The stressed-skin monoplane
The last chapter ended with a review of the most up-to-date types of 1936 because that was a critical year in the development of the shape of the aeroplane. It is also a critical stage in this book, for since this was the starting point for the aeroplanes of today, each specially developed for its purpose, with the appearance of each class diverging more and more, it seems the time to change the form of this book. From now on, each chapter will be devoted to the progress and changes of thought from which the characteristic shapes of today have been evolved. But first, it is necessary to review the general design and construction of this new generation, for without an understanding of the basic principles underlying design it is impossible to appreciate fully the whys and wherefores of the many curious shapes in the air today.
Perhaps it may even seem that this description comes rather late, but if it had been earlier it would have broken the thread of narrative.
First, there is this question of the stressed-skin structure: many designers revised their ideas of what an aeroplane should look like without altering the structure beneath. After all, doped linen fabric is the lightest covering you can get, and, if it is properly fastened down, it will carry quite surprisingly high loads—in fact Vickers-Armstrongs claimed that their system would stand 1,000 lb. per square foot. Since fabric is so strong and is also cheap, it is scarcely surprising that it remained in use for many years. The more thoughtful, however, realized that if they retained the old ideas of a structure and a covering they would meet many difficulties: for instance, although fabric is cheap in first cost, it has to be replaced every two or three years and "re-bagging" is an unpopular expense with operators; although fabric is easy to patch, it is also easy to tear; both in civil and military operations the increasing number of aeroplanes in service meant that hangar space was becoming scarcer and more expensive and aeroplanes had to spend most of their time in the open where fabric deteriorated much more quickly than metal. These were some of the factors leading to the demise of the fabric-covered structure.
Developing the stressed-skin structure, in which the metal outer surface carries a large part of the loads on the aeroplane, was no easy task—in fact even today there are many different styles, indicating different opinions as to how the job can best be done. Let us look at the problem facing the designer of a simple single-engined fighter. In the first place he has a wing to give lift, a fuselage to carry the load and the engine, a tail for stability. Forces on the wing and tail are what is known as spread loads, being due to air pressure over the whole of each surface; while the fuselage has to carry concentrated loads from the engine, pilot, fuel, undercarriage and so on. Now these two types of loading obviously require different treatment, just as do the essentially different shapes of the fuselage and the flying surfaces.
You can think of the fuselage as a tube. Anyone who has used modern furniture or tried to bend a cardboard roll realizes how much strength there is for little weight in a tube. The reason for this is that, although you cannot see it, when you try to bend a bar or tube it is the metal nearest the surface which resists your efforts most strongly. That is why you nick the surface first if you want to weaken the bar in order to bend it at a particular spot. With a metal-covered fuselage it is the skin, therefore, that gives the strength against bending and twisting (torsion). The amount of skin metal needed to carry the loads is very small and if these were all it had to carry skins need only be perhaps ^in. thick. However, these thin shells are liable to buckle, for instance a load from below will stretch the skin there and compress or wrinkle that on the top. To prevent this, the skin is attached to a framework made up of rings and longitudinal members—not unlike the ribs and stringers of a ship. The rings are called formers, frames, or, where they carry special loads, bulkheads. The longitudinal members are called stringers (after the shipping term, largely because they originated in flying-boat hulls) if they are continuous and pass through holes in the frames, and intercostals (also a nautical name) if they are short pieces joining each frame to the next. Generally speaking, the earlier fuselages were made with intercostals, but nowadays stringers are almost universal.
So much for the tubular, or stressed-skin part of the fuselage. Its strength depends first and foremost on the distribution of all loads, so when we come to concentrated weights, like the engine, special treatment is needed. The engine usually has four feet, which are carried on bearers made up as box members, or with steel tubes, that are attached to fittings on the front bulkhead of the fuselage. These fittings, usually four, must take the dead weight of the engine on the ground, perhaps a ton or more; its pull in flight, this also may be as much as a ton; and in aerobatic flight loads equivalent to four or six times as much as the engine's weight caused by "g", that is centrifugal accelerations . Now the loads from these fittings have to be transferred into the stressed-skin structure so that it is absorbed into as large an area as possible. This is usually done by running a rather heavier stringer back from the fitting and also by increasing the skin thickness locally, possibly with a doubler plate.
