Reducing the Risk of Its causes and potentially catastrophic effects NEAL WILLFORD, EAA 169108

58 DECEMBER 2003

PHOTO ILLUSTRATION BY JIM KOEPNICK

FLUTTER FLUTTER FLUTTER L

egend has it that the famed Wisconsin air racer Steve Wittman used to check his racing airplanes for flutter by making high-

speed passes just above frozen Lake Winnebago. His reasoning was that if he experienced flutter, he could quickly land and slide to a stop on the lake. You won’t find any flight-test manual recommending this procedure, but in the 1930s, as racers pushed the performance envelope, Sport Aviation 59

GLASTAR SPORTSMAN 2+2—PHOTO BY MIKE STEINEKE

Whether you’re building your own design or someone else’s, you need to do what you can to ensure that your airplane will never experience this destructive phenomenon.

designers knew that flutter was a dangerous thing, even though they may not have completely understood it. Flight-testing was the main way they had to show that their designs were free from flutter. Fortunately, a lot of research has been done since then, enabling designers and builders to reduce the risk of potentially catastrophic flutter. Let’s take a look at some of the causes of flutter and how to reduce the risk of it. A new spreadsheet to help estimate the amount of control surface balance needed is available for download from the EAA website at www.eaa.org. Just click on the EAA Sport Aviation cover and scroll down to the December links.

What Causes Flutter The FAA defines flutter as: “The unstable, self-excited oscillation of an airfoil and its associated structure, caused by a combination of aerodynamic, inertia, and elastic effects in such a manner as to extract energy from the airstream. 60 DECEMBER 2003

The amplitude of oscillation (at the critical flutter speed) following an initial disturbance will be maintained. At a higher speed these amplitudes will increase.” You’ll probably not want to use this definition when explaining to your spouse why a flag flutters in a stiff breeze! However, this long definition describes the causes of flutter, so let’s dig them out to see what roles they play and how to minimize the risk. The three factors that can combine together and cause flutter are: ■ Elastic properties of the structure ■ Inertia effects ■ Aerodynamic effects Elastic Properties: Most objects are flexible and will oscillate, or vibrate, at their natural frequency if acted on by a force. This natural frequency is a function of the object’s geometry, weight, and material properties. You can demonstrate this with the following experiment. Put a yardstick on a table and let 6

inches hang over the edge. While holding it down, pull up and release the overhanging end. You will see and hear the yardstick oscillate. Repeat the experiment with a 12inch overhang and you’ll see and hear that it oscillates at a lower frequency. Each type of motion that an object experiences is called a mode, and it usually takes two or more modes acting together to have flutter. The yardstick experiment is an example of a bending mode. We could have also twisted and released the yardstick and it would have oscillated at some torsional frequency—an example of a torsion mode. An airplane’s fuselage, wing, and tail surfaces can all oscillate in these two modes. An airplane’s critical flutter speed is a function of the natural frequency of its various parts, and all things being equal, a higher natural frequency leads to a higher critical flutter speed. It is the designer’s goal to ensure that critical flutter speed for each of the possible flutter modes is higher than the

maximum dive speed of the airplane. Conventional fuselages with tail surfaces attached directly to the tailcone are usually pretty stiff in bending and torsion. This is good, because higher stiffness leads to a higher natural frequency. A pod and boom fuselage, on the other hand, tends to be more flexible because the boom has a smaller smaller cross section. Designers who use a pod and boom arrangement should keep this in mind and try to use as stiff a boom as possible. Moveable control surfaces and trim tabs increase the number of modes, and the list of possible mode combinations that could cause flutter gets long quickly. Fortunately, the interaction of these different modes is pretty well understood, and following good design and building practices can minimize the risk of flutter. An airplane’s structural material also helps determine the airframe’s elastic properties. Researchers have pulled, bent, and twisted the different materials we use to build airplanes in order to determine their limits. Two of these material properties are the modulus of elasticity and the shear modulus. Each is a measure of a material’s “springiness,” with the first indicating how much the material will bend and the second how much it will twist when a load is applied. A higher modulus means higher stiffness, which is desirable for reducing the risk of flutter. Some materials have modulus values much higher than others. For example, fiberglass is a strong material, but its modulus values are less than 50 percent than the values of aluminum. This means that a fiberglass wing designed to the same strength levels as an aluminum wing would be roughly twice as flexible. One reason why composite aircraft have the plies of glass on the skins and spar webs oriented at roughly 45 degrees is to help Sport Aviation

61

SAFETY ON DISPLAY. The New KMD 250 Multi-Function Display. Sharp 3.8-inch active matrix LCD. Positioning, data link weather, traffic and terrain, all on an affordable high-resolution and user-friendly display. Optional VFR GPS. Give us a call at 1-877-712-2386 or 1-913-712-2613 or visit our website at www.bendixking.com/KMD250 for more information.

