Document
NASA
Tec hni ca l NASA-TP-2353 19840024308
Paper
September 1984
Validation of an Active
Gear, Flexible Aircraft
Take-Off and Landing
Analysis (AGFATL)
-j : ,_-
John R. McGehee
IM / A
3 1176 01328 7595 NASA ....
Technical
Paper
Validation of an Active
Gear, Flexible Aircraft
Take-Off and Landing
Analysis (AGFATL)
John R. McGehee
Langley ResearchCenter
Hampton, Virginia
Su m mary plying active load control to t he airplane main land - ing gear to limit the ground loads applied to the air- The logic and equations of a series - hydraulic active frame. A limited-degree-of-freedom, n onlinear active co n tr o l gear were inc o rporated i n to a mul t i-degree-of- control landing-gear analysis and computer program freedom flexible aircraft take-off and landing analysis ( ACOLA G ) was developed for studying active c ontrol ( FATOL A) to generate a computer progra m for ac - landing - gear c o ncepts (ref. 5 ) . In the same time frame, tive gear, flexible aircraft take-off and landing analysis a multi-degree-of-freed o m, stiff airframe take-off and (AGFATL). The results of previously c onducted exper- landing analysis computer program (TOLA) was ob- imental investigations consisting of shaker tests, drop tained from the A ir Force F light Dynamics Laboratory tests, and simulated-landing tests of passive and a c tive and modified to include flexible airframe c haracteristics versions of a general aviation airplane main gear are (FATO L A). (See refs. 6 to 12.)
compared with computed data to validate the AGFATL computer program. Computer program ACOLAG was employed to Comparisons of experimental and analyti c al data study the potential of a series-hydraulic active control for shaker tests (forced vibrations) and drop tests show landing-gear concept. Based on the favorable results of this study, an electroni c control and a hydraulic power good agreement for both the passive and active gears.
Experimental data for the landing tests were influenced unit were designed, fabricated, and tested with a mod- by large unmeasured strut-binding friction forces. The ifled main gear from a general aviation airplane to vail- inclusion of these friction forces in the analytical simu- date hardware performance. The results of the limited lations was difficult, and consequently only fair to good tests are reported in reference 13. To further demon- agreement between experimental and analytical data strate the feasibility and the potential of the series- was obtained. An overall assessment of the results from hydraulic active control landing gear, shaker tests, drop the investigation indicates that (despite the sus c eptibil- tests, and simulated-landing tests of passive and active ity to frictional forces of the gear employed in this in- versions of the general aviation airplane main gear were vestigation) the AGFATL computer program is a valid conducted (refs. 14 and 15).
tool for the study and initial design of series-hydraulic The logic and equations for the series-hydraulic ac- active control landing-gear systems, tive control gear developed in ACOLAG have been in- corporated in FATOLA to generate a computer pro- Introduction gram for multi-degree-of-freedom active gear, flexible aircraft take-off and landing analysis (AGFATL). (See In large airplanes , dynamic loads and vibrations ref. 16.) The purpose of this paper is to validate the resulting from landing impact and traverse of uneven AGFATL computer program with data from shaker runways and taxiways are recognized as significant fac- tests, drop tests, and landing-simulation tests of pas- tors in causing fatigue damage and dynamic stressing sive and active versions of the main landing gear from of the airframe structure. The ground-induced struc- the general aviation airplane.
tural vibrations also result in crew and passenger dis- comfort and, on large flexible airplanes, can reduce the Experimental Data pilot's capability to control the airplane during high- speed ground o perations. These problems ha v e been The experimental data for validating the AGFATL encountered with some currently operational transport computer program were obtained with passive and ac- aircraft, as discussed in references 1 and 2. Such prob- tive versions of a modified main landing gear from a lems will be magnified for supersonic-cruise airplanes general aviation airplane. The modified main landing be c aus e of the i n creased structural flexibility inherent gear, feedback instrumentation, and control hardware in their slender-body design, their thin-wing construc- for the active control landing gear are shown schemat- tion, and their high take-off and landing speeds. For ieally in figure 1. Closure of the hand valve permitted example, investigations of the ground-handling quali- passive operation of the gear by isolating the servovalve ties of one particular design of a supersonic transport from the modified gear. Shaker tests conducted by the conducted in the United States in the 1960's revealed Air Force Wrig h t Aeronautical Laboratories (ref. 14) extremely high vibration levels in the crew compart- compare the responses of the passive and active gears to ment during the take-off roll (ref. 3). One potential various sinusoidal and step bump profiles that might be method for improving ground operations of supersonic- encountered during the taxi / rollout mode. Drop tests cruise airplanes is the application of active control tech- and landing tests conducted at the Langley Aircraft nology to the landing gears to limit the ground loads Landing Dynamics Facility to demonstrate the feasi- applied to the airframe, bility and the potential of the active gear during touch- Analytical studies (refs. 3 to 5) have been conducted down impact and landing rollout are reported in refer- to determine the feasibility and potential benefits of ap- ence 15.
