NASA NACA-TN-2948-1953 Investigation of lateral control near the stall flight investigation with a light high-wing monoplane tested with various amounts of washout and various leng.pdf

1、.Imi#.NATIONALADVISORYCOMMITTEEFOR AERONAUTICSTECHNXCAL NOTE 2948INVESTIGATION OF LATERAL CONTROL NEAR THE STALLFLIGHT INVESTIGATION WITH A LIGHT HIGH-WING MONOPLANETESTED WITH VARIOUS AMOUNTS OF WASHOUT ANDVARIOUS LENGTHS OF LEADING-EDGE SLOTBy Fred E. Weick, Maurice S.SeveLson,James G. McClure, an

2、d Marion D. FlanaganAgricultural and MechanicalWashingtonMay 1953. .! :-.: ,. v,*”.*+.-Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-TECHLIBRARYKAFBmNM “-III. .111111111111111111111111mrIomiL Mmsom cowm FOR AERoNAUTICS 00bb175TECHNICAL NOTE 2948INV

3、ESTIGATION OF LATERAL CONTROL NXAR THE STALLFLIGHT INVI!STIGATIONWITH A LIGHT HIGH-WING MONOPLANETESTED WITH VARIOUS AMOUNTS OF WASHOUT ANDVARIOUS LENGTHS OF LEADING-EDGE SLOTBy Fred E. Weick, Maurice S. Sevelson,James G. McClure, and Marion D. FlanaganSUMMARYFlight testspossibilities forwere madeob

4、tainingwas found that satisfactoryunder conditions simulatingapproximately 2 below thatwith a typical light airplane to investigatereliable control at low flight speeds. Itlateral control occurred consistently, evenetiremely gusty air, at angles of attackfor the maximum lift coefficient (or thestall

5、 of the wing as a whole). This 2 margin was substantiallythesame both with full power and with the engine throttled and throughoutthe rqe of center-of-gravitylocations tested. Supplementary testswere then made on the control at high angles of attack under actualgusty air conditions, on the possibili

6、ty of entering spins, snd on theamount of elevator control required for normal three-point landings. Itwas found that with the original plain untwisted wing obtaining theconstant 2 margin below the stall required widely different elevatordeflections for thetesteal. Also, nonepoint lsmding.An attempt

7、 wassufficient elevatorrsmge of power and center-of-gravitylocationsof these settings was high enough to produce a three-then made to find a configuration that would providecontrol for a three-point lending under the mostcritical condition (forward center of gravity) and that at the ssm.etimewould h

8、ave insufficient elevator control to exceed the angle of attackat which reliable lateral control is obtained in flight under all of thecenter-of-gravity and power conditions. The entire series of tests wasrepeated with the wing twisted to 4 md to 8 of washout snd with fivedifferent lengths of leadin

9、g-edge slots covering the outer 30, 50, 60,70,and 90 percent of the wing span. With 8 of washout the aileroncontrol itself was satisfactoryunder all conditions tested, even atangles of attack well beyond that for the airplane maximum lift coeffi-cient. Longitudinal fluctuations occurred, however, at

10、 ell angles ofProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-2 NACA TN 2948attack above that for the initial stalling of the center of c wing.The results for the 30-percent slots were the sane aS those withoutslots. With all of the other slot config

11、urationslateral control wasmaintained at high angles of attack, but severe longitudinalfluctuationsoccurred at angles of attack above that for the stall ofthe plain wing.It was determined that the longitudinal fluctuations were caused byburbling over the upper surface of the wing at the center where

12、 it isalso the upper surface of the fuselage. The fluctuations were eliminatedby the use of a full-span slot. The slat was extended over the fuselageand modified in cross section to adapt it to the fuselage contour. Withthe full-span slot the angle for maximum lift coefficient was increased 6.The de

13、sired condition, that is, having sufficientup-elevator controlto accomplish three-point landings but insufficient to exceed the angleof attack for satisfactory lateral control, was attained under limitedconditions with both the case of the 8 of washout and the case of thefull-span slot. In both case

14、s the desired condition was attained onlywith power off and with the center of gravity forward.INTRODUCTIONSevere lateral instability at the stalJ.presents a serious hazard tothe private flyer. Although much progress has been made in improving thesafety of personal aircraft, the accident reports ind

15、icate that far toomany fatal accidents still result from stalls, spins, and lack of controlnear the stall. Records show that, previous to 1929 (ref. 1), over two-thirds of the accidents were from causes associated with spins, stalls,or landings. More recent records, the Civil Aeronautics Board accid