For more information, visit SPORT AVIATION on the Web at www.eaa.org

improve their torsional stiffness. Carbon fiber has much higher modulus values than fiberglass, which is one reason why it’s often used on the control surfaces and wing skins of high-performance composite aircraft. The torsional stiffness of control

surfaces is important for reducing the risk of flutter, so it is desirable to skin them with something that is light and stiff. From a cost and weight standpoint, aluminum and plywood are hard to beat. The next time you see the largely composite Cirrus SR-22, take a look at its con-

Figure 1

Wing Accelerating Up

Steady Flight

Wing Accelerating Down

Figure 2

Hinge Line

CG of Part

Axis Line

62

DECEMBER 2003

trol surfaces—they’re aluminum!

Inertia Effects: Inertia is the resistance of an object to change direction and the heavier the object, the higher the inertia. Inertia effects depend on the weight and location of the airplane’s individual pieces. These effects are actually closely tied to the airplane’s elastic characteristics. You can see this by repeating the yardstick experiment and attaching a pair of vise grips to the overhanging end. With the added mass the yardstick will now vibrate at a much lower frequency than before. Without the vise grips, when the yardstick reaches its maximum deflection, it immediately wants to spring back in the opposite direction. With the vise grips, the yardstick is not as anxious to change directions and consequently results in a lower oscillating frequency. This is not good; we want the natural frequency to be higher, not lower. An aviation example would be installing outboard fuel tanks (either internally or as tip tanks). Like the vise grips, the tanks would alter the wing’s natural frequency and could reduce the critical flutter speed. A T-tail is another “vise grip” example. Designers using this configuX ration must ensure that both the vertical fin and tailcone (or boom) are as stiff as possible to reduce the possibility of the fuselage fluttering in Y torsion and the vertical fin in bending. T-tail sailplanes often have the horizontal tail mounted to the forward part + X Direction of the vertical fin to put it ahead of the fin’s elastic axis and thereby reduce the tendency of flutter. So far our discussion about inertia effects has been

limited to the airplane’s fixed structure, but we can’t forget the rudder, elevator, and ailerons. Their inertia effects play a critical role in whether the airplane will have flutter problems. To see why, look at Figure 1. The middle airfoil represents the outboard portion of a wing in steady flight. The circle on the aileron represents its center of gravity, which on unbalanced control surfaces is usually aft of the hinge line. Let’s assume the airplane encounters a strong upward thermal that suddenly deflects the wing upward. Recalling that inertia is the resistance of an object to change direction, and that inertia forces act at the center of gravity of an object, the aileron would tend to lag behind the upward moving wing, as shown in the top of Figure 1. To the oncoming air, the lagging aileron looks like a deflected flap, which increases its lift and a tendency to twist the wing’s outboard section. At some point the wing’s stiffness will stop this deflection and the wingtip will rebound in a downward motion. The aileron again lags behind the wing, as shown in the bottom of the figure, and the upward deflected aileron will decrease the lift and increase the tendency to twist the leading edge up. This downward deflection will continue until the wing stiffness again prevails and sends the wing springing back upward. This combined bending and torsion motion will eventually dampen out if the airplane is flying below the critical flutter speed for this type of interaction. If the airplane is flying above this speed, the oscillation will feed on the energy from the airstream and the magnitude of the deflections will rapidly increase until the pilot slows the airplane—or worse, something breaks. Fortunately there are ways to reduce the risk of this kind of flutter that we will discuss shortly. Aerodynamic Effects: Aerodynamic effects are the final ingreSport Aviation

63

SOLVE YOUR TRAFFIC PROBLEMS. The new solid-state KT 73 Mode S Data Link Transponder enhances pilot “see and avoid” safety and efficiency in flight using FAA TIS technology — the new, free Traffic Information Service. Call 1-877-712-2386 or 1-913-712-2613 or visit our website at www.bendixking.com/kt73 for more information.

Quieter • Safer • Painless “To fly over 19,000 miles in 81 hours without squirming in the seat is priceless! I can honestly say I have never experienced such seat comfort.” “The difference in comfort and sound levels since Oregon Aero worked their magic on my helmet and headset is nothing short of fantastic!” photo courtesy B. Bohannon

Bruce Bohannon

Bruce Bohannon, Pilot, The Exxon Flyin’ Tiger, Has Set or Broken 19 Altitude and Time to Climb World Records.

• • • •

Custom & Portable Seat Cushion Systems Headset & Helmet Upgrades Aircraft Interior Upgrades Shoe Inserts

New Ways Of Thinking To Solve Old Problems. 34020 Skyway Drive, Scappoose, OR 97056-2516 800-888-6910 • 503-543-7399 • www.oregonaero.com • [email protected] Money back guarantee. Visa, MasterCard and US Government IMPAC cards accepted.

