the approach and landing maneuvers of various types of airplanes (refs. 10 to 13). Roll criteria for acceptable handling were proposed in terms of bank-angle change in 1 second, time to bank, and roll -control sensitivity rather than roll -control power and damping proposed for up-and-away flight. Although the control reqpired for the landing approach has determined the design for many types of aircraft, the intended operating envelope of present transport airplanes dictates that roll controls be designed for various parts of the flight envelope; thus , roll -control requirements are needed throughout the flight envelope. The present program was planned to provide design information of this type for transport aircraft in cruising flight in smooth-air conditions. During the study, pilots evaluated the in-flight handling of a wide range of roll time constants and levels of roll- control power. Yaw coupling was minimized, and rudder control was not used. Longi- tudinal control and trim were used as required for constant-altitude turns. The variable -stability JetStar airplane, designated the general purpose airborne simulator (GPAS), was used to provide the in-flight piloting task for evaluation. The GPAS has a model-controlled simulation system which provides the capability of duplicating a wide range of airplane characteristics. The range of characteristics studied included 2 the roll characteristics most likely to be considered in the design of either subsonic or supersonic transports for cruising flight. The results are presented as pilot evalu- ations and ratings and should provide basic data with which results giving the effects of other variables, such as turbulence and lateral-directional coupling, can be com- pared. Roll criteria are proposed and compared with referenced results. SYMBOLS e roll-rate error, deg/sec P sideslip error, deg eP “B sideslip-rate error, deg/sec e roll -angle error, deg co lateral wheel force, lb (N) FW g acceleration due to gravity, ft/sec2 (m/sec2) maximum roll -control angular acceleration or control power, rad/sec2 L6a6amax L roll acceleration due to aileron control, l/sec2 6a P roll rate, deg/sec commanded roll rate, deg/sec PC model roll rate, deg/sec Pm steady-state roll rate, deg/sec ps s maximum steady-state roll rate, deg/sec (Pss)max S Laplace operator, per sec 3 t time, sec time to bank 30", sec aileron-control deflection, deg commanded aileron control, deg model aileron deflection, deg rudder deflection, deg control-wheel deflection, deg roll time constant, sec bank angle, deg commanded bank angle, deg model bank angle, deg bank-angle change in first second, deg bank-angle change in first 2 seconds, deg frequency, rad/sec EQUIPMENT AND SIlMULATION Description of GPAS The general purpose airborne simulator is a Lockheed JetStar transport airplane with a model-controlled variable-stability system (ref. 14) installed to provide simu- lation capability. The general layout of the airplane is shown in figure 1, and a block diagram of the principal components of the model-controlled system is shown in 4 figure 2. The evaluation pilot's control inputs are routed to the airborne analog com- puter through the artificial-feel system. The computer is programed with the equations of motion to be simulated. For this investigation the equation used in transfer-function form was simply , Model response is compared with that of the JetStar, and the difference signal actuates the JetStar control surface to minimize the error. Roll rate and attitude were used as the control loops. With sufficiently high control-loop gain, the error is small and the computer model dynamics are reproduced closely by the JetStar airplane, The gains were: -PC- -1.0 e6a -- 1.0 sec Pm P 6 -V-C - 1.0 -a = 2.5 e Vm v A model was not mechanized for sideslip, but sideslip and rate-of-change-of-sideslip loops were used to minimize sideslip. The sideslip gains were: -6, = 2.0 sec eB The basic JetStar longitudinal dynamics for a Mach number of 0.55 and an altitude of approximately 20,000 feet (6,100 meters) were controlled in pitch by the pilot. The airplane's natural frequency at this flight-test condition in pitch was 2.55 radlsec, and the damping ratio was 0.5. These longitudinal dynamics have been rated satisfactory during handling-quaiities programs and so should not detract from the roll evaluation. - Displays and controls. A special set of transport-airplane types of controls and displays were used by the evaluation pilot, who occupied the left pilot station. The controls for the left pilot station (fig. 3) were disconnected from the airplane control system, and the pilot "flew1' the model on the analog computer of the simulation system. In flight, the normal horizon and other outside peripheral visual cues were used by the pilot, and basic displays of JetStar heading, bank angle, pitch attitude, rate of climb, and sideslip were presented on the left pilot's panel (fig. 4). During ground simulations, the left pilot's panel displayed either model-response quantities or simu- lated JetStar response quantities. The primary instrument display for the roll study was roll attitude. The response characteristics of the instrument are shown in figure 5. The ratio of