23.1.1 The Content of this Chapter

23.2.1 Introduction to Control Surface Hinge Moments

In a conventional, human-operated control system, the pilot must react the aerodynamic hinge moments that results from deflecting a given control surface. Consider the control surface in image A of Figure 23-1, which consists of an airfoil and a flap. Hinge moment analysis effectively returns the magnitude of the actuation force F, which multiplied by the bellcrank dimension d is the hinge moment. The force required to deflect the surface depends on parameters like the geometry of the airfoil and flap, deflection angle (δ), and the AOA of the combination. The force is applied using a stick, a yoke, a “steering wheel,” foot pedals, and, sometimes, a hydraulic or electrical actuation system. There are a number of ways to affect the magnitude of the aerodynamic hinge moment for any given geometry:

Hinge moments are highly affected by the geometry of the controls, including the hinge location and shape of the control surface. A general expression for the hinge moment is given below:

(23-1)

where

The hinge moment coefficient is then given by:

(23-2)

The tailoring of hinge moments is a process used to fine tune the hinge moments of a particular control surface so that some desirable stick forces will be sensed by the pilot. This is an essential task for GA aircraft, especially smaller ones, in which the pilot generates the actuation forces. It is also important for the development of boosted control actuation and autopilot integration.

Table 23-1 shows typical flap chord sizes by control surface type and their usual deflection limits. Of course, there are exceptions from these numbers, but typical values are between the numbers presented.

23.2.2 Fundamentals of Roll Control

The purpose of the ailerons is to provide control about the airplane’s roll axis, by modifying the rolling moment coefficient, C_l (do not confuse with the section lift coefficient, which shares the same variable name). The most effective way to accomplish this is to modify the airfoil in the outboard region of the wing by changing its camber by deflecting a control surface. This modifies the distribution of lift along the wing, un-symmetrically, which generates a rolling moment. The effect of deflecting ailerons on the distribution of section lift coefficients predicted by potential flow theory can be seen in Figure 23-2. The ailerons are deflected some 15° and the wing’s AOA amounts to 8° at 100 KCAS. Also, Figure 23-3 shows the resulting impact on the flow field behind the wing. It highlights how the wing that travels up generates higher lift-induced drag than the opposite wing, causing a phenomenon known as adverse yaw.

There are three common types of ailerons used in modern airplanes; plain flap ailerons, Frise ailerons, and spoiler-flap ailerons, also called spoileron (a combination of spoilers and ailerons). Schematics of these are shown in Figure 23-4. Other aileron types include the flaperon (a combination of flaps and ailerons) and elevon (a combination of elevators and ailerons). There also exist some highly specialized types of ailerons such as the slotted-lip aileron and some research project ailerons such as microjet ailerons.

23.2.3 Aileron sizing

With the preliminaries out of the way in the preceding section, it is now possible to present the following method to help with the sizing of the ailerons. The procedure involves estimating the steady-state roll rate using the aileron authority and roll damping derivatives.

23.2.4 Fundamentals of Pitch Control

The purpose of the elevator is to provide control about the airplane’s pitch axis, by modifying the pitching moment coefficient, C_m. Figure 23-19 shows the pitching moment coefficient, C_m, versus angle-of-attack (AOA) for an airplane with a conventional tail. This graph does not belong to any specific aircraft type, but rather shows typical results from a wind tunnel test for such an aircraft. The three curves represent the C_m for the complete aircraft with different elevator deflections and show how deflecting the elevator will allow the pitching moment curve to be moved around. The top curve shows the C_m with the elevator deflected 20° TEU (δ_e = −20°) as would happen for a nose pitch up (climb or stall). The center curve depicts the elevator in a neutral position (δ_e = 0°), and the bottom depicts the elevator deflected 15° TED (δ_e = +15°), as would happen if the airplane’s nose were pitched down (e.g. a dive). The separation between the curves represents the effective elevator authority for the aircraft. The curves show the classic linear downward slope required for a naturally statically stable aircraft and then a transition into nonlinear curves associated with the post-stall AOAs. When the airplane is operated at low AOAs, the airflow is mostly attached and the C_m changes linearly. However, with higher and higher AOAs larger areas of the airplane become covered with separated flow and this is the main cause for the non-linearity.

Now, consider the linear section of the curves as they intersect the horizontal axis at different AOAs. For instance, the top curve (labeled with δ_e = −20°) intersects the AOA axis at a α ≈ 15°, the center curve at a α ≈ 5.5°, and the bottom curve at a α ≈ 0°. Each of the three points indicates that the airplane’s C_m has reached zero – in other words, the aircraft is trimmed at that particular AOA. For that reason, they are called trim points. Trimmed means that as long as the slope of the curve is negative, the airplane will “want” to stay at that attitude (AOA) until disturbed again – the trim point is an equilibrium point. It is important to keep the following in mind. In order for the airplane to remain trimmed in this fashion, two things have to happen: (1) The slope of the C_m must be negative, and (2) the value of C_m must be zero. If the slope is positive, the airplane will continue to pitch and will not stop at C_m = 0. The whole purpose of putting a tail on an airplane is to force the pitching moment curve for the airplane to have a negative slope. And the whole purpose of featuring an elevator is to allow the pilot to move the trim point back and forth. This way, there exists one and only one trim point for any airspeed and the airplane must also have enough thrust to maintain flight at that airspeed, whether it is generated mechanically or by gravity (as for sailplanes). Note that for real airplanes there is only one curve, and not three as shown in the graph, and the pilot can move it back and forth at will. The three curves are only shown to indicate the trim range.