Shaker Tests provide s a comprehen s ive simulation of aircraft t ake-off and landing pha s es of operation wi t h conventional or The experimental setup for the shaker t e s ts is s hown serie s -hydraulic active gear s . Effec ts s imulated in the in figure 2. The shaker te s t s were conduc t ed for the taxi / rollout mode with the landing gear rigidly a tt ached program include (1) aircraft aerodynamic con t rol and performance during glide s lope, flare, landing, and t ake- to the 9 3 -slug ( 3 000-Ibm) drop- t ower bucket. The off when s ubjected to condition s s uch a s varying wind landing-gear strut axi s was orien t ed ver t ically, and the drop- to wer bucket was res t rained to vertical motion, speed s and direction s , engine failure s , brake failure s , landing-gear strut failures, control variable limit s , and Thi s restraint minimized the binding friction in the con t rol respon s e time s ; (2) landing-gear load s and dy- g ear. The s h aker was pr o grammed t o provide various namics for aircraft having a minimum of t hree gear s and haver s ine and step inputs to the gear t hrough the t ire. a maximum of five gears; (3) aircraft with a maximum of The haversine inpu t s consis t ed of five forcing cycles four engines; (4) s elec t ive engine rever s ing; (5) ground at frequencies varying from 1 to 20 Hz wi t h double effect aerodynamics; (6) drag chute, speed brakes, and ampli t ude s varying from 2.0 in. at the lower frequencies s poiler aerodynamic s ; (7) con s tant or skid-con t rolled to 0.5 in. at the higher frequencie s . The s t ep input braking; (8) t ake-off from or landing on runways or con s isted of one forcing cycle, a compressive step input, and (when strut motion sub s ided) a st ep return to the aircraft carriers; (9) inclined runway s and / or runway initial table posi t ion. The step initial input had a 0.5-in. elevation perturbation s ; (10) rudder steering, nose-gear s teering, or combined rudder and no s e-gear st eering; amplitude, and the amplitude wa s increased by 0.5-in. and (11) conventional oleo-pneuma t ic and / or series- increment s to a maximum of 4.5 in. or until either the hydraulic ac t ive control landing-gear shock s truts. The t ire or the strut bottomed during the compres s ive input, construction of the program i s modular so t ha t glide The forcing function was applied vertically in all case s slope, flare, landing, and take-off pha s es may be evalu- (no horizon t al input s ), ated separately or in combina t ion. The program al s o includes options for rigid or flexible airframe struc- Drop Tests tural characteri s tic s and pas s ive or serie s -hydraulic ac- The experimental setup for the drop tests and land- ti v e gear types.
ing t e s ts i s shown in figure 3. The drop tests of passive De t ailed analytical s tudies of aircraft ground opera- a n d acti v e versions of the modified main gear were made tions generally require a large amount o f computer oper- with a pitch at t i t ude of 0° (pi t ching beam locked in a ational time because of the large oscillatory variations horizontal posi t ion), zero gr o und speed, and t ouchdown in t he loads and motions of the aircraft. To alle v iate sink rate s from 3 to 5.5 fps. Since the gear longitudi- thi s problem, the AGFATL compu t er program also has nal axis was aligned vertically and no horizon t al forces s t aging and re st art capabilities. The stagin g capability were involved, these tes t s demons t ra t ed the effective- permi t s the user to change values of the input data to he ss of the ac t ive gear when strut-binding fric t ion wa s represent varia t ions of the input variables which may minimal, occur during the landing or take-off simulations. The restart capability permi ts the user to restart the pro- Landing Tests gram from any point in the t ime history where data have been s t aged in t o the program. This t echnique The landing t est s were conduc t ed for touchdown i s u s eful and efficient if the program fail s to execute pi t ch attitude s of the pitching beam from 2° to 13°, to a normal termination or for the analysi s of landing ground s peeds from 8 to 8 0 knots, and sink rate s from rollout over various simulated runway roughnes s con- 3 to 5.5 fp s . The landing test s provide a more realistic ditions. For example, to inve st iga t e runway roughness repre s entation of the l o ads and mo t ions imposed on an condi t ion s , the program can be restarted at a time in airplane than t ho s e obtained during ver t ical-drop te s ts, the simulation jus t prior to airplane encoun t er wi t h the The principal difference between these types o f test s de s ired runway roughnes s without rerunning the entire is that moment s re s ulting from vertical and horizontal landing simulation.
forces developed at the axle because of aircraft pi t ch attitude and wheel spin-up at t ouchdown induce s t ru t - Analy t ical Simula t ion Techniques binding fric t ion force s in the landing tests.
Since the AGFATL compu t er program is de s igned
AGFATL Capabili t ies for analyzing during ground opera t ions the loads and
motions of an airplane wi t h a minimum of t hree landing The capabili t ies of the active gear, flexible aircraf t gear s , the inpu t da t a were modified to simulate the t est t ake-off and landing analysis (A G FATL) computer pro- da t a for the s ingle gear. The following sec t ion s di s cuss gram, along with a s chema t ic representa t ion of the ac- the modifica t ions for the s haker t es t s, drop t es ts , and tire shock s trut, are shown in figure 4. The program landing t e sts .
S h ake r Tes t s bypassed by introducing a nominal positive vel o cit y value greater than the tire-ground-plane velocity. This Since no aerodynamic, control, or engine forces ex- procedure results in bypassing the table of friction-slip isted during the shaker tests, the aerodynamic, auto- ratios and in setting the longitudinal and transverse pilot, thrust, and engine subroutines of AGFATL were gr o und rea c ti on f o r c e c omp o nents to z e r o .
not called. A minimum of three gears is required in the The step bump tests were simulated by using the computer program; however, since only a single main restart capability of the AGFATL computer program.
gear was tested, the nose-gear attachment to the air- The time history of the landing was restarte d subse- frame was specified as a large negative value so the nose quent to the time at which the strut and airplane touch- gear would not contact the surface. Furthermore, since down impact motions had subsided. The staging capa- the drop-tower-bucket center of gravity and the single- bility was used to remove the runway sinusoidal pertur- gear strut longitudinal axis were aligned, this condition bations and to introduce step elevation perturbations was simulated by placing the two main gears in the to the runway surface.
