FLIGHT TESTING FOR CONTROL SURFACE FLUTTER by Phi/lip Augustus Meyer

P guide in this particular phase of flight testing it is mandatory that all RIOR TO USING this article as a

other phases of design, preparation, and low-speed flight test have been satisfactorily accomplished and deficiencies corrected in the particular aircraft to be flight-tested, as described in the "EAA's Aircraft Builders Handbook" and other relevant documents. This article is primarily an elaboration for Chapter 12, Item 8.0, b, "Flutter", of the "EAA's Aircraft Builders Handbook", and is not to be used instead of diligent application of A & EE report No. 45 or similar

flutter-proof design and construction techniques. The dynamics of Control Surfaces are complicated, seldom fully understood, but vital to the flight safety of aircraft. The modern commercial and military aircraft being built today are designed in detail using sophisticated mathematical models and extensive analysis before the first part is fabricated. As the parts are assembled in various stages they are subjected to dynamic laboratory testing to confirm the mathematical models, and to determine precisely the natural frequencies of vibration and their modes. During flight-testing of the aircraft, final checks are made to prove the validity of the design. Notice that I say prove, and not determine. The state of the art of aircraft design today is such as to permit the calculation of safe airframes which are totally airworthy in the design envelope. It is wholly impractical to design primarily for esthetic reasons ignoring the immutable complex and interdependent aerodynamic principles and expect to have an airframe of any aerodynamic quality much less to have acceptable paramenters of flight. We must design within our capabilities, and then fabricate the best airframe we possibly can. 6

JUNE 1973

Assuming we have a reasonable objective in degree of airframe sophistication, sufficient capability, and have then produced an acceptable airframe by all standards, we must test it to prove the validity of our work. Even for the case of a proven design which is manufactured separately by a different agency (individual), minor differences in fabrication tolerances and techniques can produce considerably different characteristics in what is basically the same airplane. On this basic understanding we now take off and prove that our control surfaces are of proper proportions and characteristics to satisfy completely their function without the deleterious side effects of shaking, flutter, or disintegration. For the purposes of this unlimited article we shall consider only controlsurface flutter, we cannot hope to

delve into wing flutter and control abnormalities at this time or any of its causes or reasons. Let us simply say that flutter is a motion of the control surface which is of a harmonic frequency determined by any or all of airflow or energy input, shape and weight distribution of the total system, and friction or energy dissipation of the system. Let us simply consider: 1) Air flowing over a surface may impart energy to that surface in a rhythmic manner such as in forming waves on water; 2) The rhythmic energy input may be in phase or harmonic with the material, such as in waves, an organ pipe, or child's swing where appreciable energy conversion may be realized, sometimes resulting in violent noise or motion seemingly out of proportion to the apparent energy input; 3) The dissipaSample cuve of a typical surface (a)

Sample curve of a typical surface (b)• Damping critical / Flutter occurs above this level

r\ Safe flight limit A% below V c i critical cs

z

CO

s v

INCREASING AIRSPEED L

lb

V,net.

cl,

V,nea

Vne=Velocity never exceed=Vci-(Vci x 0.15) Vci=Velocity where damped oscillation begins (Region C) VC2=Velocity where sustained oscillation exist; VC3=Velocity where divergent oscillation begins Fig. 1

Vc3a

tion of the energy similar to the child dragging his feet on each cycle of the swing, or waves breaking in shallow water. Now let us consider surface flutter in flight. Due to the dynamic nature of air flow, all surfaces experiencing air

flow want to flutter, flags, leaves, clothes drying on the line. Laminar vs. turbulent airflow for our purposes here are the same. The surface primarily wants to flutter at its natural