The importance of the three curves is that they also indicate the controllability or the control authority of the aircraft. For instance, if the pilot moves the elevator to the neutral position (the center curve), the aircraft would be trimmed at approximately α ≈ 5.5° as stated earlier. If he deflects it to −20° it would trim at α ≈ 15° and so on. The airplane might have a maximum elevator deflection of −25°, allowing the pilot to bring it to an even higher AOA and to stall if necessary. The curves show that the pilot can control the trim AOA by simply deflecting the elevator and that way move the trim point to a higher or a lower AOA. This, of course, is elementary information that any student of aerospace engineering is exposed to.

The required size of the elevator should be based on the investigation of the capability of the aircraft at the extreme conditions stipulated in Sections 11.2, 23.3, and the additional tools provided in Appendix C1.6. Such an investigation should plot the required AOA and elevator deflection over the entire speed range of the aircraft and include the extremes of the CG envelope. The sizing should ensure the required elevator deflection is less than what is available, per Section 23.3.

23.2.5 Fundamentals of Yaw Control

The rudder sizing depends generally on two conditions: crosswind requirements and, if the aircraft is multiengine, on one-engine-inoperable (OEI) asymmetric thrust requirements. Additional requirements would be spin recovery; however, this is particularly challenging to analyze because of the separated flow field the rudder operates in. For this reason, the rudder is generally sized for the crosswind and asymmetric thrust and there is a good chance it will be adequately sized for spin recovery, provided the fin is not blanketed by a horizontal tail or fuselage.

Figure 23-20 shows the yawing moment coefficient, C_n, versus angle-of-yaw (AOY) for an airplane with a conventional tail. Like Figure 23-19, this graph does not belong to any specific aircraft type either, but rather shows what results one would expect from a wind tunnel test. The three curves represent the C_n for a complete aircraft with the rudder deflected left, neutral, and right. As with the elevator, it shows how deflecting the rudder will shift the yawing moment curve up or down, so the airplane is trimmed with the nose pointed to the left or right. The left-most curve shows the C_n with the rudder deflected 20° trailing edge left (TEL) (δ_r = +20°) as would happen for a nose-left slip. The center curve depicts the C_n with the rudder in a neutral position (δ_r = 0°), and the bottom one depicts the rudder deflected 20° trailing edge right (TER) (δ_r = −20°), as would happen for a nose-right slip. The separation between the curves represents the effective rudder authority for the aircraft. The curves show the classic linear upward slope required for a naturally statically and directionally stable aircraft. It also shows that below a certain AOY (typically around 12–15°) the C_n is linear, but beyond becomes nonlinear curves associated with the post-stall AOYs.

As discussed in Section 11.2.8, Requirements for static directional stability, the slope of the C_n curve must be positive as shown in Figure 23-20 for the aircraft to be directionally stable. This is indicated by the magnitude of the stability derivative C_nβ, i.e. the directional stability. The required size of the rudder should be based on the investigation of the capability of the tail at the extreme yaw conditions stipulated in Sections 23.3. Such an investigation should evaluate the AOY the airplane can be trimmed to with a given rudder deflection over the entire speed range of the aircraft and take into account the extremes of the CG envelope. The sizing should ensure the required rudder deflection is less than what is available, per Section 23.3.

23.3.1 Crosswind Capability at Touch-Down

The airplane should be capable of maintaining straight and level⁵ flight while yawed 12–15° (β) at an airspeed amounting to 1.2 × V_S0, where V_S0 is the stalling speed in the landing configuration, with sufficient power to maintain the airspeed. This corresponds to a crosswind component of approximately:

(23-15)

23.3.2 Balked Landing Capability

The airplane should be capable of climbing in the landing configuration with full power at the reference airspeed and the CG at the most adverse location (forward for conventional aircraft). At this condition the dynamic pressure is low, which will require large control surface deflection. If met, this requirement prevents the HT from being undersized (also see Section 23.3.3, Take-off rotation capability). Paragraph 14 CFR 23.73 defines reference airspeed as follows:

Normal, Utility, and Aerobatic piston-powered aircraft with W₀ ≤ 6000 lb_f:

(23-16)

Normal, Utility, and Aerobatic piston- and turbine-powered aircraft with W₀ > 6000 lb_f:

(23-17)

Commuter category aircraft with W₀ > 6000 lb_f:

(23-18)

where

23.3.3 Take-Off Rotation Capability

The aircraft must offer a large enough elevator authority to allow the airplane to be rotated at airspeeds below the liftoff speed, V_LOF, with the CG in the most adverse location (forward for conventional aircraft). If met, this requirement (also see Section 23.3.2, Balked landing capability) will prevent the HT from being undersized for elevator authority.

