NASA Technical Memorandum 88289

P

Flight Test Techniques for the X=29A

John W. Hicks, James M. Cooper, Jr., and Walter J. Sefic

February 1987

(NASA-TH-88289) €LIGHT TEST 'ICCllNIQUES FOR N87-2 1908 IEE x-29A AIBCRAfl (NASA) 14 p Avail: 131s HC A02/flF 801 CSCL 01c Unclas Hl/OS 0070057

National Aeronautics and Space Ad minist rat ion NASA Technical Memorandum 88289

. .

Flight Test Techniques for the X-29A Aircraft

John W. Hicks, James M. Cooper, Jr., and Walter J. Sefic Ames Research Center, Dryden Flight Research Facility, Edwards, California

1987

National Aeronautics and Space Administration Ames Research Center Dryden Flight Research Facility Edwards. California 93523-5000 FLIGHT TEST TECHNIQUES FOR THE X-29A AIRCRAFT

John W. Hicks,* James M. Cooper, Jr.,** and Walter J. Sefic+

NASA Ames Research Center Dryden Flight Research Facility Edwards, Cal iforni a

Abstract pEst parameter estimation program

The X-29A advanced technol ogy demonstrator RAV remotely augmented vehicle is a single-seat, single-engine aircraft with a forward-swept wing. The aircraft incorporates Introduction many advanced technol ogies being considered for this country's next generation of aircraft. This The unique X-29A forward-swept-wing aircraft unusual aircraft configuration, which had never required a specially tailored flight test program been flown before, required a precise approach to to efficiently, yet safely, expand the 1-g and flight envelope expansion. Special concerns were maneuver envelopes whi 1 e gathering research data static wing divergence and a highly unstable air- to evaluate several highly integrated advanced frame. F1 ight envelope expansion included the technologies. Conducting a multidisciplinary test 1-g flight envelope and the maneuver envelope at program on a single experimental aircraft of radi- higher normal load factors and angle of attack. cal design was very challenging. The program Real -time analysis was required to clear the required tightly coordinated flight test planning envelope in a safe, efficient manner. In some coupled with specially developed maneuver tech- cases, this led to development of a unique real- niques and real-time and postflight data analysis time analysis capability. The use of real-time capabilities. Heavy reliance on ground and in- displays to provide for quick analysis of stabil- flight simulation and extensive predictive pre- ity and control, dynamic and static structures, flight analysis were also essential to program flight control systems, and other systems data success. Several specific technical areas, such greatly enhanced the productivity and safety of as classical flutter, aeroservoelasticity, flight the flight test program. control system (FCS) stability, structural loads, , propulsion, performance, and This paper describes the real -time analysis basic stability and control characteristics, had A_ iileihsds afid f:tgkt t~~tt~tk~!.;.i~ use.' .'U~!C? LU be C:Fared ofi a sing:e ;:;plane. ?. flight rate the envelope expansion of the X-29A aircraft, of up to four flights per day (high for an X-type including new and innovative techniques that pro- airplane) and two flight days per week required an vided for a safe, efficient envelope expansion. emphasis on real-time analysis. Specialized real- The use of integrated test blocks in the expansion time analysis techniques had to be developed to program and in the overall flight test approach get critical answers while in flight to clear the will be discussed. aircraft to the next envelope point.

Nomenclature The X-29A project was begun in 1977 by the Defense Advanced Research Projects Agency (DARPA) ACC automatic camber control and the Grumman Aerospace Corporation. Because of the obvious safety-of-flight advantages of the AR analog reversion Edwards Air Force Base complex, the decision was made to ship the aircraft to Edwards AFB for its BFF body-freedom flutter first flight and conduct the enti re envelope expansion program there. It was shipped through DR digital reversion the Panama Canal, arriving at Edwards AFB in October 1984. A flight test team, located at the FCS flight control system NASA Ames Research Center's Dryden Flight Research Facility (Ames-Dryden) and made up of Ames-Dryden, FDMS flight deflection measurement system Air Force Flight Test Center, and Grumman person- nel, successfully completed the first flight on FM frequency modulation December 14, 1984. Pmes-Dryden was the respon- sible test organization, with safety-of-flight and F SW forward-swept wing operational responsibility. (Further details of programmatic issues can be found in Ref. 1.) The 9 normal load factor flight envelope expansion was completed by the end of 1986. The prime objective was the expansion ITB integrated test block of the 1-g flight envelope in Mach number and altitude and the angle of attack and normal load LED light-emitting diode factor maneuvering envelope to the limits deter- mined by the FCS and structural ground tests. MCC manual camber control With the restriction of a single airplane, PCM pul se-code modulation several technical discipline test requirements and envelope clearance tasks had to be integrated *Chief Engineer, X-29 Program. AIAA member. into the flight program. These included controls, **Aerospace Engineer. structural loads, structural dynamics, aircraft tproject Manager , X-29 Program. AIAA member. systems, and propulsion. More research-oriented