In the earlier metal fuselages, tradition preserved the longerons of the old "stick and string" era, because the presence of four rather heavier stringers was useful to support the engine and other dead weights. The strength of each part must be very accurately calculated, so that no unnecessary weight is carried. To do this the engineer works first with the loads he knows, engine weight and thrust, plus the anticipated flying loads due to speed and manoeuvring. These latter are less predictable since they depend, among other things, on atmospheric conditions such as bumpiness, and on the ability of the pilot, who may be hamhanded. To allow for these variables the designer applies a factor to his estimated needs, making his aeroplane perhaps as much as eight or nine times as strong as the maximum anticipated load if it is a fighter. This precaution
is called the "factor of safety" and is laid down by each country 's military or civil licensing authority.
A word about the strength calculations may be of interest. If the fuselage, say, is a simple tube the work is simple arithmetic that anybody almost could do. But for every change in shape, for every concentrated load superimposed on a spread load, for every window opening, the work becomes more and more complicated and involves mathematical tricks to reach a solution. When it comes to wings, the variation in pressure over the surface, their complicated shape and the drastic effects of
vibration and flutter require the most advanced applications of the calculus for their solution—in fact the equations for wings are generally recognized as being the most difficult in engineering.
Why should the wing be such a difficult structure to stress? First of all, it carries the whole weight of the aeroplane in flight, spread over its surface in varying proportions, and acting upward through the narrow width of the structure. If you think of the wing in terms of a very flat streamlined tube it is obvious that the main loads are acting at right angles to the strongest part of the structure. This is, in essence, the problem of the wing designer. Everyone today is familiar with the way an aerofoil gives lift and how the lift is greater than
the drag. Usually, however, when looking at this, one is more concerned with its aerodynamic effects. In fact, of course, the designer must also see it in terms of structural strength—as sheet metal, bar, bolts and rivets.
Before going on to explain how this is done, it is as well to look at some more of the aerodynamic effects, for they are very important in influencing the designer. First, there is the movement of the centre of pressure, the point at which the loads on the wing act. The c.p. moves back quite a long way as the incidence of the wing increases, which naturally means that it puts a twisting load on the wing. When the ailerons are deflected , they too impose twisting loads, one upward the other
downward. The ailerons are also the root of one of the most dread troubles of all—flutter. At high speed the air starts to move the aileron on its hinge, first up then down, and the movement builds up, getting larger all the time until, if the ailerons do not break off, the wing starts to vibrate in sympathy . It is the same story as the soldiers marching over a suspension bridge, where if they did not break step the bridge would start a harmonic vibration, each surge of which would add to the next until the structure broke. Simpler still, it is like giving a push to a swing at the end of each sweep.
It was in the early thirties, after a series of accidents to fighters, that a British scientist, A. W. Pugsley of the R.A.E., provided the fundamental solution to the flutter problem. This was quite a simple device, as is so often the case. It simply consisted of hanging a weight ahead of the hinge so that the "moment" of the weight equalled that of the control surface, that is the weight multiplied by the distance from the hinge is the same as that of the aileron weight times its distance from the hinge. The effect of the weight is to damp down the nutter as soon as it starts—like equal weights on a balance scale. These mass balances, as they are called, were very successful and were soon made compulsory for all controls on British aeroplanes, other countries following fairly quickly because no patent rights were claimed on this safety invention. The mass balances were usually lumps of lead mounted on the ends of long tubes and for ten years were a distinctive feature of controls. By simple arithmetic, the longer the mounting tube, the lighter the mass balance; but where speed counted more than weight the mass balance was submerged inside the leading edge of the control surface. Mass balancing was the reason why control surfaces were fabric-covered for many years after the stressed skin was established, since this helped to reduce the weight aft of the hinge.
To return to the wing itself. In the cantilever monoplane, the whole strength of the structure must come from the wing. The lifting forces produce upward loads which tend to bend the wing. Now the "tube" is too shallow for the skin to resist these loads very much and spars have to be inserted to do the work. These spars are sometimes an inheritance from the old biplane, usually there are two, but there may be only one, or there may be several, or, again, there may be a single box spar. With two spars, the single spar and the multi-spar the wing loads are divided into units as it were. The spars take the bending, while the ribs and skin are combined to support the drag and twisting forces.