23.3.4 Trim at Stall and Flare at Landing Capability

The aircraft must be designed with enough elevator authority to drive it to the desired stalling speed in the most adverse configuration as detailed below. The stalling speed is the lowest airspeed at which the airplane can maintain level flight. If the airplane has insufficient elevator authority it will not experience a normal “stall-break” (i.e. the sudden nose-drop that indicates the stall event). Instead it will descend rapidly at airspeed higher than the stalling speed. As far as aviation authorities are concerned, this becomes the airplane’s minimum airspeed. This can result in a serious design predicament, especially if the aircraft is designed to comply with regulations such as 14 CFR Part 23, which has a stall speed limit of 61 KCAS for single-engine aircraft, or Light-Sport Aircraft regulations, which have a stall speed limit of 45 KCAS (14 CFR 1.1).

Another easily overlooked characteristic is the ability for the pilot to flare the aircraft in the most adverse configuration during touch-down. There are examples of aircraft that have been certified with insufficient elevator authority. Such airplanes pose a serious risk to unsuspecting pilots. It is the purpose of regulatory authorities to ensure such aircraft don’t slip through the cracks. Any certified aircraft discovered to suffer from insufficient elevator authority should have a forward CG limitation imposed on it.

Strictly speaking, the stall speed is the airspeed below which the airplane’s nose pitches down uncontrollably. When we say trim at stall (see below), we mean the airspeed just above the stall speed.

23.3.5 Stall Handling Capability

The aircraft should be designed with forgiving stall characteristics. This means it should offer natural resistance to wing roll at stall (which often leads to inadvertent spins) and be devoid of deep stall tendencies. The designer can tailor the wing to help promote resistance to roll off (see Section 9.6.4, Tailoring the stall progression). This can be accomplished with geometric wing twist, or washout, but also by selecting a high-lift airfoil as a tip airfoil, even at the cost of cruise performance. While such tailoring is debatable for aircraft such as high-speed passenger airplanes equipped with stick-pushers or shakers, it should be the norm for General Aviation aircraft. Deep stall tendency is prevented by a proper location of the horizontal tail. T-tails are prone to this condition; however, they are often desirable due to stylish appearance or some other reasons, such as to introduce end-plate effects for directional stability or lengthing the tail arm of short fuselages. Such airplanes should be tested to validate that the condition does not exist and, if it does, they should be fitted with adequate ventral fins to remedy it.

23.3.6 Stall Margin for Horizontal Tail

Airplanes of conventional tail configuration may suffer from a horizontal tail stall if subjected to extreme wing downwash conditions, such as those that result from the deployment of high-lift devices. This can increase the angle-of-attack on the stabilizing surface so it stalls. This may cause nuisance such as a broken nose gear and ground-struck propeller, if the HT stalls during flare, to dangerous in-flight handling problems that may cause the airplane to pitch over, nose-first, when flaps are deployed.

23.3.7 Roll Authority

The airplane must meet the desired or the most appropriate roll capability that suits the class the airplane is being designed for. This will prevent the ailerons from being undersized. An aerobatic airplane will have a steady-state roll rate exceeding 130°/sec at its cruising speed and sometimes in excess of 360°/sec. A responsive GA aircraft can roll as fast as 90°/sec and a sluggish one at perhaps around 30°/sec. Roll requirements are stipulated in 14 CFR 23.157, Rate of Roll. The paragraph is split into two parts: take-off and approach for landing, as follows.

23.3.8 Control System Harmony

Design the control system so the harmony in control forces will be aileron:elevator:rudder = 1:2:4. This means that the effort required to actuate the elevator will be about 2 times that of the aileron and the rudder about 4 times that of the aileron. This calls for awareness of how to estimate hinge moments of control surfaces and control system gains.

23.3.9 Climb Capability

With all engines operating, the airplane must have enough power to maintain the following steady climb gradients (also see Section 18.3, General climb analysis methods, on how to calculate V_Y) at sea level:

Normal, Utility, and Aerobatic piston-powered aircraft with W₀ ≤ 6000 lb_f:

These are legal minimums⁶ at maximum continuous power (MCP) conditions at sea level. The airplane may have landing gear retracted and flaps in the take-off position. The value of V_Y for multiengine airplanes shall be greater or equal to the greater of 1.1V_MC and 1.2V_S1. For single-engine aircraft V_Y shall be greater or equal to 1.2V_S1.

Normal, Utility, and Aerobatic pistons and turbine powered aircraft with W₀ > 6000 lb_f:

These are legal minimums at take-off power at sea level. The airplane shall have landing gear extracted unless it can be retracted in fewer than 7 seconds. Flaps shall be in the take-off position.

23.3.10 One-Engine-Inoperative Trim and Climb Capability

Multiengine aircraft, in particular twins, must offer safe trim capability and enough power to climb with one engine inoperative (see Chapter 14, The anatomy of the propeller, for important details). Such aircraft must be capable of straight and level flight (albeit with control surfaces deflected).

23.3.11 Natural Damping Capability

The airplane should be designed such that dynamic stability modes are naturally damped. Diverging non-oscillatory modes may be acceptable, albeit undesirable, provided the time to double amplitude is long enough to allow the pilot to effortlessly correct them. However, oscillatory modes should always be convergent.