frequency or a harmonic of that frequency, just as a tuning fork will "ring" only at its natural frequency or a harmonic of that frequency. Note also that waves on water vary in frequency and wave length from ripples to storm swells. The frequency and wave length of the waves depend primarily on the velocity of the wind, its length of sweep and the density and depth of the water. When we have present conditions 1 and 2 (air flowing over a surface having a specific natural frequency) and their dynamic natural frequencies or harmonics in cycles per second are nearly the same, we can have flutter or flapping. The determining factor as to whether in fact we do experience flutter is the strength of 3 in damping or absorbing the energy input of 1. When the energy absorption capability of the structure is greater than the evergy input of the airstream at that speed and altitude, we will not experience flutter. When the energy input of the airstream is greater than the energy absorption of the structure and the frequency of the airstream energy input is harmonic with the natural frequency of the structure, we can have flutter which will rapidly increase in amplitude until the structure catastrophically fails or the structure deforms to the extent that its natural frequency is changed and therefore the airstream and structural frequencies are mismatched at that moment. We can also detune the structure by changing the mass distribution or shape of it which in turn changes the speed at which flutter will occur. Proper juggling of the various parameters will considerably elevate or eliminate the flutter airspeed. Also note that the dynamic characteristics of the fixed surface to which the control surface is attached (wing, stabilzer, or fuselage) have a considerable influence on the characteristics of the control surface itself. Assuming we have done everything right in design and construction we now want to flight-test for flutter. I cannot overemphasize the seriousness of this phase of flight-testing. The records are filled with deaths due to airframe failures caused by flutter or its coupled effects. The onset of flut-

ter can be sudden and unforewarned even in regimes of flight previously flown many times with a particular airframe. Just because you have flown at a particular speed, altitude, and configuration many times before with no problem does not guarantee continued flutter-free operation. A single factor, or combination of

factors, such as turbulence or buffet

excitation, recent lubrication of or

freeing-up due to breaking in of the

control system, may considerably enhance the onset of flutter. Water or ice

accumulations in the trailing edges are especially effective in all aircraft in accommodating flutter. The characteristics of flutter must be pursued with great diligence and extreme caution throughout and well outside of the normal flight envelope. A minimum of ten and desired 15-percent test beyond the imposed flight limits should reliably provide flutterfree operation for the airframe so long as no influencing modifications are made afterwards. Should any influREGION A & B

encing modifications be made, it will be necessary to re-fly the entire flutter flight-test program to requalify the aircraft. It is recommended that a professional test pilot, well-trained in testing be engaged for the flight testing; however, if you wish to assume the risk, here is a suggested way of testing. We must be extremely careful in our procedure, and even so we may encounter divergent flutter, or coupled divergence which can catastrophically damage the airframe before recovery can be made. Simply shaking or displacing the stick at various airspeeds can be either dangerous or ineffective. We must perform a systematic and thorough investigation of the flight envelope. Normally, flutter will become evident with an increase in airspeed but this is not to be considered a rule, as flutter for a particular combination may occur at an intermediate airspeed and be absent at higher airspeeds. With proper conditions it is (Continued on next page)

end of excitation

f

decay of displacement (return to neutral)

-NEUTRAL

TIME REGION C

-end of excitation - decay envelope

CATASTROPHIC FAILURE-^

/

oscillation of surface

NEUTRAL

REGION (LINE) D,

continuing oscillation of surface

NEUTRAL

divergence

REGION E

NEUTRAL

increasing amplitude

Fig. 2 SPORT AVIATION

7

FLIGHT TESTING. . .

(Continued from preceding page)

entirely possible for an aircraft with marginal flutter characteristics to be

repeatedly flown through the flutter regime without incident and to encounter serious trouble at some later time. Matching frequency turbulence or slight deterioration of damping characteristics may suddenly cause divergent flutter at speeds where it was previously unnoticed. Testing of this nature is best performed at sufficient altitude in smooth air well away from other air traffic. Starting at a speed just above stalling with the airplane in trim and at stabilized power, smartly excite the controls individually in each direction and

simultaneously observe the respective surface. A disciplined Karate-type chop with the hand to the stick and a sharp kick with the foot to the rudder will normally suffice. It is important that the excitation be abrupt, result in positive displacement of the control and surface, and instantly the control be left free to return to neutral handsoff. This should be performed on each of the controls in both directions at five knots (mph.) or less increments up to the maximum design speed. At conditions of maximum, cruise, idle power and zero, one, and maximum "G" within the limits of the flight envelope at that point. Check also all other configurations within their limits; gear, flaps down, etc. The six or more conditions should each be tested prior to increasing to the next five-knot (mph.) increment. The controls must be left free as far as possible within trim capability of the airplane as the hand or foot will add to the damping of the control system and lead to satisfactory decay of some flutter whereas a free control may continue to flutter. In a suitably designed airplane the controls will immediately return to neutral without any oscillation (see Fig. 2, Region A). As a critical flutter speed is approached the surface and/or control will continue to oscillate several times about its neutral position as it damps to the neutral position (see Fig. 2, Region C). When critical flutter speed is reached, the control will continue to oscillate or flutter (see Fig. 2, Region D). When a flutter is observed, it should be stopped as rapidly as possible as abnormal stresses are being imposed on the structure and serious damage to the structure can occur. The control surface will continue to oscillate or flutter until stopped by an appropriate effort by the pilot, or the control surface destroys itself if the conditions are not changed. The presence of flutter may assume various sensory 8

JUNE 1970

characteristics. It is commonly thought to shake the control about

like a dog shakes water from its back.