16、entreports for 1948, show that of 850 fatal accidents i.nnon-air-carrierflying, 45 percent involved stalls.Research was begun in the early 1930s to find methods of improvingthe low-speed flying qualities of light aircraft, the first report beingpublished in 1932 (ref. 1). There have been more recent

17、 projects(refs. 2 and 3) conducted for the ssme purpose. These reports furnishedqualitative values but did not present quantitative results adequate fordesign purposes.Thus it is seen that the aircrsft designer has no convenientmethodwith which to determine the variables in design in order to insure

18、 satis-factory handling qualities at or near the stall. This is proved by thewide variation in low-speed handling qualities of the various light air-planes now in existence. Of the current personal airplanes, three typeshave been designed with the aim of maintaining lateral control near the.Provided

19、 by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NACA TN 2948 3.,stall and, under normal conditions, preventing stalling. These threetypes are generally referred to as “stall-resistant”aircraft. For theyear 1948 the CAB points out that “The over-all average f

20、or single enginedaircraft was 1 fatal stall out of but 186 aircrsft. This is more thanfour times the rate of the stall-resistantaircraft.” These facts provethat the designer can do much to decrease the rate of fatal stall acci-dents. Much of the danger associated with flying could be removed if thed

21、esigner had quantitative information with which to design low-speedhsndling qualities into the personal airplane.This investigation is based upon the hypothesis that satisfactoryrolling control is obtainableby a human pilot only if the lateralstability factor, dsmping in roll, is positive. This in t

22、urn is depend-ent on the slope of the lift curve, where an increase in angle of attackis attended by an increase in lift. It then follows that, in order toretain sufficient rolling control under all conditions, the outboardelements of a wing must be prevented from stalling.Flight tests have shown th

23、at, when an airplane is in stalled flightand autorotative moments are present together with violently changingburbled flow, a pilot cannot maintain satisfactory lateral control evenwith special devices such as spoilers which will give sarplerollingmoments for control. The difficulty is that the auto

24、rotative momentsbuild up so rapidly that the pilot cannot react quickly enough to maintainthe airplane at the lateral attitude desired (ref. 4).The aim of this project is ultimately to furnish the designer withquantitative design information from which the proper combination ofvariables may be selec

25、ted to insuresatisfactory control near the stall.This involves determining the highest angle of attack at which satis-factory lateral control can be maintained end comparing this angle ofattack with that for the maximum lift coefficient. From the comparisonan estimate can be made of any possible sac

26、rifice of low-speed perform-ance which might be entailed by ltiiting the up-elevator travel to thepoint where the critical angle of attack is the maximum that csn bemaintained.It has been fairly common design practice to twist the wing or toequip it with slots along the leading edge to control the s

27、panwise loca-tion of the initial stall point. In the case of a rectangular wing withslots, the optimum effect in countering autorotation is attainable whenthe slot covers approximately the outboard 50 percent of the semispan(ref. 5). Most designers have employed slots of considerably shorterlength.

28、Slots of less than 35-percent length, however, while preservingaileron effectiveness behind the slots, do not eliminate the autorotativemoments at angles of attack above that for the stall of the unslottedportion of the wing. In the present paper, both twist and slots areProvided by IHSNot for Resal

29、eNo reproduction or networking permitted without license from IHS-,-,-k NACA TN 2948considered as means of obtaining satisfactorylateral control at higherangles of attack near the stall. Whereas the data of reference arederived from wind-tunnel tests of a model wing, the informationpresentedherein i

30、s obtained from full-scale flight tests wti”bhtake into account,among other things, the effect of body interferencewhen the wing is inthe high-wing position.This work was conducted at the Personal Aircrsft Research Center,Texas A. descriptive characteristics are given intable I. Special fittings wer

31、e made to replace the upper attachments,of the lift struts so that the smount of washout could be varied to0, 4, and 8 at the tips. Later the airplane was flown with slotscovering various portions of the span, always with 0 of washout.External riblets were fabricated from sheet aluminum and riveted

32、to thenose ribs to adapt the slats to the existing airfoil, the contours sndinstallation of which are shown in figure 2. It was not found practicalto extend the slats around the curvature of the tips, but the effects.were unimportant since these regions remained essentially unstalled atthe critical