23.3.12 Fuel Tank Selector

Airplanes equipped with multiple fuel tanks and a fuel tank selector should always place it inside the cockpit in the field of view of the pilot. Never place it so the pilot has to look away or “feel” to figure out the currently selected fuel tank. A large number of aircraft crash every year because of fuel starvation attributed to a pilot running out of fuel after “thinking” he had selected the correct tank. Also, when flying by Instrument Flight Rules (IFR), vertigo can result if the pilot is forced to turn away from the instrument panel to view the fuel selector. The convenience of the pilot takes precedence over the convenience of the engineer or the technician routing or installing the fuel selector valve.

23.3.13 Control System Stretching

As the primary controls (ailerons, elevator, and rudder) are deflected in flight, the aerodynamic loading will stretch cables and loads will be reacted by pulleys and other parts of the control system. This can result in a substantially smaller deflection than what is measured on the ground (see Figure 23-22). The consequence is a serious reduction in control responsiveness.

Some control systems are so poorly designed that perhaps some 25% of the ground deflection is available in flight. The result is sluggish responsiveness, if not outright dangerous handling characteristics. Consider an airplane initiating a flare maneuver just before landing at a forward CG. This maneuver may require, say, 10° of elevator deflection. A flexible control system could only result in, say, 5°, with the stick fully aft. An unsuspecting pilot might discover this at the time of touch down and the result may be a broken nose landing gear if not worse. This would most probably be discovered in the flying prototype, where a mishap could jeopardize the success of an otherwise promising design.

A typical design mistake is displayed in Figure 23-23. A couple of pulleys, called A and B, have been mounted to brackets (or intercostals) that are attached to a bulkhead. The bracket for pulley A is considerably longer than for pulley B. As a consequence, when the control system is loaded up, pulley A will be displaced considerably, which will cause cable slack, which will results in a less control surface deflection. It is imperative that the control system designer is aware of such pitfalls so they can be avoided.

23.3.14 Control System Jamming

Control system jamming is a serious threat in brand new airplane designs. It is easy to overlook how control surfaces could get stuck (or jam) as they flex due to air loads, or how the deformation of the control system might get cables and pushrods into positions that might get things stuck.

23.3.15 Ground Impact Resistance

Airplanes with forward-facing firewalls should feature a canted firewall (see Figure 23-24). This may prevent the aircraft from “digging in” during a possible nose-down-attitude crash and improve the vulnerability of the occupants that might result from the sudden deceleration from an otherwise survivable mishap.

23.3.16 Reliance Upon Analysis Technology

The current advances in computer analysis methods present a particularly serious pitfall for the aircraft designer. Extremely compelling images generated by such software make it easy for even the most seasoned designer to forget to second-guess the results. However, such results may be way off. Figure 23-25 shows an example of two computational methods as compared to an experiment for the NACA 4415 airfoil as reported in the report NACA R-824 [5]. One uses the widely used code Xfoil and the other is the vortex-lattice code SURFACES. Two important observations can be made by studying the figure.

First, inviscid computational methods such as the vortex-lattice method (VLM) do not predict flow separation. While the VLM results are in good agreement with the experiment, this is only true in the linear range. The speed and flexibility of the VLM make it a great tool for the aircraft designer, but only as long as the airflow is mostly attached. For instance, in this particular example, the VLM results are acceptable over the AOA range of −12° < α < 10°.

Second, while Xfoil predicts a “gentle” stall behavior as shown in the experiment, its predictions are off in slope, and both C_Lmax and α_CLmax are too large, just to name a few. The important point is that it is easy for the aircraft designer to take such predictions at face value and, in this case, greatly underestimate wing area. The best advice is to check results for similar airfoils first and fully understand the limitations of the computational method. For this particular example, it is possible the input data was inadequate or some other explanations apply. Perhaps the user did not select the proper options in the program. These words are not intended to judge the software itself, but rather the operator. Just because a computational method offers viscous approximation does not mean the method is bullet-proof.

Figure 23-26shows stall progression predictions by two Navier-Stokes solvers. The left wing half was meshed using a structured grid, and the right one using an unstructured grid. The image shows flow separation patterns at an AOA of 18°. The disparity in predicted pattern is clearly evident. How do you know which prediction better resembles reality? Neither? Obviously it is impossible to say without experiment; however, an aerodynamicist armed with only one solver has no choice but to rely on it – often erroneously. Be careful – your development program may be at stake.

Always check the results of any computer analysis software by conducting validation tests. When dealing with aerodynamics design, validation is only acceptable when it consists of comparison to actual wind tunnel tests. When possible, select wind tunnel test results for geometry that most closely resembles the one being analyzed. If such wind tunnel test results cannot be found, select a different geometry as long as it is trustworthy. When dealing with structural tests, validation is only acceptable when compared to actual load tests. Do not compare one analysis method to another one, unless it is validated this way. Such comparison will build trust in software and expose the limitations of the methods. Only an understanding of the limitations of a particular computational method will lead to a proper use.

23.3.17 Weight Estimation Pitfalls

The most common pitfall in the estimation of weight is underestimation. This usually stems from the omission of systems whose weight is poorly defined, but sometimes simply from a human trait – optimism. This is a peculiar human condition that, if left unsupervised, can lead to trouble. Inevitably, an inexperienced designer asked to guess the weight of a component inside some weight range is likely to pick the lower end of the range if the component is destined for use in his own design. Thus, if a component is estimated to be within a weight range of, say, 80 to 120 lb_f, the designer is likely to assume it will weigh 80–100 lb_f in his aircraft. Of course, this is less of a problem among experienced designers, but a possible pitfall it is nevertheless. The remedy? Realism enforced by thorough checks by people other than the weights engineer.