transverse vertical plane containing the center of grav- ity of the airplane. Since it was necessary to use two Drop Tests main gears, the force applied to the airplane mass would be double the force of the single gear used in the shaker The drop tests were conducted with the gear at- tests. Therefore, t o obtain similar dynamic character- tached n o rmal to the pitching beam that was restrained istics, the mass of the simulated airplane was input as at a pitch attitude of 0°. The lift force was simulated twice the mass of the dr0p-tower bucket. (b y use of the staging capability) by introducing a gen- Angular degrees of freedom were not present during e ralized vertical-for ce v e ctor e qual to values of the lift the shaker tests; therefore , the pitch , yaw , and roll recorded during the tests. Since no ground speed was attitudes of the simulated airplane were input with zero involved in these tests, the following inputs were modi- valu e s , and significant changes in these attitudes were fled to simulate the drop test: the flight path angle was prevented by inputting very large moments of inertia input as -90 ° , the drop velocity at touchdown was in- about the pitch , yaw, and roll axes. put as the airspeed , and the pitch angle was input as 0 °" To simulate the sinusoidal and step inputs of the shaker, the runway elevation perturbation capability was used; however , the AGFATL computer program is Landing Tests not c apable of simulating vertical-force inputs to the The landing tests were conducted with the single gear at zero ground speed. In the shaker tests, the gear main gear attached to the pitching beam of the test rests on the shaker table and supports the mass of the fixture. The pitching beam had a mass of one-half drop-tower bucket; ther e for e , to us e th e runwa y eleva- the airplane mass (ref. 15). Aerodynamic lift and drag tion perturbation capability , the airplane must have a forces, aerodynamic elevator control force, and nose- ground speed with the gears supporting the airplane gear force were simulated in the test program; however , mass at the designed static deflection of the shock strut aerod y namic lift and drag for c es in the test program and the tires. To reach this condition , the simulated were not compatible with the aerodynamic coefficients airplane was placed on a glide slope with the main gear required as input to the computer program. Therefore , tires slightly above the runwa y with a ground speed of these forces were repr e sent e d in th e c omputer program 71 knots and a sink rate of 3 fps. To expedite reach- by generalized z-axis (lift) and x-axis (drag) body- ing the desired rollout condition , a large damping force oriented force vectors modified as a function of time was input to the gear struts to provide rapid damping by use of the staging capabilit y of the program.
of strut and airplane motions resulting from the im- The aerodynamic elevator control force and the pact phase. When the gear struts were supporting the nose-gear force (which were internal forces relative to airplane mass at the designed static strokes, the large the test fixture) were represented in the program by a damping force was removed from the struts (by use of generalized moment vector about the pitch axis. This the staging capability) , and the sinusoidal runway ele- moment vector was modified during the time history ration perturbations were introduced. To simulate the to reflect changes in the simulated test values of the 1.3-Hz forcing function of the shaker , five cycles of el- elevator and nose-gear forces.
evation perturbations with double amplitudes of 2 in. The presence of large binding-friction forces in the and wavelengths compatible with the 71-knot ground main gear strut during the landing tests was discussed speed were staged into the computer program, in reference 15. In addition, the test fixture was also The wheel had no rotational velocit y during the subject to frictional forces which would not be present shaker tests; therefore, the table containing ratios of during an aircraft landing. The light aircraft main ground friction to tire slip in the program had to be gear us e d in the test program reported in reference 15 was very susceptible to the generation o f large binding- loads on the mass from the gear. The circled numerals friction forces due to the large ratio of stroke to pis- are compressive peaks of the five cycles of excitation.
ton diameter. The gear was particularly susceptible to After the third cycle of excitation, the experimental large binding-friction forces during the touchdown im- accelerations have reached limit values of approximately pact phase, since the spacing between the shock-strut -2.25 g and l g . The value of -2.25g is attributed bearing surfaces is a minimum when the strut is fully to a force from a snubber employed in the gear to extended. Also, this gear had previously been subjected prevent damage at compression bottoming. The l g to severe test conditions (ref. 13)that resulted in large limit occurs as the gear fully extends and the tire bending moments and elastic deformations of the gear, leaves the shaker table. The analytical accelerations which may have caused bearing wear and aggravated increased throughout the excitation cycles and for one the susceptibility to binding friction, cycle beyond excitation but did not reach a limit, The 'computer program AGFATL simulates the probably because the snubber force was not simulated strut-binding friction resulting from tire spin-up and in the analysis. However, the computed accelerations encounters with changes in runway elevation but is not did reach the l g limit after the third excitation cycle.
readily adaptable to binding friction resulting from de- Subsequent to the five excitation cycles, the experi- formations in the strut due to wear. The program also mental data show that the limit accelerations continue simulates the Coulomb type of friction resulting from for approximately two cycles before appreciable damp- the fit of the bearings relative to the cylinder and the ing occurs. The computed data show that the reso- piston. However, a smoothing technique (ref. 1 2 ) is in- nant effect continues for approximately one cycle be- cluded to prevent sudden changes in magnitude and fore damping is evident. The experimental data also direction of the friction force when the strut velocity exhibit greater damping than that which occurs with passes through zero during changes in direction, the computed data.
Analytical simulation of the strut and test fixture Experimental and analytical strut strokes resulting frictional forces was accomplished by staging changes to from the same five cycles of excitation are shown in fig- the strut Coulomb friction and the generalized vertical- ure 5(b). Z ero stroke is the static stroke of the gear force vector to represent strut-binding friction and test (approximately 5 in.) required to support the mass of fixture friction, respectively. Since these frictional the drop-tower bucket. The effect of resonance is again forces were not separable from the inertia forces mea- very graphically illustrated by both the experimental sured in the test program, a trial-and-error method of and computed data. After approximately two excita- including these forces had to be employed in the ana- tion cycles, the experimental data show that the strut lytical simulation, is stroking between the limits of maximum compression (influenced by the snubber) and maximum extension (5-
Results and D iscussi o n in. negative stroke). The computed strokes do not reach
a compressive limit but increase during the five excita- To validate the A GFATL computer program for pre- dicting the loads and motions of aircraft with conven- tion cycles; however, on the third excitation cycle, the tional passive gears and / or series-hydraulic active con- gear is fully extending (5-in. negative stroke). For two trol gears, analytical results are compared with data cycles following the excitation, the experimental data show that the gear continues to stroke between the lim- from shaker tests, drop tests, and landing tests of its of maximum compression and maximum extension a modified single main gear from a general aviation airplane, before damping occurs. On the other hand, the com- puted compressive strokes increased for one cycle and continued to fully extend for two cycles before damping Shaker Tests reduced the strut stroke.
