Category Model Aircraft Aerodynamics

In Figure 9.5 is shown a typical curve of profile drag plotted against ci for any of the NACA 6 series aerofoils. At the design ci, drag is much lower than for an orthodox or old – fashioned section. On either side of this value there is a low drag range or ‘bucket’ in the graph, so that small departures from the ideal operating conditions cause no change in profile drag coefficient At either side of the low drag bucket, on one surface or another, the velocity distribution changes and the boundary layer becomes turbulent, with an associated sharp rise in drag. In the NACA designations of these aerofoils, the third digit,

of thick and thin profiles

usually written as subscript thus: NACA 643418, indicates the width of the low drag bucket, in this case 0.3 ci on either side of the ideal designed value for the profile. A profile with the subscript 3 as above, designed for a ci of 0.4, will work efficiently at c) down to 0.1 and up to 0.7. (Note, however, that constant drag coefficient does not mean constant drag force – the higher speed associated with lower Cl of the wing increase drag force at a constant Cd – see Chapter 2.) The fourth digit of the aerofoil number gives the design ideal lift coefficient, in tenths, and the final two figures give the profile thickness as a percentage of the chord.

As already noted from the velocity profiles of the thick profiles of Figure 9.3, favourable flow conditions are preserved on thicker aerofoils over a greater range of ci than on thin ones. The absolute minimum drag of a thick profile is slightly more than for a thin section of similar camber, but the drag bucket of the thick profile is wider. This is indicated in Figure 9.3. Such a thick wing has a wider speed range, and, in the case, for example, of a racing model, will be less affected by slight inaccuracies of flying, and less slowed down in steep turns, than a model with thin wing and higher maximum speed straight and level.

As the Reynolds number is reduced, so long as flow remains super-critical (i. e. re­attachment after the separation bubble), the natural tendency for laminar flow to persist shows up. The minimum drag of the laminar flow profile is slightly higher (because the relative viscosity of the air at low speeds is greater compared with the density-speed – chord factors), but the boundary layer, after passing the maximum velocity point on the wing, remains laminar for a greater distance and this has the effect of widening the drag bucket slightly. The result is shown diagrammatically in Figure 9.6.

Aerofoils are no longer designed by ‘cut and try’ methods, but are worked out to fit their special purposes. The first substantial gains achieved were the NACA ‘6’ series aerofoils developed before and during the Second World War. They were used, in slightly modified form, first on the P-З I ‘Mustang’ fighter. These aerofoils were designed to achieve very low profile drag by preserving laminar flow over as much of the wing as possible. The improvements in practice were less than hoped for, because of the inaccuracies of the wings in service, but there were genuine overall benefits. The main method of achieving

Fig. 9.2 cont.

NACA 0009 LE RADIUS 0.89 PERCENT NACA 0010 LE RADIUS 1.10 PERCENT

Note that the ordinates of the 9% thick profile are exactly 90% of the 10% profile. NACA four-digit symmetrical sections may alwaysbe scaled up or down to different thicknesses by simple arithmetic.

the lower drag was to employ aerofoil thickness forms similar to those shown in Figure 9.3. As the velocity diagrams show, at zero angle of attack the maximum velocity/: minimum pressure point on these profiles is at 40% or 50% of the chord. Other thickness I forms were designed with this point further back or further forward. The second digit of an, NACA profile designation, such as the 4 in 643618, indicates the position of the maximum velocity point The boundary layer, on a suitably smooth wing, will remain : laminar to a point somewhere aft of the 40% chord position on such an aerofoil. (By і suitably smooth is meant a wing free from ripples, humps or hollows rather than one which is highly polished.) As shown in Figure 9.3, at small angles of attack the velocity ‘ distribution on both surfaces is favourable for laminar flow back to the 40% point, the 9% thick profile at ci of.06. The thicker profile, 652015 shows that laminar flow is preserved on both sides up to a ci of 0.22. This is of great importance. A thicker wing at a high angle of attack may have greater percentage of laminar flow and hence lower drag than a thin profile. (Compare also the 653018 profile.) This applies equally to cambered profiles. If j one of the symmetrical sections of Figure 9.3 is cambered round the NACA a = 1 mean line, the basic character of the velocity distribution, and hence laminar flow, is not І changed. Figure 9.4 shows the results in graphical form. The amount of camber given to ‘ the profile is determined by the desired operational Cl, as described in Chapter 7. At this value, laminar flow prevails over both upper and lower surfaces, up to the peak velocity ■: position and slightly beyond it. At a higher angle of attack, the velocity graph resembles that of Figure 9.4(b). Transition on the upper surface occurs futher forward. There may even be flow separation further aft, but this is not important since the aircraft is not intended to operate far from its designed Cl – The result in terms of drag at the design ci of the profile is very substantial improvement