For more information, visit SPORT AVIATION on the Web at www.eaa.org

VAN’S RV-10A—PHOTO BY LEEANN ABRAMS

When it comes to preventing control surface flutter, there are two types of control surface balance to consider: static and dynamic.

dients of the flutter recipe. The air rushing past an airplane in flight provides a large source of energy, which an oscillating wing or control surface can tap into. This energy source is proportional to the airspeed squared, so a slow airplane will have a reduced tendency to flutter. (Notice I didn’t say zero tendency.) Wind gusts and buffeting due to separated flow can also provide a pulsing force to initiate flutter. As long as these disturbances happen below the critical flutter speed, aerodynamic forces and structural stiffness should dampen any resulting oscillations out. At higher altitudes the air is less dense and consequently the aerodynamic dampening is also reduced. Also, an airplane flying at high altitude and high airspeed can start experiencing compressibility over portions of the wing or tail due to the local flow reaching a high Mach number. This can result in aerodynamic buffeting that can cause control surfaces downstream to start vibrating. 64

DECEMBER 2003

Reducing the Risk of Flutter Whether you’re building your own design or someone else’s, you need to do what you can to ensure that your airplane will never experience this destructive phenomenon. Years of research and cumulative flight experience have provided us with good recommendations for reducing the risk of flutter. Aerodynamic: The aerodynamic shape of the control surface can influence an airplane’s tendency to flutter. Control surfaces with flat or concave sides are more flutter resistant than those with convex (slightly bulging) contours. Another trick some designers use to reduce the risk of control surface flutter is to bevel the trailing edge slightly. Control System Rigidity: Free play in the control system should be reduced to a minimum. Hinges and control connections should not have excess slop, and control cables should be at their proper tension. From a potential flutter standpoint, a control system that uses large

diameter push/pull tubes has an advantage over a cable system. Reference 1 recommends that with one aileron clamped to the wing, the total free play at the trailing edge of the other aileron should not exceed a distance equal to 2.5 percent times the average aileron chord (measured aft of the hinge line). Excessive slop in the aileron control system can aggravate the condition shown in Figure 1. Similarly, the total free play in trim tabs should also not exceed 2.5 percent times the average trim tab chord. The trim tabs should also be irreversible, meaning that they should only move by using the trim wheel or switch, not by applying a load to the tab itself. Finally, the interconnection between elevator surfaces should be as torsionally stiff as possible. Using a large diameter elevator interconnect tube is a way to help achieve this. Control Surface Balance: The most effective way to reduce the risk of control surface flutter is to use an appropriate amount of control sur-

face balance. Flaps are normally not balanced because of their inboard location, but the ailerons, elevator, and rudder often have some amount of leading edge balance weights. It’s true that some airplanes don’t have any leading edge balance weights on their control surfaces, and they don’t have flutter problems. Often they are slower airplanes with fabric-covered control surfaces whose center of gravity tends to be close to the hinge line. But, just because some “slow” airplanes haven’t had flutter problems doesn’t mean that others won’t, so designers of slow airplanes can’t just assume their designs will be flutter free. The FAA’s predecessor, the CAA, provided information (References 13) that is still useful to help designers estimate the amount of control surface balance needed to reduce the risk of flutter. The FAA still allows Reference 1 to be used for certified airplanes, provided that it meets a whole slew of criteria (see FAR 23.629 for more details). When it comes to preventing control surface flutter, there are two types of control surface balance to consider: static and dynamic. Static balance deals with whether a control surface balances about its hinge line. A surface is completely balanced if it’s balanced about its hinge line. It’s underbalanced if its trailing edge is heavy and overbalanced if its leading edge is heavy. This is the type of balancing that most builders or restorers are familiar with, and the plans or manual for their airplane should provide information on how the control surfaces are to be balanced. Don’t forget that repainting your airplane can change the balance on its control surfaces due to the added weight of the paint, so be sure that they are properly balanced before flying. Because a control surface’s structure is mostly aft of its hinge line, airplanes that require static balanced surfaces often locate the necSport Aviation 65

essary balancing weight in the surface’s leading edge or on an extended arm. Designers, however, like adding weight to their airplanes as much as pilots like paying user’s fees, so they often try to make the control surfaces as light as possible. This is one reason why so many World War II fighters, like the F4U Corsair, had fabric covered control surfaces. A control surface is dynamically balanced about an axis (like the centerline through an airplane) if an angular acceleration about this axis does not tend to cause the control surface to rotate about its hinge line. For example, if a wing is oscillating up and down in bending, the outer portion of the wing is experiencing a larger angular acceleration than the inner portion. If the aileron is not dynamically balanced, this wing flapping can cause the aileron to start rotating about its hinge line. Now there’s two modes

of vibration, and you may recall that it usually takes two modes interacting to start flutter. Another way to reduce the amount of balance weight is to design the control surface to be statically underbalanced and ensure that it has a proper amount of dynamic balance. Reference 3 states that this is okay as long as the maximum dive speed is less than 300 mph. Above this speed, it recommends that all control surfaces be both statically and dynamically balanced. Dynamic balance depends on the weight of each control surface part, as well as their center of gravity’s location with respect to the hinge line and axis of rotation. The amount of dynamic balance can be determined experimentally (as shown in Reference 1) or calculated if you know the weight and CG locations of each of the parts. Figure 2 shows the dimensions needed to