23.3.18 Drag Estimation Pitfalls

There are two common pitfalls when estimating drag: over- and underestimation. Drag is notoriously hard to estimate correctly. Overestimation may make an otherwise promising idea appear as a slug on paper and lead to the termination of a project before it even begins. Underestimation can lead to a devastating disappointment once the airplane flies, if not program cancellation and possibly a financial catastrophe. It is often caused by significant protrusions being “left out” of the calculations during the conceptual design phase (for instance see Section 15.5.1, Cumulative result of undesirable drag (CRUD)). The remedy? Ensure the drag estimation is thoroughly checked, if not performed, by experienced people, although this does not always suffice. Estimate the drag of a number of airplanes that appear similar to the one being designed. If the drag estimate of the design deviates greatly from those this should raise a flag. While this discussion pertains to the state of the project before a wind tunnel model is even considered, it is possible the right answer won’t emerge until it has been conducted.

23.3.19 Center of Gravity Travel During Flight

Most airplanes experience a movement of the center of gravity (CG) during flight. This is most often caused by the fuel being consumed, but sometimes by mass being purposely dropped as a part of a mission; for instance a military aircraft dropping ordnance, or an aerial firefighter dropping water, or parachutists exiting. The designer should aggressively plot the CG travel as a function of such weight reduction and ensure that at no time will it travel beyond the prescribed CG limits. An airplane with a wide CG limit will be far more forgiving than one with a narrow one. This is discussed in more detail in Section 6.6.13, In-flight movement of the CG.

23.3.20 Wing/Fuselage Juncture Flow Separation

The juncture where the wing joins the fuselage can often bring a surprising increase in drag. It is one of those areas often left to flight testing to evaluate, but by then it might be too late and require expensive design changes. Established companies will spend resources on developing proper wing root fairings; however, the less established ones will often wait until flight tests to assess the severity of the flow separation in the juncture. The initial geometry of the fairing can be designed keeping some straightforward rules of thumb in mind, although ultimately a Navier-Stokes solver or wind tunnel testing is required to complete the shape, as the flow field in this region is simply too complicated to assess by other means.

23.4.1 Stability and Control – Dorsal Fin and Rudder Locking

As stated in Section 11.2.11, The dorsal fin, the primary reason for the installation of a dorsal fin is to prevent rudder locking, a phenomenon that may occur if the yaw angle of the aircraft becomes too great. The phenomenon is discussed in great detail in the section.

23.4.2 Stability and Control – Ventral Fin and Deep Stall

As stated in Section 11.2.12, The ventral fin, there are two kinds of ventral fins: for deep stall or for Dutch roll. The T-tail configuration of the Lockheed F-104 Starfighter was intended to increase its effectiveness through an endplate effect (Whitford [6]) but other airplanes feature T-tails for other reasons, for instance aesthetics. Some of those require a ventral fin to fix a deep stall tendency as a consequence of the design decision. Reynolds [7] presents a good discussion of the development of ventral fins on the Learjet 55. Figure 23-27 shows the ventral fins on a Learjet 60.

23.4.3 Stability and Control – Ventral Fin and Dutch Roll

A ventral fin can also be introduced to increase the vertical or side silhouette of an airplane for the purpose of improving its Dutch roll damping or increase the directional stability of a multiengine aircraft when operating with one engine inoperative (OEI). Ventral fins intended for this purpose are installed in a far more vertical position than those intended to fix deep stall.

23.4.4 Stability and Control – Forebody Strakes

As was discussed in Section 23.3, Preliminary aircraft design checklist, conventional sizing of the vertical tail is contingent upon critical asymmetric flight conditions: cross-wind landings, balked landing, one-engine-inoperative trim capability, and many others. The resulting fin size can be large and increases both drag and airframe weight. The large tail can, therefore, be costly for transport aircraft in terms of fuel costs. Consequently, alternative methods that do not increase drag are attractive and of great interest to the aircraft designer. Additional benefits of such methods are realized when time comes to “stretching” or lengthening the fuselage of an existing design in order to develop a variant aircraft that can carry more passengers or payload. Such changes are frequent in the commercial aircraft industry. Extending the fuselage will destabilize the aircraft, longitudinally and directionally, and can easily call for enlarged stabilizing surfaces.

Research on forebody strakes has been active for a long time. One of the best-known efforts was run by NASA’s Dryden Flight Research Center, Edwards, CA, in which an originally retired F-18 Hornet fighter was restored and equipped as a high angle-of-attack research vehicle (HARV) in a flight research program lasting from April 1987 until September 1996 [8]. The program focused on AOA well above 30° and identified the shape of the nose as a key player in the lateral stability of the test vehicle at high AOAs [9]. The resulting cross-flow on the forebody is considerable and the resulting asymmetry in pressure on the nose can cause a significant side force. This force acts across a large distance between the forebody and the CG, creating a large yawing moment. The effort successfully demonstrated that directional control could be exerted using deployable strakes installed in the nose. This research led to a cooperation between NASA Langley Research Center and McDonnell Douglas Corporation (now Boeing) to develop strake technology for use with transport aircraft. Of course, such aircraft operate at much lower AOAs, something closer to 8° during approach for landing, so a different kind of strake had to be developed.