The box spar is a more homogeneous unit and it is gradually superseding the individualistic spar and skin systems. The usual name for the box spar is a torsion box, which indicates that it is made to resist twisting. In its elementary form the box consisted of two spars with a thick skin between them. As knowledge has grown about the behaviour of these structures the sides of the box have become simpler and the upper and lower skins thicker, so making the box a more uniform unit.
Advantages of the structure are that, with the thicker skin, ribs are reduced in number and the hollow box is used as a fuel tank—an almost universal arrangement for modern air liners —while the thick skin resists damage when refuelling and during ground handling generally. To the torsion box are attached relatively light leading and trailing edge sections. BULKHEAD
The Spitfire had an unusual application of the box spar. The main structure was an enormously strong D, made from a heavy spar with a heavy curved skin forming the leading edge. The wing aft of the spar was a light metal-covered version of the old biplane structure with a light, or "false", spar to which the flaps and ailerons were attached. This wing of the Spitfire was one of its great features and was, in its day, an ingenious way of making a very thin elliptical wing of great strength.
Complications are added to the designer's work when making a wing by the need for attaching it to the fuselage and catering for ground loads. If the wing has plain spars, the loads are usually coaxed into them by means of large fittings at the roots and these are then bolted to mating fittings attached to specially strong fuselage bulkheads. Sometimes, as in the Spitfire, the wing is bolted directly to the fuselage, and in other cases, the Hurricane is an example, there is a centre section, to which the wing is bolted, and this in turn is attached to the fuselage. It is largely a matter of preference, though if there is a centre section the undercarriage, which is attached to it, makes the fuselage mobile—a useful feature during manufacture and maintenance.
The practice, which really came in with the low-wing monoplane, of mounting the undercarriage under the wing adds to the designer's troubles. The upward shocks when landing apply concentrated loads to the spar, which try to sever it at the points of attachment of the undercarriage (called for this reason shear loads), and a bending force around the wing root. This means reinforcing the wing locally, and this at a point where large openings have been cut in the lower skin to take the retracted undercarriage. In the Spitfire the strong D torsion box was reinforced to take these loads and the legs retracted slightly backward into the non-load carrying part of the wing. In the Hurricane the two spars of the centre section took the landing loads, the wheels lying in the space between them. When there are engine nacelles on the wing the designer is faced with another set of concentrated loads, although the rear part of the nacelle makes a convenient resting-place for the retracted undercarriage without cutting holes in the wing skin.
Such then, is a very bare outline of the problems faced by the engineer who is designing a stressed-skin monoplane. In the subsequent chapters it is hoped to show how the changes are rung on the essentials in order to meet the different needs of the operators.
Before leaving the general subject of the new structure, I should like to make clear one or two even more general points on design. The mathematics of aerodynamics and structures are, as I have said, very complicated and very advanced. This would lead one to think that the science of aeronautics is an exact science. Yet such a supposition would be very far indeed from the truth. Despite the hundreds, or thousands, of millions of pounds spent in research every year, the designing of an aeroplane today remains largely a matter of guesswork. One has only to think for a moment of the number of breakages of new aeroplanes (since the last War alone), the extraordinary differences between two aeroplanes designed for the same work, the number of types abandoned because they are unsatisfactory, to realize the truth.
It may seem fantastic that with all the mental effort that has been expended, not to mention experiments in wind tunnels and in flight, tests on materials and on structures, that it is not yet possible to design an aeroplane precisely and with absolute certainty that it will do exactly what is required of it. It is true that for simple duties this can be done, but even the simple aeroplane seldom goes into service without some important modification.
One reason for the uncertainty of the aeronautical science is that it is always advancing; rarely is a designer called upon to perfect his work, he is usually asked to go one—or even two— better! Speed, climb, range, payload, invariably one or more of these must be increased. And always round the designer's neck is the millstone of lightness. If there were no weight penalty an aeroplane could easily be made strong enough for anything, but every new part has to be designed so that there is as near as possible no surplus weight. In fact, today, it is common practice to calculate the sizes of a part so that, when put in a testing machine, it will fail about 5 per cent below the design figure, after which it is strengthened until it meets requirements. The reasoning behind this is that if you make a part stronger than necessary there is no way of knowing for certain which places can be safely weakened, whereas it is comparatively easy to strengthen a weak point that has collapsed. Perhaps an outline of the designing of an aeroplane would be interesting? It certainly helps one to understand some of the results.