This paper is declared a work of the U.S. Government and therefore is in the public domain. Investigations of aircraft performance, local Incorporated) data bus are combined in an instru- aerodynamics. handliny qualities, and stability mentation component known as an interleaver and and control derivative extraction were also are output as a serial PCM stream. In addition, included. Structural loads had to clear both the a constant-bandwidth frequency modulation (FM) basic loads envelope in the automatic system is installed to acquire high-response camber control (ACC) winy configuration mode and acceleration and vibration data. The output of the static winy diveryence in manual camber con- the FM mutiplexer is routed to a prernoaulation trol (MCC) mode. mixer where it is combined with the pilot's voice. Air-to-ground telemetry is the only source of air- Ai rcraft Descri pt+ craft data since there is no onboard recordiny system. Instrumentation provisions include .The X-2YA advanced technoloyy demonstrator measurements for structural loads and dynamics, (Fiy. 1) is a single-sedt aircraft desiyned to flight controls, stability and control, aircraft evaluate the syneryistic effects of several inte- subsystems, propulsion, and performance. The grated technoloyies to advance the state of the F404-6E-400 enyine contains an F-18-style fliyht art of this country's next generation of fiyhter test thrust instrumentation system, but only the c aircraft. Its most notable feature is the for- GE basic enyine instrumentation portion was put ward-swept winy (FSW) having a 33.73O quarter- into operation for the envelope expansion phase. chord winy sweep. The primary wingbox is covered More discussion of engine instrumentation can be with aeroelastically tailored yraphite epoxy found in Ref. 2. b covers attached to conventional aluminum and tita- nium spars. The load-beariny wing covers give a Externally, 12 infrared light-emitting diodes 1 iyhtweiyht,.yet strong, structure that controls (LEDs) were mounted across the top of the right winy deflection and the tendency of the FSW toward wing as part of the flight deflection measurement winy divergence. The winy has a 5-percent-thick system (FDMS). The LEDs transmitted light to a supercritical airfoil, a fixed , and dual receiver mounted in the right side of the full-span, dual-hinyed trailing , which above the wingroot. The left wing, yield high lift duriny takeoff and landing and , , and strake upper surfaces provide the sole source of lateral contro.1 for the were instrumented with 179 flush-mounted static aircrdtt. The flap is also used to vary the wing pressure taps to measure aerodynamic pressure camber either automatically with the continuous distribution. Finally, each wing pontoon, con- ACC mode or manually by the pilot using the taining the mid- and outboard hydraulic discrete MCC mode. Flaperon deflection ranye is actuator, was fitted with a special aft-mounted from 10' trailiny edge up to 25" flaperon eccentric-rotary-mass structural excita- down. The winy aerodynamics are coupled with a tion system to investiyate the supersonic flap-tab full-authority, high-rate canard located forward single-degree-of-freedom flutter potential. The of the winys at the leadiny edye of the enyine FDMS and structural excitation system are tempor- inlets. The travel of the canards is 30' leading ary external devices used specifically to assist edge up to 60° leading edge down at a maximum rate in the structural clearance of the flight envelope. of 100 dey/sec. Aft-mounted fuselaye strake flaps with a *30° travel work in conjunction with the Real-Time Processiny and Display canards and symmetrically deflected flaperons to provide a unique three-surface aircraft pitch The NASA Ames-Dryden Western Aeronautical Test control. A single-piece provides yaw Range provided the real-time monitoring and analy- control. The aircraft forward of the engine sis capability for the envelope expansion work. inlets consists of a standard F-5A nose section, Since all data from the X-29A aircraft were avail- cockpit, and nose gear. able only over a telemetry downlink during flight, all data had to be recorded, and in some cases The nominally 35-percent negative static analyzed, in real time on the ground. The telem- margin of the aircraft is stabilized by a highly etry downlink and data processing setup is out- auymented triplex di gital-analog fly-by-wi re lined in Fig. 3. The range provides telemetry control system. The FCS has three modes: normal acquisition and processing, real-time data analy- (primdry) mode, digital reversion (DR) mode, and sis and display, voice communication links, radar analog reversion (AR) mode. tracking for space positioning, and video displays of the aircraft in flight. This information was Propulsion is provided by a single General made available to two mission control rooms, the Electric F404-6E-400 afterburning engine, rated "blue room" and the spectral analysis facility, at 16,000 Ib thrust at sea level. The engine is which worked in parallel to direct the flights. mounted in a fuselaye with two side-mounted, fixed- The blue room consists of 12 strip charts, numer- geometry inlets that were desiyned for transonic ous CRT displays, and a terminal console hookup performance. Aircraft takeoff gross weight is to a computer providiny real-time data proces- 17,800 lb with a 4000-lb fuel capacity in the siny. Gould SEL 32/55 and 32/77 minicomputers fuselaye and strake tanks. are used to process telemetry information, to calibrate it, to convert it to engineering units, Instrumentat ion and to display it to the control room. This con- trol room contains the test conductor, ranye com- The fliyht test instrumentation system munications personnel, project management, and (Fiy. 2) was tailored for envelope expansion discipl ne engineers from controls, propulsion, and some flight research measurements. Because systems 1oads , and aerodynamic stabi 1 ity and of airframe internal volume constraints, a 10-bit control remote unit pu 1 se-code modul at ion (PCM) system was used for data aquisition. The outputs of the five The spectral analysis facility consists of PCM units and the ARINC 429 (Aeronautical Radio, six str p charts, CRT displays, a communications