In figure 5, time histories of drop-tower-bucket mass The higher-than-computed negative experimental acceleration and strut stroke from the shaker tests accelerations were attributed to the greater experimen- are compared with those computed with the AGFATL tal compressive strokes (fig. 5(b)), which generated a program, snubber force that was not simulated in the analy- sis. Furthermore, the larger experimental compres- Passive gear response t o excitation near resonant sive strokes probably resulted from friction between the frequency . Figure 5(a) compares experimental and drop-tower bucket and the guide rails that also was not analytical acceleration responses of drop-tower-bucket simulated in the analysis. In spite of these differences, mass to a forcing frequency near the 1.25-Hz strut the analytical simulation very graphically illustrates the resonant frequency. The applied forcing function was gear response to excitation near gear resonance, and the a 1.3-Hz, 2-in-double-amplitude, five-cycle sinusoidal computed and experimental response data are in good waveform. N egative acceleration represents upward agreement.
A ct i v e g ear response to exci tation near r esonant h i gh p r o babil i t y of fr iction e x i st ing betwee n t he d r o p- fr equency . E x p e rim ent al and analy t ical acc e l e ra t ion s t o wer b uck et and guid e ra i l s mak es the d i s par it y l ess of the b uck et ma ss ar e p r ese n te d in fi gur e 5 (c ) for the a c - s igni fic an t.
t ive g e ar r es pon se t o an e x ci t a t ion fr e q u ency near g e ar r es onan t fr e q ue ncy. Th e fi v e cycl es of the forci n g func- Passive gear response to 2 . 5 - i n . step bump .
t ion ar e indica te d b y t h e c ircl e d nu m e ra ls. T he m o st F igur e 5 (e ) pr ese n ts ex perim e n t a l and analy t ical drop- o b viou s c h arac te ri st i c of these da ta rela t iv e t o the pa s - t o we r- bu ck et ma ss acc e l e ra t ion s r es ul t i n g from the r e - s iv e g e ar da ta (fi g. 5 (b)) i s t h e a bse nc e of any r es onan t eff ect . T he ex p e rim e n t al and analy t i c al ac ce l e r at ion s s pon se of the pa ss iv e g ea r t o a 2 . 5-i n - h ig h s t e p b ump.
Th e ex perim e n t al acc e l er a t io n i s gr e a te r th an the ana- for th e ac t ive g e ar are limi te d t o appro x i m a te ly 0.1 g and 0 . 2 g , r es p e c t iv e ly . F ollo w ing the ex ci t a t ion, the ex - ly t ical during the ste p-up, t h e r eb ound, and t h e com- p r ess ive p o r tio n of the fir st os cil l a t io n. Th e d amping in p er im e n t al r es pon se i s ra p i d ly a tte nua te d, an d a s w a s the ex p e rim e n t a l test appara tus w a s gr eate r th a n t h e the ca se w i th the pa ss iv e g e ar, the damping pr ese n t i n the ex p e rim e n t w a s gr e a te r tha n th a t analy t ically damping i n t h e analy s i s , a s e vid e nc e d b y t h e fac t t ha t s i mu l ated, the experi m enta l a c celerat i o n had dr o pped t o zero a t a For the a c tive gear, experimental and analytical time of about 6 seconds, but the computed acceleration strut stro k es resulting from the five-cycle, 1.3-Hz, 2-in - was still oscillating. Even though the analyti c al a c cel- double-amplitude excitation are shown in figure 5(d). eration was lower than the experimental in response to The five ex c itation cy c les of the forcing function are in- the step-up and the first oscillation, the c haracter of the dicated by the circled numerals. The resonant buildup response was the same.
For the step - down input , the analytical acceleration i n respo n se which occurred with the passive gear is ab- sent, and the rapid attenuation of strut response after was greater than the experimental be c ause the com - puted a cc eleration was still oscillating from the step-up the last excitatio n c y c le is obvious for both the exper- i n put, but the experimental a cc eleration had damped imental and analytical strokes, although the analyti c al to zero. During the first rebound oscillation, the exper- data indicate less damping. The experimental compres - imental acceleration was again greater than the c orn- sire strut strokes are generally greater than the analyti- puted, but following the step inputs, the magnitude and cal simulation , and c onversely , the experime n tal strokes the frequency of oscillation (1.3 Hz) was the same for during gear extension are less than the analytical ex- both the experimental and computed data.
tension strokes. The for c es applied to the drop-tower bu c ket by the active gear are small , and fri c tion between Passive gear experimental and analytical strut the bu c ket and guide rails c ould be a dominant fa c tor, strokes in response to a 2.5-in-high step bump are shown in figure 5(f). The experimental strut stroke during the For example, if the fri c tion force was large relative to the gear for c e, the displa c ement of the bu c ket would step-up input and strut extension during rebound were be smaller and the gear would stroke more. Also, if greater than those predicted by the analysis. As pre- the displacement of the upper mass was restrained dur- viously dis c ussed, the friction for c e between the drop - tower bucket and guide rails could result in greater ing the compressive stroke of the gear, gear extension would be r estrained by the shaker table displacement, strokes than would o cc ur with the unrestrained vertical motion in the analytical simulation. The same hypoth- The experimental data indicate that this may have been esis would apply to the strut stroke response to the the case; however, no bucket displacement data were presented in reference 14 t o substantiate this be h avi o r, s t ep-down input.
Fri c tion between the drop-tower bucket and guide rails The maximum experimental and analyti c al acceler- was not simulated in the analysis, and the bu c ket was ations are different during the step-up input, rebound, unrest r ained in vertical motion. Therefore, during corn- and c ompressive portio n of the first oscillation; however, pressive stroking of the gear, the bucket would displace the chara c ter of the accelerations is the same.
more and hen c e reduce the gear stroke. Conversely, be- The experimental and analytical strut strokes are cause o f the greater bucket displaceme n t, the gear w o uld different, but the presence of friction between the dr o p- extend to a greater stro k e even though it was restrained tower bucket and guide rails during the experiment may by the shaker table displa c ement, account for the difference.
Both the experimental and analytical active gear re- sponses to five cycles of excitation de m onstrate the ef - A ctive gear r esponse to 2 . 5 - in . step b ump . The fectiveness of the active gear in eliminating gear reso - experimental and analytical drop-tower-bucket acceler - nant r esponse. Excellent agreement was noted betwee n atio n s f o r the ac t ive gear are shown in figure 5(g) fo r the experimental and computed accelerations of the response to a 2.5-in-high step bump. The agreement bucket mass. Although the agreement between experi- between experimental and computed accelerations for mental and computed st r ut strokes was not as good , the the step inputs is excellent. H owever, as has bee n the case for all the shaker tests, the damping in the test perimental pressures were generally higher than the apparatus is greater than that used in the analysis, computed values.