33、angles of attack.Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-6 NACA TN 2948Progression of the stall was observedby telltale tufts affixed tothe upper surface of the wing. Figure 3 shows the dome from which thetuft behavior was examined.The cabin

34、fuel tank, located forward of the instrumentpanel, wasdisconnected from the fuel system and only the wing cells were used,thereby minimizing the change of center-of-gravitylocation during flight.This procedure also sfforded the facility of varying the center-of-gravity location merelyby filling the

35、cabin tank with water.The human element in maintaining and reproducing steady conditionsduring the tests was eliminatedby means of a device (fig. 4) which,when adjusted in position, permitted the elevator to be held fixed whileallowing unrestricted aileron action. This device was mounted over therig

36、ht-hand-side control-wheel shaft and hinged, as shown, to a reinforcedarea of the instrumentpael. The front face of the slide was indentedto match the protrusion on the plug secured to the end of the shaft.Slide posftion was calibrated against elevator deflection after themanufacturers limit stop wa

37、s removed.Instrumentation.-All the instrumentswere nonrecording,the databeing read and recorded by the pilot in most instsnces or by an observer.Both the altimeter and the airspeed indicatorwere sensitive instruments,the latter registering in l-mph increments in the range from 10 to 80 mph,and were

38、piped to an AN 5816-1 pitot-static head mounted 1 chord lengthahead of the leading edge and 58 percent of the semispan out on the leftside of the airplane. This installation is shown in figure 5. Thecalibration of the airspeed-indicatorgage is shown in figure 6. Theposition element of the error for

39、the calibration curve was derivedbycross-plottingthe angle-of-attackcurves with the curve of the equationfor upwash mgles presented in reference 6.Figure 5 also shows the vsne-type angle-of-attackindicator and itsmounting smymgement. Its location in the field of flow about the air-foil is one in whi

40、ch the error is comparatively small and varies almostlinearly through the range of angles tested (ref. 6). 1% was chosenalso for its convenience of observationto either the pilot or theobserver. The instrumentwas calibrated in fltght, the curve (fig. 7)being derived from the sagle of climb and the s

41、inglebetween the rootchord w the horizontal.The yawmeter, which consisted of a vane mounted high enough abovethe fuselage to reduce the propeller slipstrearaeffects to atinimum andwhich was connectedby a long sheft to an indicatormounted in the roofof the cabinj was read by the pilot through a mirro

42、r system. Roll angleswere measured either by a standard gyro horizon or by a grid read byvisual.reference to the actual horizon. Pitch angles were measuredProvided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-NACA TN 2948. through the medium ofwindow of th

43、e cabin.The flight tests7a graduated quadrant painted on the right-hand-sideTEST PROCEDURE AND METHODSconductedmay be collected in the following groups.With each wing coiguration the entire group of tests was repeated withthrottle full open and throttle closed and for three locations of theairplsae

44、center of gravity.Variation of Lift Coefficient With Angle of AttackFlight tests to determine the relation of lift coefficient to angleof attack were conducted in order to correlate flight test data withthe calculated wing snalysis and to determine the sngle of attack formaximum lift coefficient.The

45、 power-off lift curve was obtained by gliding at various anglesof attack and observing the airspeed at each angle.For the power-on lift curve a series of timed 200-foot climbs atvarious airspeeds was made, in order to plot a curve of rate of cltibagainst airspeed. From the rate-of-climb curve the li

46、ft component ofthe engine thrust could be evaluated, and by combining this with power-onairspeed and sze-of-attack data the power-on lift curve was plotted.Since it was necessary to average the results from several climbs ateach airspeed in order to obtain a smooth rate-of-climb curve, it wasconside

47、red unnecessary to make refined calculations of engine output andthe slight changes in rate of climb as the fuel load decreased.6Visual Observation of Progression of Stall by Use ofTufts on Upper Surface of WingThe aircrsft was flown at a steadily increasing angle of attackuntil some portion of the

48、wing was observed to be stalled. Then a con-stant airspeed was held until all the tufts in the stalled portion couldbe plotted. A number of runs were made at constant airspeed, decreasingthe airspeed 1 mph for each run, until the stall was reached. Then theaircraft was stalled a number of times, wit

49、h aobserved during each stall, until the behavior.9 plotted.few of the tts beingof all the tufts was.Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-8 NACA TN 2$)48Determination of Maximum Angle of Attack Below Stall .at Which Lateral Control is Still Available .When an aircraft encounters gusts, substantial