2 link to the blue room, and two real-time com- maneuvers were designed to obtain special flight puters. One of the computers, an HP 5451C Fourier research data to aid in evaluating the advanced Analyzer, is used to extract structural dynamics technologies. The actual envelope expansion took frequency and damping data for the primary air- place in two phases. The first phase was the frame modes. The second computer, a Gould Concept limited flight envelope expansion below Mach 0.60 Series 9780, is used at a flight controls station and 30,000 ft. After checking out the aircraft to analyze controls system frequency response to in this region and making some FCS modifications, determine system stability margins. Reference 3 the rest of the full 1-g envelope was expanded contains a more complete description of the NASA to 40,000 ft and supersonically to maximum speed. Ames-Dryden range facilities. Integrated Test B1 ock Approach Basic F1 ight Test Approach The 1-g envelope clearance was accompl ished by Several technical areas of concern had an performing a series of maneuvers at a given alti- impact on the basic X-29A flight envelope clear- tude and Mach number. Two integrated test blocks ance. The primary objective, of course, was to (ITBs) were developed to clear the aircraft enve- clear the aircraft in speed and altitude and in lope in several discipline areas. The ITB-1 was angle of attack. Another goal was to attain some used for the 1-g envelope clearance and consisted level of positive normal load factor capability on of the following four maneuvers: the aircraft, not only to obtain loads data but also to be able to maneuver the aircraft at a 1. Structural dynamics block given Mach number and altitude. Minimum stability 1-min stabilized point margins of the aircraft FCS had to be maintained Three-axi s control raps throughout the flight envelope and were a concern 15-sec hands-off trim point in the transonic flight regime. The actual FCS stability trends for such a highly unstable air- This block was used to clear the aeroservo- plane were uncertain. Structural static wing elastic and flutter envelopes. The three-axis divergence and the coupling of the pitch short- raps and natural turbulence excitation put energy period mode with the first wing bending mode into the structure. The hands-off point was used (this coup1 ing is known as body-freedom flutter) to accomplish airspeed and angle-of-attack cali- were strong influences on the envelope expansion brations. approach as dynamic pressure increased. Canard loading effects in the transonic area at high 2. Longitudinal block dynamic pressures also had a strona influence. A possible supersonic flutter mode in the dual- FaSt-SIOW pitch doublet hinged wing-mounted flaperons was another concern; Repeat doublet the flaperon rotary-mass shaker system was used Stick rap in an attempt to identify this phenomenon. Over- Frequency sweep all, aircraft aerodynamic stability and control surface effectiveness, especially transonically This block was used to gather data that were and at higher angles of attack, affected the entered into the real-time phase and gain margin approach to parameter estimation of the aerody- controls clearance routines. Approximately 15 sec namic derivatives. after completion of this block, a clearance could be passed to the pilot. This block was also used The basic approach used in expanding the to gather data for the longitudinal stability and X-29A flight envelope was to first clear the control analysis that generated aerodynamic deriv- 1-g flight envelope at a given Mach number and atives in postflight processing and to clear the altitude. The second step was to fly at the 1-g static. loads envelope. next lower Mach number at that given altitude and clear the maneuver envelope in angle of attack 3. Lateral block and normal load factor. Typically, Mach number Roll -yaw doublet increments for envelope expansion were 0.05 sub- Repeat doublet sonically, 0.02 transonically, and 0.1 superson- Full-stick 0' to 60' rol s, left and right ically. Figure 4 shows the 1-g flight envelope Stick rap test point matrix and a typical sequence (solid Frequency sweep symbols) used to clear the envelope. Primarily because of structural clearance considerations, This block was used to ga her lateral- a specific dynamic pressure level was first ap- directional stabilitv and con rol derivatives. proached at a higher altitude, which allowed for The frequency sweep was used to generate lateral- a more gradual increase in dynamic pressure with directional handling qualities parameters. Mach number than at a lower altitude. The expan- sion then continued at a constant dynamic pressure 4. Directional block by flying to the corresponding Mach number at the Yaw-roll doublet next lower target altitude, usually 10,000 ft Pedal rap below the previous one. This was followed until the low-a1 titude, high-Mach-number region was This block was used to gather data for static reached, where constant-a1 titude expansion was loads and stability and control clearance. substituted for constant dynamic pressure due to considerations of static wing divergence clear- ance. In addition to the test maneuvers designed With the successful completion of the ITB-1 to expand the flight envelope, the angle-of- at any given Mach number, the aircraft was slowed, attack and normal load factor expansion maneuver usually by 0.05 Mach number, and a maneuver clear- sequence contained some added maneuvers. These ance block, ITB-2, was performed. This block