For more information, visit SPORT AVIATION on the Web at www.eaa.org

estimate the dynamic balance of a control surface, which can then be calculated using the following equation: Dynamic Balance Coefficient =

Σ Part Weight x X x Y Σ Part Weight x X x X

The top half of the equation is called the product of inertia and the bottom half the mass moment of inertia. The Σ symbol represents the total sum for each part in the control surface. X values are positive for distances aft of the hinge line, and Y values are positive when they are above (or outboard) the axis line. A control surface is dynamically balanced when the sum in the top of the equation is equal to zero. A negative coefficient means the surface is overbalanced and a positive coefficient when it is underbalanced. According to the equation, when a balance weight is located out at the tip, the value of X times Y is very negative, and consequently is a very effective way to make the top part of the equation equal to zero. When the balance weights are outboard, the result can be a control surface that is dynamically balanced yet statically underbalanced. However, there is a limit to the amount of static underbalance that a control surface should have, even with the correct amount of dynamic balance, and Reference 3 recommends that the center of gravity should never be aft of 15 percent of the control surface chord. Through flight experience and testing in the 1930s through the 1950s, researchers determined the acceptable amount of dynamic balance for the various control surfaces, which are presented in References 1–3. Note that these criteria are based on airplanes using construction methods popular in that era and having “conventional” tail surfaces and fuselages. Those interested in designing an all-composite airplane—one with pod and boom fuselage or a T tail— 66

DECEMBER 2003

should, for preliminary design purposes, make sure the control surfaces are both statically and dynamically balanced. Then it would still be advisable to seek out an experienced flutter expert to run a ground vibration test on your design. Modern structural analysis programs can also be useful to do preliminary flutter estimates for unconventional designs. We saw that locating the balance weight at the tip will usually result in the least amount of weight needed, but such an installation does require that the control surface be very stiff in torsion. Some designers instead prefer to locate the balance weight in the outboard portion of the leading edge. Wherever the location, it is extremely important to make sure that the balance weight attach structure is both strong and stiff. A balance weight attached to a flimsy arm can end up causing more problems than it solves. The FAA requires that the attach structure for balance weights be able to support that weight times the following LIMIT load factors: ■ 24 g normal to the plane of the control surface ■ 12 g fore and aft ■ 12 g parallel to the hinge line These should be the minimum limits for homebuilt airplanes also. Finally, if you’re building a proven design and can verify that the designer or manufacturer has demonstrated that the airplane is free from flutter at least 11 percent above redline speed (VNE), don’t change the control surface balance weight or their locations. If you do, you might inadvertently increase the risk of flutter in an otherwise safe design. Don’t second-guess the designer unless you know what you are doing!

Flutter Testing There are two types of testing used to check if an airplane will have flutter problems. The first is a ground vibration test, where special

equipment vibrates various parts of an airplane at different frequencies. The results can be used to identify the possible flutter modes and frequencies and also to correlate dynamic computer models. The second involves diving the airplane to a speed 11 percent faster than the redline speed while, according to FAR 23.629, “proper and adequate attempts” to induce flutter have been made. See References 5 and 6 for more information on both of these types of tests. Flutter is serious business, but your chances of ever experiencing it can be reduced by using good design practices, or by following the designer’s or manufacturer’s balance requirements for your airplane, and finally by flying at or below redline speed. Design, build, and fly safely!

More at www.eaa.org Click on the EAA Sport Aviation cover and scroll down to the December links.

References Airframe and Equipment Engineering Report #45, “Simplified Flutter Prevention Criteria for Personal Type Aircraft.” Civil Aeronautics Manual (CAM) 04, “Airplane Airworthiness,” 1944, http://dotlibrary.specialcollection.net/ ANC-12, Vibration and Flutter Prevention Handbook, 1948. Aircraft Vibration and Flutter, Freberg, C. and Kemler, E., John Wiley and Sons, 1944. “The New Look on the Turner T-40 and Thorp T-18”, Thorp, John and Turner, Eugene, EAA Sport Aviation, August 1969. Understanding Flutter, Thorp, John and Sunderland, Lu, EAA Sport Aviation, September 1976.