First, the customer—air force or civil operator—issues a specification, which says broadly what is wanted: the type of aeroplane, its duties, what it must carry, its maximum, cruising and landing speeds, climb performance, ceiling, range, and so on. Depending on the duties of the aeroplane the limits of its capabilities will be emphasized differently, e.g. in a transport range and payload will come first, in a fighter it may be speed or perhaps climb and manoeuvrability.
The designer and his team study the specification and set about meeting it rather on these lines. First they make a rough estimate of the wing area and the power needed to do the job. From this can be decided the approximate overall size and the engines and fuel. Next come some rough layouts to see how the parts fit together, where the guns, bombs or passengers will go. Almost certainly, the designer will have some of his pet ideas to incorporate—a particular type of construction, multi-spar wing, or torsion box, double-slotted flaps, even a "signature" tail shape—which will give the design his individual stamp. Approximations of weights, usually based on past experience, follow and from these more exact performance calculations can be made. Then preliminary stressings of the main components are carried out and from these the structure weights can be checked. It is a sort of house that Jack built, each stage depending on the previous one, with frequent back checking. Finally, all the work is embodied in a set of drawings, performance calculations, sketches of installations, descriptions of the finished aeroplane, price and delivery estimates. These are usually presented as a brochure and, quite often, a model is prepared to impress the less technical minds on the customer's staff.
All the foregoing work will have taken perhaps three months or more and the brochure will be laid before the customer, who will study it together with tenders from other manufacturers. This study may take six months—at the end of which time the customer, particularly if he is a Service one, will probably have changed his mind about the load, or the range, or the speed! Possibly, when all the interested departments were consulted about the specification someone was overlooked; or a new and extremely efficient piece of equipment is expected to be ready in time for the aeroplane going into service; or some trouble has arisen with a similar aeroplane. Anyway, the order is given to a manufacturer for a prototype and the real work begins.
Now the designing becomes too involved to be handled by the designer and his small nucleus of project engineers and it has to be given out to the main design departments, drawing office, aerodynamics and stressing. In this country perhaps fifty men will now be on the job, in the U.S.A. probably several hundred. Models are made for testing in the wind tunnel and from there the finer details of nacelle shapes, wing root fairings, and so on are carefully sorted out and the drag and lift of the aeroplane, as well as the effect of control movements and the forces on them, are measured. It may be that a flying scale model, with a pilot, is built if there are any very special features to check—such as the revolutionary Avro Vulcan delta bomber, which was preceded by the three Avro 707 one-third-scale models.
General design of the structure is carried out and estimates of the strength are made, from which the actual sizes and shapes of details are evolved. Strength calculations are checked by airworthiness authorities. Sometimes models of parts of the structure are made to check theories, in any case a fuselage and wing are built specially for strength tests to destruction. In addition to the design of the main structure, there is all the design work on cockpits, cabin and installations, for which a full-size wooden mock-up is built and fitted with all equipment . This work is usually the subject of many meetings with experts from outside the firm. Usually, during the design of the aeroplane the customer finds his original estimate of what he wanted was too modest. Almost invariably the all-up weight grows and the designer finds he needs more engine power to keep up the performance, which, of course, means more fuel, and this in turn leads to more weight, requiring more power. . . . This has always been the tendency in aeroplane design, but today it has become a chronic complaint!
All the foregoing has been connected with the prototype. It may be that one or two are ordered and carefully test flown and evaluated against competitive designs, after which an order is placed. If this is done, the original design, which was hand-made, will have to be "productionized", that is the structure will have to be modified for easy manufacture in quantity and jigs and tools will have to be designed. It is this transition from individual to quantity manufacture that causes the delays of three or four years usual between the appearance of prototype and production aeroplanes.
Sometimes, as with the Short Empire Boats, a bold step is taken and an aeroplane is ordered "off the drawing-board". In this case it is usual to hurry one or two aeroplanes ahead of the rest, so that all the flight testing can be carried out on them before the true production aeroplanes start to come off the line, thereby making it possible to incorporate any necessary modifications during manufacture.