3 expanded the angle-of-attack, normal load factor, 1. Symmetric pullup and maneuver capabilities of the vehicle by 2. Symmetric pushover including the following maneuvers: 3. Left and right turn reversal 4. Left and right rudder kick 1. Steady-heading sidesl ips 5. Left and right rudder reversal 3" sideslip Full-pedal or limit sideslip Structural Dynamics Techniques

These were used to gather static loads and The primary real -time analysis technique used lateral-directional stability and control data. to track the structural frequency and damping of five main modes was the random-data auto-power- 2. Roll -yaw doublets at half-maximum sidesl ip spectrum analysis. The five modes included the first symmetric and antisymmetric wing bending 3. Full-stick 0" to 60" rolls, left and right modes, the first fuselage vertical and lateral bending modes, and the first vertical fin bending These rolls were used to gather loads, con- mode. Monitoring the symmetric first wing bending trols, and hand1 ing qual ities data. These were modal frequency in real time was especially impor- buildup maneuvers to the 360" roll. tant in tracking the onset of the body-freedom flutter (BFF) phenomenon. While actual mode 4. Full-stick 360" roll, left and right coupling occurs at about 2 Hz, onset was predicted to begin as early as 5 Hz. The BFF phenomenon These rolls were used to gather loads, acts as a precursor to the static wing divergence controls, and handling qualities data. problem. At a given Mach number and altitude con- dition, data from an approximately 60-sec interval 5. Roll-pitch step inputs were required for analysis. The required maneuver consisted of a stabilized point with random air These 0.5-in step inputs and the resultant turbulence and three-axis stick rap inputs. The control and aircraft responses were used for sim- eccentric rotary-mass flaperon excitation system ulation correlation work. was also used at times as an input for the entire airframe. The primary function of the flaperon 6. Constant-altitude windup turn excitation system, however, was to provide wing excitation to track the supersonic flap-tab Data from these maneuvers were used in the single-degree-of-freedom flutter. loads, controls, and wing pressure disciplines for 1 oad clearance. These maneuvers were accomplished A fast Fourier analysis was performed on to the cleared loads envelope to establish data the data with a minicomputer using a frame size that would allow further load clearance. of 1024 data samples at 100 samples/sec. On the ground, an analog bandpass filter was used 7. Constant-thrust windup turns to isolate each structural mode. In a similar manner, postflight analysis of the same maneuver Data from these turns were used in the per- points was performed on the canard random data formance discipline to derive drag polars. using a frame size of 2048 samples at 200 samples/ sec. The inverse Fourier transform was computed 8. Pushover-pullups to obtain the autocorrelation function from which a data cutoff time could be selected and smoothing This wings-1 eve1 sequence consi sted of 1-g performed to obtain a smoothed auto-power-spectrum stabilized flight, a pushover to 0 g followed by display (Fig. 5). Each mode in the auto-power- a pullup to 2 g, and a return to 1-g stabilized spectrum display was fit with a least-squared- flight. These maneuvers were used to develop error parabolic curve whose modal frequency was drag polar shapes and to fill in the 0- to 1-g the maximum amplitude of the curve, and the loads range. damping was extracted using the ha1 f-power tech- nique. More specific information on the method can be found in Ref. 4. The actual frequency and The ITB-1 and ITB-2 maneuvers cleared the damping of the five main modes were compared in ACC flight control mode. The MCC flight control real time with precomputed predictions (Fig. 6). mode was then evaluated by a full ITB-1 followed Damping or modal frequency that were signifi- by pushover-pullups and windup turns. These data cantly lower than predictions caused a halt in were used primarily in wing divergence analysis the envelope expansion until the problem was to establish the static wing divergence dynamic understood. pressure predictions. Stability and Control By using the integrated test block approach, the X-29A aircraft was taken through the fl ight envelope while simultaneously clearing areas of Aircraft aerodynamic characteristics were concern in various disciplines. analyzed by extracting the nondimensionalized aerodynamics derivatives, using an interactive Having cleared the 1-g and maneuvering enve- parameter estimation program known as pEst. lopes, the emphasis shifted to gathering research Special three-axis stick doublet maneuvers were data. Numerous maneuver blocks were set up for used as data input to the computer program. To the different disciplines. For example, the loads increase the information content of the data, research group established the following set of these maneuvers consisted of pairs of fast and maneuvers to build up their X-29A data base: slow doublets in each axis, which were repeated