Figure 5(h) presents experimental and'analytical shock - strut strokes for theactive gear when subjected to Dr op test of active gear . Figure 6(c) presents a 2.5 - in-high step bump. The experimental strokes are comparisons between experimental and analytical mass - greater than the computed strokes, and following the center forces and shock - strut strokes for the active gear step - up and step - down inputs , the experimental static drop test. The agreement between the experimental stroke is biased relative to the designed static stroke to and analytical mass - center forces is good. The agree - a greater compressive stroke. Following the step - down ment between the experimental and analytical shock - input, the experimental stroke data quickly return to strut strokes is good during touchdown impact and re - the biased static stroke, but the computed stroke re - bound, but the maximum computed stroke is greater quires a greater time to return to the static stroke, than the experimental stroke during secondary impact These discrepancies may result from the frictional ef- and damping to static deflection. This difference may fects previously discussed, deviation in shaker perfor- be caused by the relatively small differences in mag- mance, or differences in the response of the analytically nitude and timing of the forces which occurred dur- simulated controller relative to those which occurred in ing the touchdown impact and rebound phases of the the experiment, simulation.
The agreement between experimental and analytical Experimental and analytical strut-hydraulic pres- accelerations of the drop-tower bucket in response to sures and servo-spool displacements are shown in fig- the 2.5-in. step bump was excellent. The agreement ure 6(d) for the active gear drop test. The general between experimental and computed strut strokes was trends of the experimental and computed pressures and not as good, but the trend of the stroke data in response servo-spool displacements are in good agreement. How- to the step-up input was the same. The reason for the ever, the maximum value of the computed pressure is difference between the strokes following the step-down greater than the experimental pressure during touch- input is not understood, down impact and is less than the experimental pressure In summary, the agreement between experimental during rebound. These discrepancies may be attributed and analytical data obtained for the shaker tests of to small differences between response of the servovalve the passive and active gears was good despite some used in the experimental program and that simulated differences, which may be attributed to snubber or in the analysis. For example, the computed servovalve frictional forces in the experimental apparatus that were displacement had greater positive and negative values not represented in the analytical simulation, during rebound than those which occurred during the experiment.
Dr op Tests In summary , alth o ugh t h ere are differences between Comparisons of experimental and analytical results the experimental and analytical data obtained from the obtained for vertical-drop tests of the passive and active drop tests of the active gear, the agreement between experimental and analytical data is good.
gears are shown in figure 6 for a ground speed of zero, a sink rate of 5.5 fps, and a pitch attitude of 0 ° . The strut-binding friction is a minimum for these touchdown Landing Tests parameters. However, friction forces were present be- Comparisons of experimental and analytical results tween the drop frame and the standoff structure used for landing-simulation tests of the passive and active in the experimental investigation of reference 14. The gears are shown in figure 7. The tests were conducted tests analytically simulated for the passive and active for a ground speed of 80 knots, a touchdown sink rate of gears were tests 49 and 51, respectively. 5.5 fps, and a pitch attitude of 2°. The analytical simu- lations for the passive and active gears are tests 40 and Dro p test of passive gear . The experimental and 42 of reference 15, respectively. The inclusion of friction analytical mass-center forces and shock-strut strokes for forces in the analytical simulations of these tests was a the vertical-drop test of the passive gear are shown in very difficult task, since these forces were not specifl- figure 6(a). The computed mass-center forces and strut cally defined during the test program. In addition, the strokes are in good agreement with the experimental strut-binding friction, which was of considerable mag- data during touchdown impact, rebound, and secondary nitude as evidenced by the fact that the strut would impact. Experimental and computed strut-hydraulic stop stroking during the period of maximum stroking pressures shown in figure 6(b) are in good agreement velocity, had to be applied in conjunction with the fric- during touchdown impact and rebound. During sec- tion forces of the test fixture and the moments from ondary impact and damping to static pressure, the ex- the elevator control and nose-gear force simulators. As a result, these forces and moments had to be staged into Landing test of active gear. For the active gear, the program by a trial-and-error method. figure 7(c) compares experimental and analytical mass- center forces and shock-strut strokes, and figure 7(d) compares strut-hydraulic pressures and servo-spool dis- Landing test of passive gear. Figure 7(a) com- placements. The experimental and computed forces, pares experimental and analytical mass-center forces strokes, pressures, and servo-spool displacements were and shock-strut strokes for the touchdown impact phase in good agreement during the compressive phase of the of the landing. The agreement between the experimen- initial impact, with slightly less agreement during the tal and analytical mass-center forces was excellent. The rebound phase. The experimental stroke data during agreement between experimental and computed shock- rebound indicate that large cyclic frictional forces (side- strut strokes was not as good, but the characteristic by-side symbols) were generated which were not analyt- shape of the analytical data was the same as that of the ically simulated. Consequently, the gear extends at a experimental data. The differences between stroke data greater rate analytically than it did experimentally, pro- resulted from the difficulty of introducing the proper duces an earlier transition of the servo-spool displace- proportions of strut-binding and test fixture friction ment from removing fluid to adding fluid, and results forces and elevator control and nose-gear moments. Ad- in a greater analytical shock-strut pressure. These dif- ditional factors, which influence the strokes during the ferences during rebound resulted in greater differences latter stages of the computed time history, are small er- between the experimental and analytical data during rors in simulating forces and moments during the initial secondary impact and transition to the static stroke.
impact phase, which result in cumulative errors dur- Despite the difficulties with analytically simulating ing subsequent portions of the time history. This is friction forces, the agreement between experimental and indicated by the differences in strut strokes during sec- computed data during the compressive phase of the ini- ondary impact and transition to static stroke. Despite tial impact was good. As a resuIt of the inability to the unusually large friction forces encountered with this analytically simulate the strut friction during the re- gear during the experiment and the related difficulties bound phase of the initial impact, particularly for the encountered during the analytical simulation, the com- active gear in which the servovalve control responds to puted data represent the dynamics of the system very force fluctuations, only fair agreement was obtained be- well during the touchdown impact phase of the landing.