However, at the higher frequencies

which may exist the control lever may

not shake and remain relatively quiet; instead the surface may be vibrating so as to make a buzzing sound in the airframe or to appear blurred to one's vision. Since the entry into the flutter region was essentially accomplished by increasing airspeed, a decrease in airspeed will again normally place the aircraft in a non-fluttering or safe

region. However, the time required to decrease airspeed to a safe region may be excessive from the standpoint of damage and structural fatigue. There are several ways to decay rapidly the flutter of a control surface: 1) Reduce air speed; 2) Manual freezing of the affected control; 3) Appreciable displacement of the affected control; 4) Change of flight conditions by judicious use of the other controls. It is best to effect simultaneously the three (l,2,and 3) primary means of stopping flutter, as the most rapid means of recovering from the dangerous onset of any flutter. These should essentially be performed simultaneously within the limits of the airframe; however, reduction of airspeed is paramount. The manual freezing will add considerably to the mechanical damping within the limits of strength of the control and pilot. This will generally reduce the flutter to a minimum value until reduction of airspeed has time to take effect and alleviate the flutter. Fig. 1 is two exemplary curves of typical surfaces (a) and (b). Surface (a) has the characteristics of an intermediate flutter speed beginning at Vela reaching critical at Vc2a and then experiencing decreasing excitation so that at a slightly higher speed the surface is flutter-free, only again to experience flutter as the speed continues to increase. Surface (b) is very critical and progresses from flutter-free to divergent flutter with very little increase in airspeed. Fig. 2 is a progression of curves showing the motion of the surfaces at

the speeds and energy levels depicted

in Fig. 1, Regions A, B, C, D, and E

(along the y-ordinate axis). Referring to Fig. 1, Region D defines the lower limit of Region E. A slight increase in airspeed will cause entry into Region E. A reduction in airspeed will progressively enter Region C and proceed to safe Region B. While I refer to all five Regions in the above description, they are for explanation only. AT NO TIME ATTEMPT TO EXPERIENCE FLUTTER BEYOND THE VERY BEGINNING OF REGION C. The increase in airspeed necessary to progress from Region C

through D to E may be less than one knot. Also, repeated excitations in Region C may cause damage and entry into Region E with no increase in airspeed. In the testing procedure, if the increases in airspeed are of too-large increments or damage occurs such as inadvertent loss of balance weight or hinge-linkage, one may suddenly find oneself in a divergent condition whereupon excitation the control will very rapidly increase the amplitude of its oscillation going beyond the control stops of the airplane and destroying the control and surface by deformation or loss (see Fig. 2, Region E). Should this happen due to circumstances beyond the control of the pilot, several things may simultaneously be performed to dampen the oscillations: 1. Reduce airspeed as rapidly as possible. The decrease in airspeed will stop the flutter by slowing the airplane to a stable airspeed. 2. Prevent the controls from fluttering by manually freezing the position of the control. This means to oppose physically the oscillations by manually trying to lock the control in one position. 3. Force the fluttering control to an extreme position, however, considering the "G" loads being imposed. 4. Change the loads and positions of the other controls. For example, in case of a single fluttering aileron, yaw, skidding away from the damaged surface to lengthen the effective chord and possibly blanket the airflow. 5. Change configuration to a previously proven safe one if possible; flaps, trim tabs, gear, canopy, etc. (use configuration change with discretion, do not cause a more serious problem but again sometimes the oddest things will help). As previously mentioned, and shown in Fig. 1, it is necessary to test for flutter to approximately 15 percent beyond the normal operation red-line speed to assure reasonably flutter-free operation under all laterencountered conditions. The speed is properly determined in the reverse manner, i.e., determine the maximum speed the airframe is capable of and test for characteristics to that speed. Then red-line your aircraft at 87 to 91-percent of the maximum speed tested to. This red-line is then the maximum never-exceed speed for your airplane. (Note: The red-line for production aircraft is not necessarily determined in this manner and having the tolerance of the flutter red-line. There are other red-line factors which may be more restrictive and having less tolerance, i.e.,fasteners, fabric, skin thickness, rib-stitching, windshield, etc.).