4 for the longitudinal and lateral cases. The tions for any significant aircraft response was highly augmented three-surface integrated pitch cause to Slow or ha1 t envelope expansion. control made the extraction of the three separate control surface derivatives impossible during the In-F1 ight Deflection Measurement envelope expansion phase of the program. For this reason, only an effective pitch control power The FDMS was utilized for wing static struc- derivative could be extracted, representing the tural divergence clearance; data were compared combined pitch control of all three surfaces. The with aeroelastic wing deflection predictions. The Cram'er-Rao bound was applied to determine the data were also used in analysis programs designed uncertainty levels in the individual derivatives, to derive the estimated wing static divergence and the data scatter was used to determine the dynamic pressure. The entire system was located validity of the derivatives. on the right-hand side of the aircraft, with the 12 LEDs on the upper right wing surface (Fig. 9). The high longitudinal static instability of These LEDs were focused on two diode array the X-29A aircraft requires artificial stabiliza- receivers that were mounted in the right-hand tion by the high-response FCS. As a result, the side of the fuselage behind a 5- by 7-in glass aircraft response is dominated by the FCS rather window, just above the wingroot. One diode array than by aerodynamics. Safety-of-fl ight consid- focused on the inboard set of LEDs and the other erations and the dominance of the FCS necessi- on the outboard set. Typical results, obtained tated the development of a real-time analysis primarily in postflight data reduction, can be technique to monitor the longitudinal-axis control seen in Fig. 10. Real-time computation and dis- system stability. A fast Fourier transform tech- play of the data were possible but not deemed nique was used to measure open-loop frequency necessary for envelope clearance. Wingbox twist response and the corresponding phase margin at data obtained from deflection measurements are the gain crossover frequency and gain margins at plotted as a function of wing semispan. Chordwise the high and low phase crossover frequencies. and spanwise relative displacement of the LED tar- Real-time calculations were compared with pre- gets yielded very reliable measurements of wing flight computed predictions as the envelope twisting and bending, respectively. expanded from point to point in Mach number and altitude. The stability criterion used to halt Structural Loads envelope expansion to the next Mach number point was a minimum gain margin of 3 dB or a phase The structural loads clearance of the X-29A margin of 22.5', or both, based on MILSPEC-F-9490D aircraft was accomplished for both ACC and MCC (Ref. 5). Significant adverse trends toward those wing modes primarily by postflight analysis. 5trtcttr;? 1 c;d 1 ;mi+ nn..nl nnnc ..I_Cn An,,nl nnnA minimum diw siuwtd ui halted the efivc:~p~~npan- 8 11111 .. c Il.C. "yc' ..-I c ..'.L."r-" fer sion process until the situation was understood 80 percent and 100 percent of limit, based on the and a remedy was developed. This real-time capa- 80-percent design load ground proof test. These bility was of great value to the program. Using limits were monitored in real time on a color CRT postflight analysis, only one expansion point presentation for the left canard, left lateral could have been flown per flight and only one fuselage, left wing, and vertical tail loads. flight every two to three days. The real-time A typical real-time presentation is shown in implementation of the stability margin analysis Fig. 11. Exceeding any of these structural loads and real-time clearance of the FCS performance envelopes would have caused the pilot to break off allowed the clearance of two or three Mach number a maneuver, halting any further envelope expansion and altitude points per flight at up to four to the next higher Mach number. Some results were flights per day. This greatly accelerated the used to modify any further maneuvering that might envelope expansion phase. cause airframe loads to exceed their limits. This usually resulted in a real-time call to decrease The real -time open-loop frequency response or limit the normal load factor target condition analysis required about 52 sec of flight data of a given maneuver. For each airframe component, recorded at 40 samples/sec, or a total of torque loads were compared with bending loads com- 2048 time history data points. The maneuver puted from in-flight strain gage measurements. block consisted of a stick rap, a doublet, and a frequency sweep in the pitch axis. Typical The canard actuator load limit was the only results of the real-time plot output are depicted nonstructural load limit observed during real in Fig. 7. Further details of the technique time. The monitoring of this lilnit was necessi- development can be found in Ref. 6. In addition tated by the single-system hydraulic actuator to determining the FCS phase and gain margins, a capability. The limit was set at 80 percent of real-time time history overlay of the actual air- the actuator limit defined in terms of an equiva- craft flight response and predicted response was lent load measured at the actuator. The common made from the same data. The predicted response load factor expansion maneuvers, such as the was generated in real time by entering the control constant-thrust pushover-pullup maneuver from deflections measured in flight and the initial 1-9 stabilized flight and the constant-altitude flight conditions into the mathematical models windup turn, were used to obtain loads data. and calculating the predicted response. A typical example is shown in Fig. 8. Additional details of Static wing divergence clearance was accom- this method are contained in Ref. 7. This real- plished by use of both strain gage and FDMS data. time comparison was an additional check on the FCS Separate postflight analysis of the data with the stability analysis in that it detected any air- Southwell technique8 gave an estimate of the craft stability degradation when clearing the air- divergence speed, which was compared with predic- craft to the next test point. An overall degra- tions. The canard strain gage data were analyzed dation in aircraft stability relative to predic- in a similar way to determine canard divergence