tween experimental and analytical data during rebound, Experimental and analytical mass-center forces and secondary impact, and transition to static stroke.
shock-strut strokes for the passive gear during traverse of the step bumps are shown in figure 7(b). The force
Concluding Remarks
and stroke time histories are plotted from an arbitrar- ily selected time (zero time in the figure) before en- A computer program for multi-degree-of-freedom counter with the first bump through traverse of the sec- active gear, flexible aircraft take-off and landing anal- ond bump. Since the gear was stroked to the static po- ysis (AGFATL) was developed for studying series- sition prior to encounter with the bumps and the gear- hydraulic active gears. To validate the AGFATL com- bearing spacing was greater, the strut-binding friction puter program for predicting the loads and motions of force was much smaller than that which occurred dur- aircraft with conventional passive gears and/or series- ing the touchdown impact phase. The friction force hydraulic active control gears, analytical results are between the drop frame and the standoff structure compared with data obtained from shaker tests, drop was still present and presumably of similar magnitude.
tests, and landing tests of a modified single main gear No attempt was made to analytically simuIate friction from a general aviation airplane.
forces due to gear wear or test fixture during traverse of Comparison of experimental and analytical the bumps. Although the computed mass-center forces, responses for both passive and active gears indicates both positive and negative, were greater than the exper- good agreement for shaker tests (forced vibrations) and imental values during the bump encounters, the general drop tests (although there were some differences). For shapes of the experimental and computed force time his- the simulated-landing tests, the passive and active gears tories were the same. The incremental changes in the were influenced by large strut-binding friction forces.
experimental and computed shock-strut strokes and the The inclusion of these undefined forces in the analyti- shapes of the stroke time histories were the same. Al- cal simulations was difficult, particularly for the active though no attempt was made to simulate strut-binding gear in which the servovalve control responds to force and test fixture friction forces, the agreement between fluctuations; consequently, only fair to good agreement mass-center forces and shock-strut strokes for the pas- was obtained between experimental and analytical re- sive gear was good during traverse of the step bumps. sponses for the landing tests.
An overal l assessment of the results from the investi- 7. L ynch, U r ban H. D.; a n d Du e weke , J oh n J .: Tak e offand gati o n indicates that ( d espite the susc e ptibility to fric- LandingAnalysis (TOLA) C o m put e rP r ogra m. Part II -- P ro bl em tional forces of the gear e mployed in this investigation) Fo r mu!a t ion _ AFFDL-TR - 71 - 155, Part II, U.S. Air Force,May 1974.
the A G FATL computer program is a valid tool for the 8. Lynch, Urban H. D.; and Dueweke, John J.: T ake off and study and initial design of series- h ydraulic active con- Landing Analysis Computer Pro gr a m (TOLA ) . Pa r t Ill-- Us e r ' s trol landing-gear systems. Manua l AFFDL-TR- 7 1-1 5 5, Part III, U.S. Air Force, Apr.
1974.
9. Y o ung, Fay O.; and Dueweke, J ohn J.: T ak eoff and Land - Langley Researc h Center i ng Analysi s Compute r Prog r a m (TOLA) . Pa r t IV -- Prog r a mme r's National Aeronautics and Space Administration Manual . AFFDL-TR-71-155, Part IV, U.S. Air Force, Jan.
19 7 5.
Hampton, VA 23665 10. Dick, J. W.; and Benda, B. J.: Addition of Fl e xibl e Body Option July 13, 1984 t o th e TOLA Compu t e r P r o gr am . Pa rt I -- Fina! Report . NASA CR-132732-1 , 1975.
11. Dick, J. W.; and Benda, B. J.: Addition of Flexibl e Body Op tion t o th e TOLA C omput er P r og r a m. Part II -- U ser and P r o g r a mmer Docum e ntation. NASA CR-132732-2, 1975.
Re f ere n ces 1 2 . Car d en , H uey D .; a nd M cG e h e e , J o h n R . : Valida t ion o f a Fle xible Aircra ft Take - Off and La nding An al ysis (FATOLA) . N A S A 1. DC-10 Landing Gear M o dified. Aviat . Week _ J Space Techno L, TP-1025, 1977.
vol. 98, no. 12, Mar. 19, 1973, p. 181. 13. Ross, Irving; and Edson, Ralph: An El e ct r onic Cont ro l fo r an 2. Ropelewski, Robert R.: Airbus Test Tempo Quickening. Elect ro hyd r aulic Ac ti v e Cont r ol Ai r c r aft Landing Gea r. NASA CR- Aviat . Week _ Space Te c hn ot, vol. 98, no. 10, Mar. 5 , 1973 , 3113, 1979.
pp. 32-35. 14. Morris, David L.: Active La nding Gea r Respon se Testing .
3 . W ig'not , Jack E.; Durup , Paul C .; and Gamon , Max A.: D e- AFWAL-TM - 82 - 177-FIEM , U.S. Air For c e , Apr. 1982.
sign Fo r mulation and Analysis o f an Ac tiv e La nding G ea r. Vol - 15. McGeh e e, John R.; and Dreher, Robert C. : Ex p er i me ntal u me I . An al ysi s . AFFDL-TR-71 - 80, Vol. I, U.S. Air Fo rc e, A ug. 1971 . (Available from DT I C as A D 88 7 1 27 L . ) Inves ti gation o f Ac t iv e Load s Co nt ro l fo r Air c raft La nding Gea r.
NAS A TP-2042, 1982.
4. Bender, E. K.; Berkman, E. F.; and Bieber, M.: A Feasibil it y 16. Mall, G erald H.: I m p r ove me nts to the FATOLA Co m p uter P r o - Study o f Activ e Landing Gea r. AFFDL-TR-70-126, U.S. Ai r gr a m In c luding Added Ac ti vely C ontrolled La nding Gear S ubrout i nes .