actual roll 5 attitude displayed to sine-wave inputs is shown in figure 5(a). Note that the response follows well but tends to be flat-toppedy which could be interpreted by the pilot as a lag during the reversals at near peak oscillation amplitude; however, no pilot commented on the effect of the flat-topped response. The actual amplitude ratio (reflecting the flat-top effect) was constant over the frequency range of interest for this program (fig. 5(b)), and the phase lag (measured as if the response were sinusoidal) appears to be acceptable; it was less than 40" over the rsage of frequencies of interest for this program. During the evaluations, the evaluation pilot maneuvered in roll and controlled in pitch as required. The artificial-feel system provided for him was an electrohydraulic control system designed to provide the capability of simulating a wide range of control- system characteristics. Applied force was detected by strain gages which commanded hydraulic servo position which, in turn, moved the control wheel to correspond to the applied force. The control position can be a function of preselected force gradients and nonlinearities; however, for these tests no breakout or hysteresis was simulated, and a roll-control gradient of 0.4 lb/deg (1.8 N/deg) was selected, with increasing force gradient at a wheel deflection of 60" (fig. 6(a)). (During one flight, the pilots selected force gradients for several flight conditions. ) The frequency-response characteristics of the roll-control feel system (fig. 6(b)) were determined by harmonic analysis of a pilot controlling with randomly varying frequencies to indicate the adequacy of the roll control for the program. The measured response can be approximated by an overdamped second-order system with a natural frequency of 10 to 12 rad/sec, which is typical of hydraulically actuated control sys- tems. The force gradient in pitch was 22 lb/in. (3.85 N/cm), which was described by one pilot as being lighter than that of most transports. Other pilots did not comment on the longitudinal force characteristics. A delay time, which was a function of the time lag simulated, was noted and is discussed in some detail in a later section. Data-acquisition system. -On each flight, approximately 40 parameters of more than 150 available were recorded on two 50 -channel oscillographs. A 7 -cps filter was used to attenuate high -frequency noise on the oscillograph recordings. Analog- computer model and JetStar responses, as well as pilot inputs and selected model- control systems parameters, were recorded. Some quantities were recorded twice with different scale factors for better resolution. A 12 -channel direct -writing oscillograph was used for in-flight analog computer and GPAS following checks. A voice tape recorder was used to record all pilot comments. GPAS Roll Simulation A GPAS validation program indicated that the airplane/system was capable of highquality reproduction of large -airplane model dynamic -response characteristics; however, a delay time was noted between model roll response and GPAS following in roll. Figure 7 presents examples of GPAS response to step commands of the model aileron control for the roll time constants investigated. Following requirements during actual piloting tests were expected to be less severe than the step commands of 6 figure 7. Step control commands are not used by pilots in normal maneuvering. For a roll time constant of 0.1 second, figure 7(a) shows a delay lag of about 0.1 second between model roll rate and airplane following. Shifting the model roll rate and com- paring it with the roll-rate response (second trace, fig. 7(a)) indicated following more typical of a higher order system than a simple time-constant response. However, the pilots were unable to recognize the order of the response, but did appreciate the fast response and commented favorably on the roll simulation. Evidence of slight turbu- lence is also apparent in the airplane roll rate. Figure 7(b) illustrates roll-rate model following by the GPAS for a roll time con- stant of 0.35 second. Again, the model roll-rate trace has been shifted in time to give a better indication of the quality of reproduction of the model first-order response. A delay time of approximately 0.1 second is apparent before the JetStar responded in roll and, after a shift of an additional tenth of a second, the airplane response followed the first-order response of the model. However, shifting the model response an ad- ditional tenth of a second in figure 7(a) did not provide agreement between airplane and model response. As roll-rate model response approached and became longer than that of the basic JetStar (the time constant of the JetStar was about 0. 