5 speed and were compared with predictions. The block approach, coupled with real-time analysis MCC wing mode allowed for temporarily fixed wing for immediate aircraft clearance, was essential to geometry during a constant-altitude windup turn, the successful conclusion of the expansion phase. from which the loads data were generated. This The X-29A program incorporated a number of new and mode allowed the pilot to manually set discrete innovative real-time display techniques to provide 5' increments of flaperon from 5' trailing edge for safe and efficient envelope expansion. The up to 25' (full) trailing edge down. use of real-time displays to provide for quick analysis of stability and control, dynamic and Remotely Augmented Vehicle static structures, flight control systems, and other systems data greatly enhanced the produc- To assist in the collection of data of higher tivity and safety of the flight test program. quality and larger quantity, the remotely aug- mented vehicle (RAV) system (Fig. 12) was incor- Refe rences porated into the X-29A aircraft. The pi1 ot-assi st version of the RAV system was tested in this lSefic, Walter J., and Cutler, William, phase. Aircraft state variables calculated from "X-29A Advanced Techno1 ogy Demonstrator Program the telemetered data are input to the computers, Overview," AIAA-86-9727, Apr. 1986. which generate guidance and control information. ZHicks, John W., Kania, Jan, Pearce, Robert, In The RAV system has two operating modes. and Mi11 s , G1 en, "Chal 1enges in Model ing the X-29A the first mode, a ground-based computer drives a Flight Test Performance,' AIAA-87-0081, Jan. 1987. set of airborne guidance needles (Fig. 12(b)) through a radio uplink. The ground-based computer calculates the flightpath required to accomplish 3Moore, Archie L. , "The Western Aeronautical the maneuver and then computes error signals from Test Range of NASA Ames Research Center," NASA the desired and actual values in the form of TM-85924, 1985. pitch, roll, and position. These errors are telemetered to the vehicle and displayed as 4Kehoe, Michael W., "AFTI/F-16 Aeroservo- commands to the pilot using the instrument landing elastic tnd Flutter Flight Test Program - system needles and the speed bug as indicators. Phase I, NASA TM-86027, 1985. This guidance assists the pilot not only in flying maneuvers more precisely but also in the tran- 5"Military Specification - Flight Control sition to new test points. Systems, Design, Installation, and Test of Piloted Aircraft, General Specification for," The maneuvers currently envisioned for this MILSPEC-F -9490D, June 1975. mode include (1) Mach number and altitude capture, (2) constant-angle-of-attack level turns, and 6Bosworth, J.T., and West, J.C., "Real-time (3) level