Force, J ul y 1971. (Available from DTIC a s AD 887 4 51L.) NASA CR-166069, 1983.
5. McGehee, John R.; and Ca x den, Huey D.: A Mathe mat ica l Model of an Ac ti ve Co ntr ol Landing Gear for Lo ad Control During Impact and Roll - Out . NASA TN D-8080 , 1976.
6. Lynch, Urban H. D.: T ake off and Landing An al ysis (TOLA) Co m pv J e r P r o g r a m. Pa rt 1 -- Capabilitles o f the T ak eoff and La nding An al ysis Co m put er P r o gr a m. AFFDL-TR-71-155, Part 1, U.S.
A i r Force, Feb. 197 2 . ( Av a ilable from DTI C as AD 741 942.)
-ACCELEROMETER
LAND l
GEAR
STRUT
RESERVO l R ACCUMULATOR
Figure 1. Schematic of active control landing gear employed in experimental program.
"0 E o_ 4 -1 e,i U-
_"-"---_:__To -- o r
IMPACT RUNWAY IMPACT
LOADS INDUCED LOADS LOAD S
FLEXIBLE AIRFRAME
OPERATIONAL LOADS CHARACTERI STICS
LOADISOLATION
• THRUSTREVERSAL AND
ANTI SKIDB R AKING
• AERODYNAMIC
• STEERING
• A S YMMETRIC • MODAL MASS
• MODAL FREQUENCIES
• MODAL DISPLACEMENTS
F i gure 4. Capab i l i t i es of act i ve gear . flexib ( e aircraft take-off and #anding analysis.
-3m -_- Experimental Analytical _ Forcing function
-1 @
acceleration, _ g units
z I I ! I I
0 2 4 6 8 I0 Time, sec (a) Mas s a cc el e r a ti ons f o r pas si ve gea rt es t e d a t a f re qu en c y o f 1 . 3 H z an d a doubleampl i t ud e o f 2 i n .
Fi gu r e5 . Ex pe ri menta l an d ana l yt ic a lr e s ult s f o r sh ake r te s t s.
4 -- Forcing (_) (_) (_ -'(}'- Experimental function (_ (_ _ _) (_ _ A n alytical ) Stroke at which contact with snubber occurs 2-- ) I
0 >
i O C _ _, I Strut stroke, in. I t Y !
¥ o
!
i t
-6 I I I I I
0 2 4 6 8 10 Time, sac (b) Strut strokes for passive gear tested at a frequency of 1.3 Hz and a double amplitude of 2 in.
Figure 5. Continued.
-3- --0-- Experimental Analytical -2 -- -I- Mass acceleration, g units C) (_) (_) (_) C) Forcing function
2 I t I l l
0 2 4 6 8 10 Time, sec ( c ) Mass a cc el er ations fo r a ct ive gea r tested at a f r equ e n c y of 1.3 Hz and a double amplitude of 2 in.
Figure 5. Continued.
4- --0-- Experimental Analytical (_ (_) (_) (_) (_ Forcing function 2-- # Strut stroke, in.
-2 -4m
-6 I I , I ,I I
0 2 4 6 8 10 Time, sec (d) S t rut st ro kes fo r .active gear tested a t a f r e qu e nc y of 1 .3 H z and a double ampli t ud e of 2 in.
Figure 5. C o n ti n ued.
- 1.0 - -43- - Experimenta 1 Step-up _ Analytical 9 , _ ,, !
-.5 Mass acceleration, O( g units
.5- _ t
Step-down
1.o I I _I I I
0 2 4 6 8 10 Time, sec (e) M ass ac c e lerations fo r passiv e gea r subje c ted to 2 .5-in- h ig h s te p bu mp .
Fi g ure 5. C o n tinu e d.
- -(Z_- Experimental Analytical Step-up Str u t str o ke, in.
_ { { { { t
0 2 4 6 8 10 Time, sec (f) S t rut st r ok e s for pass i ve g e a r s u bjec t ed to 2.5- i n-h i g h step b u mp.
F i gu r e 5. Continued.
- 1.0 _ -_Z)- - Experimental Analytical Step-up -.5 - Mass acceleration, O ( g units
t
• 5 -- Step- down
I I I I I
I. 00 2 4 6 8 I0 Time, sec (g) Mass acc e l e rati ons f o r acti ve g e a r s ub je ct e d t o 2. 5- i n- hi g h s t ep bu mp.
Fi g ur e 5. C o n ti n u ed .
1 9 m --(}- Experimental Step-up _ Analytical
2-
I I
O - - 0 - --0- O- _ -( > . r _ 0- < )-_)
Strut str°ke' 0'! (in. V V V _ _' i z _ j_ v v
St e p-down
-6 I I I I I
0 2 4 6 8 10 Time, sec ( h ) S t ru t st r ok e s f or a c t i ve gea r su b je c ted t o 2.5- i n- hi g h step bump , Fi g ure 5. C o n clud e d.
2 O _8.__x103 0 Experimental -4i i_ _Analytical force, I bf I 0 Mass-center _. 0 / /0__ O_
_ _-'_-
4 I I I I
m Strut stroke, 4i - in.
_ d , ..y I I I i
0 .4 .8 1.2 1.6 2. 0 Time, sec (a) Mass-center forces and shock-strut strokes for pass i ve g ear. Test 4g of reference 14.
Figure6. Experimental andanalyticalresultsfrom vertical-drop tests. Ground speed.0 knots;touchdo w n s i nkrate.
5,5 fps. pitch attitude, 0 °.
2 1 12 , ._x102 O Experimental _Analytical m pressure, psig Strut-hydraul ic A
4 _ _o ° o- o o oo o o o o o
I I I I I
0 .4 .8 1.2 1.6 2.0 Time, sec ( b) S trut - h y draulic p ress u r ef or p assi ve g ear. T est 49 of re f ere n c e14 .
Fi g ure 6. Co n tinued.
x 103 0 Experimental Analytical force, Ibf Mass-center -4 0 _ _ --
V
4 ! I I I !