Shah and Granda [10] studied the use of forebody strakes to improve the directional stability of a wind tunnel model of a generic commercial jetliner. This was done as part of a cooperative research between NASA and Boeing referred to as Strake Technology Research Application to Transport Aircraft (STRATA). The strake is a flat plate that is mounted on the forward part of the fuselage, of which an example can be seen in Figure 23-28 and Figure 23-29. These were mounted to the 50-series of the McDonnell-Douglas DC-9 when it was stretched from the 30-series aircraft. These were also installed on the DC-9-80, also known as the MD-80. Their findings can be summarized as follows:

23.4.5 Stability and Control – Taillets and Stabilons

Taillets and stabilons are terms coined for the modification made to the twin-engine Beech 1900 commuter aircraft (see Figure 23-30). The design dates back to the late 1970s and early 1980s, developed from the Beech Super King Air to compete with the Swearingen Metro and the British Aerospace Jetstream. It accommodates 19 passengers in a pressurized cabin and is powered by two 1100 SHP Pratt & Whitney PT6 turboprops driving four-blade propellers, capable of 235 KTAS at 25,000 ft. The Beech 1900D is a variant of the 1900 with an enlarged fuselage that offers a stand-up cabin. This was accomplished without changing the dimensions of the original tail.

The increase in drag required more powerful engines, which, combined with the reduction in directional stability due to the larger fuselage, called for the addition of ventral fins and taillets. Each taillet adds about 1.67 ft² of area to the total vertical stabilizing area and improves Dutch roll damping. This prevents the airplane’s yaw damper from being a non-dispatch critical component [11].

The stabilon refers to a couple of 7.75 ft² horizontal ventral fins that are mounted to the lower aft part of the fuselage (again see Figure 23-30). These increase longitudinal stability and allow the center of gravity to be farther aft than otherwise, making the airplane less sensitive to passenger and baggage loading. According to the source [11] (which cites Beech sources) the stabilons increase the aft CG limit from 32% MGC to 40%. They also improve “deep” stall characteristics (see discussion in Section 11.2.12, The ventral fin).

23.4.6 Stability and Control – Control Horns

Control horns are extensions that are placed on control surfaces, such as ailerons, elevators, and rudders, for reducing aerodynamic hinge moments. This is accomplished by adding area of the surface ahead of the hingeline. Control horns come in many sizes and shapes. There are two common types: unshielded and shielded. Shielded control horns are recommended for airplanes that are certified to fly into known icing (FIKI – flying into known icing) as they do not accrete ice like control horns that are more directly exposed to the elements.

23.4.7 Stall Handling – Stall Strips

A stall strip is a small triangular strip made out of metal or rubber that is bonded to the leading edge of a wing to help control the stall characteristics of an airplane (see Figure 23-31). The strip can be seen on the leading edges of most manufactured aircraft, as they must comply with regulations such as 14 CFR Part 23.201, Wings level stall.

The difficulty in maintaining production tolerances effectively renders all serial numbers of a given aircraft type similar and not precisely the same. Yes, all Cessna 172 aircraft look so similar to the untrained eye that they may appear identical. However, in fact, when compared to the intended outside mold line (OML) they all deviate in one way or another, albeit in minute ways. The same holds for all other mass-produced aircraft. For this reason, when airplanes come off the production line, they are test flown before delivery by the production flight test team and stalled to confirm stall behavior. Then, based on the observation of the test pilot, stall strips are bonded in a specific location on the leading edge.

Stall strips are detrimental in that they may cause an early separation even at climb AOA. This will be manifested as a reduction in climb performance. They can also be detrimental on NLF wings as they may destabilize the boundary layer aft of their location, causing an earlier transition than otherwise.

There are primarily two positions the aerodynamicist needs to be aware of and are important to a properly located stall strip. spanwise station and clocking (see Figure 23-32). During the development phase of a new aircraft, the location of stall strips on the wing becomes a process of trial and error. The experienced aerodynamicist will create a test matrix with potential locations and clocking positions and have the test pilot stall the aircraft and note change using terms like “good,” “bad,” “improved,” or similar. The essential task is to use information from each stall to its utmost as each flight is costly. Using CFD may help, provided there is evidence to support it correctly predicts flow separation. Such evidence should comprise pictures of the flow field and separation area as seen using a tufted wing and subsequent comparison to CFD predictions. In the absence of such evidence, CFD should be used with care great or not at all – incorrect predictions may simply muddy the waters.

23.4.8 Stall Handling – Wing Fence

A fence is a vertical panel mounted to the upper surface of swept-back wings (see Figure 23-33). The invention is attributed to research from 1937 by the German aerodynamicist Wolfgang Liebe at the Deutsche Versuchsanstalt für Luftfahrt [12, p. 224] (or the DVL – The German Aeronautical Test Establishment). This work was carried out in order to help remedy dangerous wing roll-offs during stall of aircraft such as the Heinkel He 70 and Messerschmitt Me 109.