accelerations and decelerations. Open-Loop Frequency Response Analysis of F1 ight Test Data ," AIAA-86-9738, Apr. 1986. This system will be used during the research phase of the program. Past experiences with a hauer, Jeffrey E., Crawford, David 6.. Gera, RAV system on the HiFlAT vehicle,g F-104,10 and Joseph, and Andrisani , Dominick, "Real-Time Com- F-8 digital fly-by-wire aircraft11 have proven the parison of X-29A Flight Data and Simulation Data," benefits of a more powerful ground-based computer AIAA-87-0344, Jan. 1987. in generating flight test maneuver guidance and control. 8Ricketts , Rodney H. , and Doggett, Robert V., Jr., "Wi nd-Tunnel Exeeriments on Divergence of In the second mode of operation, the ground- Forward-Swept Wings, NASA TP-1685, 1980. based computers can command any surface on the aircraft to perform a series of time-dependent movements that are preprogrammed in both frequency gDuke, E.L., and Lux, D.P., "The Application and amplitude. For example, the canard may be and Results of a New Flight Test Technique," commanded to go through a series of pulses, doub- AIAA-83-2137, Aug. 1983. lets, and sinusoidal frequency sweeps very similar to the longitudinal block in the ITB-1. This will Ioneyer, R.R., Jr., and Schneider, Cdr. E.T., allow the determination of control derivatives for "Real -Time Pilot Guidance System for Improved each surface. Reference 12 contains further F1 ight Test Maneuvers ,'I AIAA-83-2747, Nov. 1983. information on this RAV mode. IlHartman, Gary, Stein, Gunther, and Powers, Concluding Remarks Bruce, "Flight Test Experience With an Adaptive Control System Using a Maximum Likelihood Param- The X-29A aircraft has a number of unique eter Estimation Technique," AIAA-79-1702, 1979. aspects that required a carefully planned and integrated flight test program to expand its 1-g IzShafer, Mary F., "Flight Investigation of and maneuver envelopes. The particular chal 1enge Various Control Inputs Intended for Parameter was the multidiscipline approach on a single Estimation,'' NASA TM-85901, 1984. X-type airplane project. The integrated test

6 Vertical height, 14 ft 9.5 In. Bmo ECN 33297 009

Fk. 1 X-29A advrmced technology demonstmtor.

conditioning signal conditioning control input console I I + I I Telemetry I Pliot's Premoduiation I antenna I mixer Slgnei I inputs I I I I PCM I e , I I Dlgltai Signal interleaver + conditioning unit \

I Temperature subsystem Telemetry inputs transmitter

429 data bus computer interface Input 6127

Fig. 2 X-29A data acquisition system.

Altitude, fta 61 1""1 r'?Communications - 10 ,OF8 Real-time I /I control I I 1 and display centers .2 .4 .8 .8 1.0 1.2 1.4 1.8 I Mach number 6128 6129

Fig. 3 Real-time pocessing and dis kf- Fig. 4 night envelope etpansion appmach.