Strut stroke in. ' 4
o 2
- -co- I I I I
0 .4 .8 1.2 1.6 2 .0 Time, sec (c) Ma ss - c en t er f orc e s an d sh ock- strut s tr oke s for a c t i ve gea r . Te st 51 of r eferen c e 1 4.
Fi g ur e 6 . Co n ti n u ed .
x 102 12 O Experimental Analytical 8 0 pressure, psig Strut-hydraulic _O
cP
o I I I I I
.2 -- Servo-spool (_v v -_A_ r _. _ _ r _ _ r _ displacement, in. 0(_ _ "_ 00,_,.0._- v - v _... . -
-. 2 I I I I I
0 .4 .8 1.2 1.6 2.0 Time, sec ( d ) St ru t- h yd rau li c p re ss u re and se r vo-spo ol disp la cem ent fo r a c t iv e ge a r . T e s t 5 1 o f r ef er en ce 1 4 .
F ig ur e 6 . Conc l ud e d.
-8 x 103 0 Experimental Analytical -4 Mass-center force, Ibf O'
4 I I I I I
m Strut stroke in.
' 4:m 000 0" _ 0 00 O0 O0 0 O0
I I I
0 .4 .8 1.2 1.6 2.0 Time, sec (a) Mass-center fo rc es and s ho c k-s trut s t r okes fo r pas s iv e g e a r d ur ing t o uc hdown impa ct. Tes t 40 of re fe r e nce1 5.
F i gure 7. Experimental and analytical results from land i ng tests. Ground speed, 80 knots; touchdown sink rate . 5.5 fps; pitch attitude . 2° .
2 5 x 103 0 Experimental _Analytical
-2 0
Mass-center force, Ibf 0
3 I I I I J
0 o Strut stroke, in.
I I I I I
0 .4 .8 1.2 1.6 2.0 Time, sec (b) M a ss - ce nt e r f orc e s a nd s ho c k- strut str o ke s f o r pa ssi v e gea r d uri ng tr ave rs e of st e p b u mp s. T es t 40 o f r ef ere n ce 1 5 .
Fi g ure 7 . C o n ti n u e d.
2 6 -8,x 0 Experimental I _Analytical -4 Mass-center force, I bf O(
0 o
Strut stroke, in,
I I I
O' .4 .8 1.2 1.6 2.0 Time, sec ( c ) M ass < e nt e r f orce s an d s h o c k- strut s t r okes for acti ve ge ar duri ng t o uch do w n i mpa c t . Tes t 4 2 o f re fe r en c e 1 5 .
Fi g ur e 7 . Conti n ued.
12 x 102 0 Experimental _Analytical Strut-hydraul ic pressure, psig
4 _ObO
I I I
.2 Servo-spool 0 00_.) _0_01 _ displacement, in.
_.20 I I I I I
.4 .8 1.2 1.6 2.0 Time, sec (d) Strut - h yd r a u lic p re s sure and servo-sp o o l displa c emen t f or a c t i ve gea r du ri ng t ou ch down impa ct . T es t 42 of re fe re n c e 15.
Figure 7. Concluded.
2 8 1 . Report No . 2 . Government Accession No. 3 . Recipient ' s Cata l og No .
NASA TP - 2353 4 . TiUe a nd Subt i tle 5. Re po rt D ate VALIDATION OF AN ACTIVE G EAR, FLEXIBLE September 198 4 AIRCRAFT TAKE-OFF AND LANDIN G ANALYSIS 6. Perto r m ingOrg a nization Code ( AGFATL) 505 - 45-14-01 7 , Author( s} 8 . P erfor m ing O rgan i zatio n Report No.
John R. McGehee L-15807 10 . Work Unit No .
9. P er f orming Organization NameandAdd r ess NASA Langley Research Center 11 . Co ntractor G rant No.
Hampton, VA 23665 13 . Typeo f Report andPeriod Co v ered 1 2 . Spon s oring Agency Nameand Address Technical Paper National Aeronauti c s and Spa c e Administration 14 . S po nsoring A ge ncy Co de Washington, DC 20546 '1 5. Supplementary Notes 16. A b s t rac t T h e results o f a n a n alyti c al investigation using a computer program for active gear, flexible aircraft take-off and landing analysis (AGFATL) are compared with experimental data from shaker tests, drop tests, and simulated- landing tests to validate the AGFATL computer program. Comparison of experimental and analytical responses for both passive and active gears indicates good agreement for shaker tests and drop tests. For the simulated- landing tests, the passive and active gears were influenced by large strut-binding friction forces. The inclusion of these undefined forces in the analytical simulations was difficult, and consequently only fair to good agreement was obtained. An assessment of the results from the investigation indicates that the AGFATL computer program is a valid tool for the study and initial design of series-hydraulic active control landing-gear systems.
i 17. Key Words ( S ug ge sted b y Au t hor i s)) 18 . D i s tribution Statement Air c raft landing gear Unclassified - Unlimited Acti v e c o ntrols Landing loads Landing analysis Subject Category 05 19 . Security Cl a ssif, (oft h isre por tl 20. Se c urity C la ssif, (of thispa ge ) 21 , No, of Pages 22. Price Unclassified Unclassified 29 A03 For s a le bytheNational Technical In f orm a tion S ervice , Sprinp. f ield , Vir g inia 22 1 6 1 N ASA -La ngle y , 1984 National Aeronautics and THIR D- CLAS S BULK RATE Po s ta ge and Fees Paid OlJC - ace A dministration _ N a tion a l Ae ron a uti cs a nd Sp a c e Adm i ni s tr a tio n NASA-451 Washington, D.C. ' 20546 3 O fficial Bus i ness P e nal ty fo r Priva t e U s e , $300 NASA POSTMAS T ER: If Unde l ive r ab l e (Sec tio n 1 5 8 Po s t a l Manua l ) Do N ot R e tur n • , L . ..............
li
D O N O T REM O VE SLI P F R O M MATERIAL tothe li brar y .
D e l e t e y our nam e f rom th i s s li p w h enr e turnin g mater i a l N AME MS I [ f NAS A lg l e y R I AD N - 75 :