Early versions of the fence extended from the leading to the trailing edge, while the more modern configuration extends a relatively short distance back. The wing fence is a less common solution today than before and is mostly found on some business jets (e.g. Cessna Citation III, Beechcraft 400 Beechjet, Learjet 35, and Gulfstream II). Earlier commercial types featuring wing fences include the Boeing 727 (see Figure 23-33), Comet 1, Sud-Est Caravelle, and Tupolev Tu-154, to name a few. Wing fences can also be found on many fighter aircraft, such as the MiG-15, North American F-86 Sabre, and even relatively recent aircraft such as the Hawker Hawk (US designation T-45 Goshawk), Fiat G91 and SIAI-Marchetti S211. It can even be found on GA aircraft such as the German Akaflieg Braunschweig SB-13 sailplane and the Australian Eagle 150 tandem wing aircraft.

Fences are sometimes referred to as “boundary-layer fences.” However, a fence can serve in a number of important ways:

23.4.9 Stall Handling – Wing Pylons

Jet engine pylons on the wing have been shown to improve the stall characteristics of jet aircraft. Abzug and Larrabee [1, p. 174] explain it by pointing out that the bound vortex on the wing induces side-wash on the engine pylon, as shown in Figure 23-35. Consequently, an outboard side force is generated on the nacelle and pylon. The pylon generates a vortex on the upper surface of the wing, whose flow direction can be seen to oppose the outboard spanwise flow. This suppresses early wingtip flow separation.

It is alleged in Ref. [1] that the effect was discovered by Boeing, most likely when developing the B-47 Stratojet in the late 1940s. Both the Boeing 707 and B-52 Stratofortress feature two pylons per wing and neither had wing fences, so common on swept-wing aircraft of the era. This spurred the development of another aerodynamic fix – a small “pylon” without the nacelle. The device was conceived by Douglas Aircraft aerodynamicists and was mounted to the DC-9, whose engines are mounted at the back of the fuselage. The device was patented under the name vortilon. It was developed as a part of a three-part solution to a possible deep-stall scenario on the DC-9 [13] and mounted relatively inboard, where the vortex it generated at high AOAs favorably improved flow over the horizontal tail. The other two were an enlarged horizontal tail and hydraulically boosted nose pitch-down control.

23.4.10 Stall Handling – Vortilons

As explained above, a vortilon is a small pylon-shaped fence mounted to the lower surface of a wing, usually a swept-back wing. It can be mounted alone (as it is on the inboard wing of the DC-9 commercial jetliner) or in a row (as it is on the outboard wing of the ERJ-145, as shown in Figure 23-36). Its purpose is to shed a vortex at high AOAs that aerodynamically partitions the wing into several low-AR segments, reducing section lift coefficients and delaying stall (see Figure 23-37). It is not as effective as the fence, but adds less drag to the airplane [7].

23.4.11 Stall Handling – Wing Droop (Cuffs, Leading Edge Droop)

The term wing droop (also called a cuff or a leading edge droop) is used to describe the enlarging of wing airfoils in the outboard wing region (see Figure 23-38). The cuff presents a discontinuity to the leading edge of the wing. This means that as the AOA increases, the wing is effectively partitioned into two smaller segments. This can be seen in the vortex that begins to form at the discontinuity. The wing partioning results in a reduced AR of the outboard (and inboard) segments, not unlike the one obtained using the wing fence. As we have seen before, a reduction in the AR of a lifting surface means a more shallow lift curve slope and delay of the stall to a higher AOA. Consequently, the cuff provides improved roll stability at stall.

In 1982, NASA conducted an investigation on the effectiveness of leading edge devices for stall departure and spin resistance on the Piper PA-32 Lance with a T-tail [14]. The investigation revealed that the partial span leading edges eliminated the abrupt stall tendency of the 1/6th scale wind tunnel model and a tendency for autorotation, improving spin resistance. While some of the results are dependent on the geometry of the airplane and Reynolds number, they are still important indicators of the potential benefits of cuffs. Among results were:

The most serious drawback of wing droops is that they may make spin recovery harder. The feature that improves roll stability at stall – higher lift at the wingtip – is also larger during spin, requiring greater moment to stop. Among other drawbacks is higher drag and added manufacturing complexity. Cuffs for composite aircraft are typically manufactured by two means: as an integrated structure or as an add-on. Integrating the cuff into the wing skin is desirable, as it reduces part count and weight. However, the quality of the laminate layup may suffer due to fiber dryness often associated with sharp corners.

23.4.12 Flow Improvement – Vortex Generators

Strictly speaking, a vortex generator (also known as a VG) is any device that starts and maintains vortex motion in fluid flow. However, generally, the term applies to devices that generate localized vortices on lifting surfaces to improve the airflow over them. There are typically two reasons for their installation; to alleviate buffeting and loss of control effectiveness due to shock interaction at transonic airspeeds [15] or to suppress flow separation at high AOAs at low subsonic airspeeds. The latter will be the focus here. Figure 23-39 shows a common application of VGs: on the lower surface of the HT where it increases the effectiveness of the elevator by delaying separation at high deflection angles and low dynamic pressures.

There is a large amount of information available on VGs in technical journals, like NACA, NASA, AIAA, ESDU, and others, so only basic information will be given here. A good insight is given by Wentz and Seetharam [16], who investigated the effect of VGs on a GA(W)-1 airfoil. They found the VGs to be very effective in suppressing flow separation. The VGs increased the C_lα and C_lmax, but lowered the α_stall. There was a substantial increase in the section drag coefficient, of about 25 drag counts. However, the drag at higher AOAs was lower due to the delay in separation, although the cross-over point was at a C_L of about 1.1. Guidelines for the sizing of VGs are given below and are obtained from various sources, including Refs 15, 16 and 17.