Ideally, your aircraft will show no evidence of flutter throughout the test envelope (test envelope equals normal envelope plus 15 percent) and you should have flutter-free flying for the life of the airframe. But do not feel the test was in vain, for what you did

was to prove the qualities of your airframe. There is nothing more satisfying than a medical check-up with a resulting clean bill of health. But if cancer of flutter is found, cure it while you are able before it kills.

If a tendency to oscillate a cycle or

two (Region C) is noted at any speed,

or even at some intemtediate airspeed and it disappears at a higher airspeed, the controls and airframe must be

reworked until this tendency is alleviated. All tendencies to flutter (Region

C) must be removed from the flighttest envelopes, since if the aircraft at the time of flight-test shows only a tendency to flutter any slight change in damping, for example due to moisture or corrosion in the structure, may

later cause divergent flutter at that

same speed.

There are several other factors we should also consider in flight-test of a new aircraft to determine truly its aerodynamic airworthiness. The impor-

tance of these factors is relative to the particular design of the aircraft, and the type of flying and flight conditions which it will encounter. Factors as diverse as wing or surface divergence, control reversal and inertial coupling may be critical in a particular design. These are usually determined prior to the first flight by means of computeraided analysis. We must design and build as best we can, and then safely prove our aircraft to be safe. (Author's note: I ask all aerodynamicists and those knowledgeable in sophisticated aircraft design to consider the necessary over-simplification and apparent misuse of technical description which are also vernacular in relating this article to the readers' understanding. However, all comments will be deeply appreciated). About The Author Mr. Meyer resides on a farm near Lake Zurich, Illinois where he has a landing strip and keeps his own Stinson 108-3. His aviation background: 1. M.S.,B.S. Aeronautical Engineering, University of Illinois. Wrote thesis on multi-stage, three-phase plasma generation (Deep Space Drive), Designed Schileren Optical System for

you have it on is being offered by Security Parachute Company of San Leandro, California. According to the firm, the Model 150 Safety-Chute is the first parachute to be assigned specifically for use in sailplanes, low-power aircraft (under 150 mph), and helicopters. The extra-thin parachute has neither seat nor back, weighs approximately 16 Ibs., can be used in any style of seat with complete comfort, and goes on and off with two quick-flip buckles. Rate of descent of the all-Nylon steerable canopy is said to be comparable with that of a 28 foot canopy. The manufacturer also notes that the parachute is the only one on the market with a four-month repack period and is certified under TSO C-23B. For further information, write to: Security Parachute Co. P. O. Box 3096 San Leandro, California 94578

UI Hypersonic Blow Down Tunnel. 2. Conceived the jet augmentation of the Air National Guard's KC-97F&G, and worked in close liaison with the USAF, the National Guard Bureau, and other agencies to promote and develop the KC-97L. Devised and flew the initial flight-test program of the prototype KC-97L (JKC-97G). 3. Various aerospace consulting work. 4. Captain for major airline flying four-engine jets. 5. Member IFALPA Airworthiness and Performance Committee, member ALPA Boeing 747 Committee, Advisor ALPA SST, and Airworthiness and Performance Committees. 6. Tactical and all-weather interceptor jet pilot. 7. Instructor Glider Pilot, with Silver C for soaring. 8. Has done preliminary design work for a supersonic home-built capable of M 1.7 at 60,000 feet. 9. Presently developing a simplified angle-of-attack system, which will be flight tested on his Stinson to replace the airspeed indicator as a primary-flight instrument.

My Shadow Is a Silly Pup

New Lightweight A new kind of parachute said to be so light and comfortable you'll forget

the

By B. J. Rompre', EAA 32326

Parachute

10635 50th Street Edmonton, Alberta, Canada

My shadow is a silly pup Who slinks beneath the plane to see If he should come along and ride Or stay behind and wait for me! When I taxi to the strip He trots along as if to say "Today I'll tag along with you, No one can make me stay!" I flip the throttle to take off, He runs along with glee! He gives a most-impressive show. No laggard pup is he! Oh no! He doesn't stay behind Rooted to the spot, But swiftly skims along Each play-sized country lot. Then oozes over fields and ponds, And vaults over house tops. Ahead, behind, and to the sides He dodges, but never stops Till at last I near the earth. Then, he also thinks it's best To slow down a little bit And, like me, come to rest! SPORT AVIATION

9