7 power approach Left wingtip accelerometer

Gain, dB

-20 - -30 -10 1 10 30 Frequency, radlsec 6132

Fig. 7 Real-time FCS frequency response. Fig. 5 Pambotic curve fit to dete&ne frequency and damping in near mat time.

-Flight Linear simulation

FCS mode 0 Nonnal -2 0 DR 0 AR - Pltch Predictions rate, - Nmai deglSec 0 ------e_- --- DR --- AR .20 47 Pitch 2 - /--%\/' attitude, .16 0' deg -2 IY Damping, .08 accelerometer -04 output, 0 rn -2 0 2 4 6 8 10 Time. sec 6133

Fig. 8 Reat-time computed l.on@uhat aircraft 4 response. 150 250 350 450 550 650 Equivalent velocity, knots 6131

Fig. 6 Typical structumt dynamics mcdat frequency and damping.

8 Fig. 9 Fliqht deflection measurement system.

0 Grumman predicted 0 Measured (flight deflection meas- urement system) 0

0 0

0

0 0 0 0 - Semispan, percent 6134

Fig. 10 Typical X-29A deflection measurement data for wingbox twist aa a function of semispan aom- pared with predicted data.

9 Fig. 11 Typical real-time graphics display of wing, canard, fuse- lage, and vertical tail bending as a function of torsion.

Test aircraft

Vertical Vertical pointer side Roll axis indicator scale calibrations Position corrections Comparison with flight Telemetry receiver transmitter Throttle Application of guidance Horizontal algorithm scale needle LPitch axis Ground-based computer scale 6135 6136

!a, Flight test tmjectory guidance system. (b) Generic pilot disptaq device.

Fig. 12 Remotely augmented vehicle system.

10 1. Report No. 2. Government Accession No. 3. Recipient's Catalog No. NASA M-88289 I I 4. Title and Subtitle I 5. Report Date FLIGHT TEST TECHNIQUES FOR THE X-29A AIRCRAFT February 1987 I 6. Performing Organization Code

7. Author(s) 8. Performing Organization Report No. John W. Hicks, James M. Cooper, Jr., and Walter J. Sefic H-1401 10. Work Unit No. 9. Worming Organization Name and Address RTOP 533-02-51 NASA Ames Research Center Dryden F1 ight Research Facility 11. Contract or Grant No. P.O. Box 273 Edwards CA 93523-5000 , 13. Type of Report and Period Covered 12. Sponsoring Agency Name and Address Technical Memorandum National Aeronautics and Space Administration 14. Sponsoring Agency Code Washington, E€ 20546 1 I 15. Supplementary Notes Prepared as AIAA-87-0082 for presentation at the AIAA 25th Aerospace Sciences Meeting, Reno, Nevada, January 12-15, 1987.

16. AbsPract

The X-294 advanced technology demonstrator is a single-seat, single- engine aircraft with a forward-swept wing. The aircraft incorporates many advanced techno1 ogies being considered for this country's next gen- eration of aircraft. This unusual aircraft configuration, which had never been flown before, required a precise approach to flight envelope expansion. Special concerns were static wing divergence and a highly unstable ai rframe. F1 ight envelope expansion included the 1-g flight envelope and the maneuver envelope at higher normal load factors and angle of attack. Real-time analysis was required to clear the envelope in a safe, efficient manner. In some cases, this led to development of a unique real-time analysis capability. The use of real-time displays to provide for quick analysis of stability and control, dynamic and static structures, flight control systems, and other systems data greatly enhanced the productivity and safety of the flight test program.

This paper describes the real-time analysis methods and flight test techniques used during the envelope expansion of the X-29A aircraft, including new and innovative techniques that provided for a safe, effi- cient envelope expansion. The use of integrated test blocks in the expansion program and in the overall flight test approach will be discussed .

18. Distribution Statement F1 ight envelope expansion Unclassified - Unl irnited Flight test techniques Forward-swept wing

19. Security Uanif. (of this report) 20. Security Classif. (of this page) 21. No. of Pages 22. Rice' Uncl assi fied Uncl assi fied 11 A 02

*For sate by the Rational Techmcal Information Service, Springfield, Virginia 22161.