Very small VGs, called micro-VGs, are used in applications where size, or lack thereof, is of importance; for instance to improve flow over control surfaces and flaps. An example is shown in Figure 23-40. Their purpose is to delay flow separation over a specific control surface. In the case of flaps, they have the potential to increase C_Lmax while having no impact on cruise drag, as they are only exposed to the airflow when the flap is deployed. It is a drawback that they may cause an interference with the free motion and stowage of the flap. Refer to Lin [18] for recommendations regarding sizing, spacing, and location.

23.4.13 Trailing Edge Tabs for Multi-Element Airfoils

Ross et al. [19,20] investigated the installation of a special lift-enhancing tab (or a cove-tab), similar to a Gurney flap on the pressure side of a two-element airfoil (NACA 63₂–215) with 30% chord single-slotted flap and on a multi-element airfoil. The idea has been patented under US Patent 5,249,080.

A cove-tab with typical dimensions is shown in Figure 23-41. The research indicates the tab must be sized and located based on the geometry of the flap and cove. For this reason, the dimensions shown in the figure do not necessarily hold for all applications. The tab also presents an issue during flap retraction and deployment and may, therefore, have to be mechanically actuated, potentially adding complexity to the installation. It is an advantage that the cove-tab is passive – in other words, it does not require a separate power source to function, like for instance a jet-flap.

The installation and wind tunnel testing of the cove-tab on the two-element airfoil was the focus of Ref. [19]. It was found that it increased the maximum lift coefficient by 10.3% with the flap deflected at 42°, attributed to a reduction in flow separation on the upper surface of the flap. The tab was found to be less effective (even detrimental) for lower deflections. However, a combination of a cove-tab with a Gurney flap and vortex generators on the flap increased the maximum lift at a high flap deflection by 17% over the optimum flap position of the baseline.

Reference [20] investigated various flow mechanics of the cove-tab installation using a Navier-Stokes solver. Among notable conclusions is that as the flap element is moved away from its optimum position the flow separation increases (which explains the reduction in lift). Such a movement can be the consequence of tolerance stack-up or, simply, incorrect positioning during design. Adding the cove-tap will reduce the separation and restore the flap lift to within 1% of the optimum flap lift. This way, it can serve as a “fix” for incorrectly positioned flaps.

The investigation revealed the tab is detrimental to the lift at low flap deflection (T-O configuration) as it appears to reduce the suction peak on the leading edge of the flap. However, at large flap deflections (landing configuration) it delayed flow separation on the flap, greatly improving its capability.

23.4.14 Flow Improvement – Nacelle Strakes

Nacelle strakes are vortex generators commonly found on the engines of modern jet transport aircraft, civilian and military. At high AOAs the strake generates a powerful vortex that makes up for the flow separation and loss of lift due to the presence of the nacelle. The strake typically has a LE sweep of approximately 70° and it is aligned with the airstream at cruise to minimize its interference drag. It can improve the maximum lift coefficient by as much as 0.05–0.1. An example of a nacelle strake on an Airbus A319 commercial jetliner is shown in Figure 23-42.

23.4.15 Flow Improvement – Bubble-Drag, Turbulators, and Transition Ramps

The wings of small radio-controlled aircraft and sailplanes that operate at very low Reynolds numbers (60,000 < R_e < 500,000) are often subject to the formation of a laminar boundary layer separation bubble along the leading edge. An example of such a separation bubble is shown in Figure 8-18. Such bubbles can be very detrimental to the flight characteristics of the corresponding aircraft, detrimentally affecting the lift and drag (forming so-called bubble-drag). They can even lead to unpredictable and sudden changes in the magnitude of these forces. Generally, two methods can be used to reduce the bubble-drag: (1) turbulators and (2) tailoring of the transition curve.

Turbulators (also called trip strips or boundary layer trips) are often installed near the leading edge of such aircraft. They are intended to force the laminar boundary layer to transition immediately to a turbulent one to prevent the formation of or reduce the size of the more detrimental separation bubble. The presence of a turbulator has important effects on the lift and drag of the airfoil. It is difficult to accurately predict its influence, calling for trial-and-error approaches in wind tunnel or flight tests. This is further compounded by the fact that a trip strip configuration found suitable for a specific airfoil at a given AOA may be detrimental to it at another operating condition [21].

Tailoring the transition curve refers to the proactive design of airfoils to encourage a more rapid movement of the laminar-to-turbulent BL transition point on the upper surface toward the leading edge. This is shown in Figure 23-43, which depicts how the derivative is steeper for airfoil A than B. This is referred to as the shape of the transition ramp. In other words, in the lift coefficient range 0 < C_l < 1.0, the transition point, X_tr/C for airfoil A “moves faster” toward the LE than airfoil B. Gopalarathnam et al. [21] show that this results in a lower bubble-drag for airfoil A. This observation makes it possible to design airfoils for operation in low Reynolds number flow that are less likely to suffer this undesirable type of pressure drag.