9.1 VELOCITY AND PRESSURE DISTRIBUTIONS

Aeromodelling has undergone a revolution since the time of F. W. Schmitz. Free flight models still operate close to critical Reynolds number conditions but radio controlled models, especially the larger sailplanes, high speed racers and aerobatic models are usually outside the danger area. Many such models fly quite successfully with profiles similar to the Clark Y or Gdttingen 796 and it is obvious that problems of sub-critical flow separation have been left behind. Part of the reason for this is that such models do not usually fly at very high angles of attack. As Schmitz’s results showed, a profile like the N-60 operated efficiently at a low angle of attack but stalled early even when operating, nominally, above its critical Re. The same thing was found on the Go 801, separation problems at high angles of attack did not entirely disappear until about Re 170,000. Modellers have put up with the premature stall of such profiles. Performance at higher speeds is quite good. The pylon racing model is in any case operating most of the time at Re’s quite comparable with those of full-sized sailplanes and even some light aeroplanes. Very great improvements in performance have been achieved in full-sized sailplanes, and some powered aircraft, by the use of so-called laminar flow profiles. The advantage in terms of saving skin friction are very large, especially at high speeds where profile drag becomes of major importance (Fig. 4.10). Early work in this area by the Low Speed Aerodynamics Research Association has been largely overlooked by modellers, but the aerofoils LDC2 and LDC3M produced were used on some models at the time, about 1948, when they first appeared.

As described in Chapter 3, on any wing, the boundary layer will be laminar at first near the leading edge, but will make a transition to turbulent flow when it reaches the critical boundary layer Re. The value of this critical Re will depend on the quality of the wing surface. Many full-sized aircraft have poor surfaces. Even if highly polished, there are waves and ripples in the skin caused by rivet tension and humps created by stiffeners and spars. It is very difficult to preserve laminar flow over such a wing for more than a few centimetres near the extreme leading edge, and even when great efforts are made to achieve an accurate profile, the large Reynolds number associated with the high flight speed and large wing chord, promotes early transition. The smallest defect in the surface, even a fly speck or the crushed body of an insect, can cause transition. For these and similar reasons designers of full-sized light aeroplanes have not been able to achieve all the benefits of low drag, laminar flow, and many still prefer to use relatively old-fashioned profiles. However, as the Reynolds number falls, the chances of preserving laminar flow over more of the wing increase. Small defects that, at higher velocities, cause transition,

may be over-ridden by the relatively more viscous boundary layer, and if wings can be made accurately, as they are in modem full-sized sailplanes, the aerodynamicists’ predictions, based on theoretical studies and wind tunnel tests, come true. The problem for modellers is the reverse of that for the turbulent flow aerofoils. At sub-critical Re, turbulators and protrusions caused by spars may improve performance by forcing transition in the boundary layer. Laminar flow models should seek to maintain profile accuracy at least as far back as the point where the boundary layer will make its transition naturally. The standard of precision required is, because of the low Re, less than that needed for the full-sized aircraft.

Providing the wing surface is smooth, a laminar boundary layer will tend to prevail as long as the speed of flow is rising under the influence of the low pressure area above the wing (Fig. 3.5). Behind the minimum pressure point the laminar flow persists for some distance but then a separation bubble forms and (providing super-critical Re prevails), the flow re-attaches as a turbulent boundary layer. In Figure 9.1 velocity measurements made on two wing profiles are shown. These show how the speed of flow over the upper and lower surfaces vary at different angles of attack. On the Gottingen 389, for example, at 2.8 degrees angle, the flow on the upper surface increases rapidly to a maximum within the first ten percent of the wing chord. Laminar flow will persist up to this point and a little way beyond it, but then transition will occur and turbulent, high drag flow covers most of the wing. At a lower angle of attack, -3.1 degrees, the maximum velocity point on the upper side is somewhat further back, but at higher angles it moves forward so that near the stall, at 14.6 degrees, the whole upper surface is in turbulent flow. Meanwhile, on the underside the velocity peak of the upper surface is opposed by a velocity decrease close to the leading edge, and thereafter the air accelerates towards the trailing edge. Transition

will already have taken place and the boundary layer will be turbulent. At the more negative angles of attack, the roles of upper and lower surfaces are reversed as the wing begins to ‘lift’ downwards.

The Gottingen 382 profile is a thicker version of the 389, but the velocity measurements show important differences. At 2.7 degrees angle, the velocity maximum on the upper side is about ten percent further aft than on the thinner profile, and even near the stall there is a likelihood for laminar flow back to about ten percent before the velocity decrease causes transition.

The details of the velocity of airflow over a wing depend on both its thickness form and its camber. The typical, pre-1940 aerofoils of Figure 9.1 and others of the same vintage, such as the Clark Y, N-60, etc. were designed around what was then thought to be the ideal form for any streamlined body. The same basic shape, thickened or thinned appropriately, was used for strut fairings, streamlined wires, tailplane and fin profiles, wheel spats and whole airship hulls. Typical ordinates are given in Figure 9.2 which refer to the NACA ‘four digit’ aerofoils. The velocity distribution graphs given with the profiles show that at zero angle of attack, the velocity peak (and hence minimum pressure point) on both surfaces is reached at about ten percent of the chord. A short distance aft of this, transition to turbulent flow occurs. Cambering changes this basic feature only slightly; such profiles are fundamentally incapable of preserving laminar airflow over much of the wing.