8 sec), following became more like first order (fig. 7(c)f or T~ = 1.0 sec, fig. 7(d) for T~ = 3.0 sec, and fig. 7(e) for T~ = 10.0 sec). However, delay times were evident before the airplane roll-rate response matched that of the model for the low rates commanded in these examples. For the checks of the longer time constants, it was necessary to command low response rates in order not to exceed the bank-angle limitations of the JetStar, since it was desirable to obtain a recording of at least three time constants to achieve a good ap- proximation to the steady-state roll rate. The delay times (fig. 8) between the model roll-rate response and the JetStar roll- rate response were measured for the range of time constants of interest for this pro- gram and over a wider range of roll rates than could be obtained during the in-flight checkouts. Delay times were short, approximately one-tenth second, at short time constants and for high roll rates. As the time constant was increased and the com- manded roll rate was decreased, the delay time increased. Delay times were less than 1 second for commanded roll rates as low as 2 deg/sec. No roll response was obtained for roll-rate commands of approximately 0.3 deg/sec or less. As a part of each evaluation, the pilot was asked to demonstrate normal and fast roll rates for transport maneuvering. The rates demonstrated are indicated in fig- ure 8 as the crosshatched regions. For low roll rates, the delay times for short time constants approached the value of the time constant simulated. Delay times for the long time constants were a much lower percentage of time constant simulated. Question H on the pilot's questionnaire (table 1) asked whether objectionable lag existed between the control wheel force or displacement and the JetStar roll response. A summary of pilot response to the question is presented in figure 9. "No," "Yes," or "Slight" answers are summarized as functions of maximum steady-state roll rate and roll time constants. Lag in roll response was most often noted at the longer time constants than at the low roll rates, indicating that the pilots were probably commenting on the delay in roll-rate buildup or the effect of the roll time constant rather than the delay time which occurred without JetStar response. There were no specific pilot 7 comments concerning the simulation-system delay time , nor were there comments concerning the order of the roll response not being a first-order lag. The effect of lag in combined response was studied during the investigation of roll controls for transport airplanes in references 11 and 12. Control-system rate limits were adjusted during the study reported in reference 11 to allow full control in 0.4 sec- ond and 0.9 second. The change in rate limit had no significant effect on the pilot evaluation except as might be noted in the slightly reduced bank angle achieved with a given program of time-displacement wheel position. In reference 12, the control- system rate limit was changed to allow maximum control to be achieved over the range of 0.2 second to 1.4 seconds. From the investigation it was concluded that there was little effect on pilot rating up to a lag of 0.7 second, and there was a degradation in rating of only about one pilot rating number to the lowest rate limit investigated. The pilots described the degraded control -system response as an apparent increased roll time constant for the large wheel deflections. The results implied that the lag or delay time could be interpreted as a slightly increased roll time constant. From these re- sults and the results of the pilot questionnaire in the present study, it was concluded that the delay time had little effect on the pilot ratings obtained and that the simulation was acceptable. Conduct of the Experiment The prime variables of the program were level of roll-control angular accelera- tion and roll damping or roll time constant. Roll -control angular accelerations ranged from 0.05 rad/sec2 to 3.5 rad/sec2, and roll time constants ranged from 0.1 second to 10 seconds. This resulted in steady-state roll rates of 1 deg/sec to a theoretically possible 2100 deg/sec , with a control-wheel and aileron deflection of about 60" avail- able to the pilot. Five experienced test pilots participated in the program. Three of the pilots flew 90 percent of the program, and two other pilots evaluated typical conditions. Although none of the pilots had experience as airline transport pilots, all were experienced test pilots with a varied background of flight test and evaluation experience, including large transport airplanes. All were familiar with handling qualities evaluations and pilot rating scales. Four of the pilots were NASA Flight Research Center research pilots, and one was an engineering test pilot for the Boeing Company. Pilots A, B, C, and D had approximately 3000, 4000, 1000, and 500 hours, respectively, of flight test ex - perience; they had 2500, 8180, 2800, and 250 flight hours, respectively, in transport or bomber types of airplanes. The order of evaluating test conditions was selected randomly, and no pilot was aware of the test condition to be evaluated prior to the actual evaluation. Some con- ditions for evaluation were repeated, in some instances on the same flight; other con- ditions were evaluated as many as nine times during the program. A single pilot repeated evaluations of a test condition as many as four times, and each of the three primary pilots repeated evaluations of conditions at least three times to indicate pilot . v ari a b il ity During checkout of the simulation, the pilots were acquainted with the goals of the program and the pilot questionnaires were finalized. Evaluations were conducted 8