In measurements of boundary layers in low turbulence wind tunnels, it has been recognised for some time that noise alone can cause a delicate boundary layer flow to change sharply. Noise generated by the fan or motor of a wind tunnel can provoke early transition of a laminar boundaiy layer, and even the sound of someone walking past the test section may cause a change. Elaborate tests have been made with artificially generated sounds of varying pitch and volume, which show that separation and stalling can to some degree be controlled by this means. Sound is, basically, a series of small compression waves in the air and this may be enough to change the microscopic turbulence. Alternatively, the air noise may cause sympathetic vibrations in the solid wing skin, which could cause the boundary layer flow to change.

In practical model flying it is quite likely that the noise and vibrations caused by the engine and propeller cause the boundary layer to make transition to turbulence sooner than would occur on a sailplane, for instance. (It is highly unlikely that shouting, screaming or even singing at a model sailplane will have any effect!)

9

The fineness of the tape used for turbulators in all wind tunnel testing is worthy of note. Modellers have sometimes employed much thicker ones, sometimes using strips of limn or 1.5mm balsa where thicknesses one tenth of this should be sufficient Turbulators, or; invigorators, which are too thick can cause flow separation rather than improving the boundary layer conditions.

■І

8.3 ZIG ZAG TURBULATORS

There is much to be said for laying the tuibulator tape in zig-zag fashion. Tests at Delft і University have shown that a zig zag tape tuibulator just ahead of the separation point on the wing has a better effect than a straight strip. Hie best spacing of the zig zags is a matter i for experiment on a given wing. The best results are obtained when there is a definite { relationship between the natural tendency of the flow within the separation bubble to develop waves and small, chordwise vortices (see paragraph 3.9). The modeller is unlikely to know what this is without very costly tunnel tests and some calculation, so trial and error is the likely way of discovering the optimum arrangement Possibly pinking shears of different sizes could be used to produce tapes for trial.

8.4 PNEUMATIC TURBULATORS 1

Perforations, such as a row of pin holes through the wing skin instead of a tape strip, can | act as a tuibulator. 3

The air pressure inside a wing is usually somewhat greater than that on the upper 1 surface, so air is drawn through the perforations and injected into the boundaiy layer. Ibis ] in effect trips the flow and may be sufficient to make it tuibulent (The effect was first j noticed by M. M. Gates in wind tunnel tests carried out in the 1950s.) Many full-sized і sailplanes also use pneumatic boundary layer control, especially on the underside of the wing near the trailing edge, where a separation bubble commonly forms. High pressure air is taken in by a small intake, positioned under the wing, and this raises the pressure in the і hollow chamber inside the wing. Very fine holes are drilled, at small spacing, through the skin just ahead of the separation point of the bubble. The injected air blows the boundary layer off the wing altogether and reduces the profile drag. Keeping the many small holes open is a problem of maintenance.

Research by Martyn Presnell in a wind tunnel at Hatfield has shown that considerable improvements in the performance of free-flight model sailplanes and rubber driven aeroplanes can be achieved by the use of multiple ‘trip strips’ or, in Presnell’s terminology, ‘invigorators’.

Test wings using the Benedek 6356b aerofoil (see Appendix 3) were constructed from materials exactly like those used in a typical FI A (A2) sailplane model. Balsa wood wing ribs and spars were used, the framework being covered with tissue paper, doped, and in one case, the forward third of the wing was skinned with thin sheet balsa. Not only were lift and drag forces measured, but some flow-visualisation tests were done. These involved coating the test wing with pigmented kerosene to reveal the nature of the boundary layer. Where the b. l. was turbulent the kerosene evaporated rapidly, leaving a film of pigment. Within the laminar separation bubble, the evaporation was less rapid so the flow of the air nearest the wing skin could be seen as the liquid moved upstream (See Figure 3.6). In the fully laminar flow regions the kerosene remained liquid longer still and flowed in the normal downstream direction. The flow separation point and re-attachment downstream of the bubble could then be discovered for each angle of attack. (Modellers have sometimes noticed that, when flying in the late afternoon or early evening at dewfall, dew deposited on a wing before flight will still sometimes be present after the flight on the leading edges where the flow is laminar, but evaporates from the rear parts of the wing where turbulent boundary layers are expected.)

The addition of a single turbulator at 5% of the wing chord improved the measured lift and drag figures, as expected, at Reynolds numbers below 40,000, although the separation bubble was still present. The turbulator consisted of a thin strip of adhesive plastic tape 0.15mm thick and 0.75mm wide, running spanwise.

It was then found that the addition of further strips of the same thin tape at various positions on the chord aft of the turbulator resulted in further improvements of lift and drag figures. The best results at Re below 70,000 were found with five of these invigorators in the positions shown in Figure 8.8. The original 5% turbulator remained in place throughout

Presnell noted that placing an invigorator within the separation bubble, as revealed by the kerosene, made no detectable difference. The first invigorator must be placed just aft of the re-attachment point and the others spaced over the rear part of the wing in the turbulent boundary layer. The exact mechanism of the invigorators is not fully understood at present. It may be that they aid the already turbulent boundary layer to remain attached to the wing after the bubble has been passed. Presnell points out that several leading contest model fliers have used invigorators with success.