As the treatment of the different subjects must be brief and general this paper can only deal with the complex subject of aircraft fatigue in very broad terms, attempting to trace the development of the subject over the past ten years or so, to present some of the problems and the methods being adopted to solve them and to emphasise the very wide extent to which fatigue affects aircraft design today.

The fatigue of metals was recognised as a separate form of failure as long ago as 1854 and is now the subject of a very extensive literature. Mechanical engineers have long been familiar with fatigue as a criterion of design and are apt to be critical and even amused at the aeronautical engineer's recently found interest.

In a recent study of the buckling of pin-ended struts of a certain light-alloy material the author found it necessary to use ball-end fittings which supported rigidly an appreciable length of the strut. For some of the struts—having slenderness ratios of the order of 20—the unsupported length was only about a quarter of the overall length between pinned-ends, (Fig. 1). The effect of rigid end-fittings can easily be taken into account in calculating the elastic buckling load of this system. The author expected to find this end effect important for shorter struts; however, this is not the case. In fact, buckling of the strut is governed largely by the flexibility of the unsupported length, and even large end fittings may not affect appreciably the buckling load.

A review is given of developments in the field of aeroelasticity during the past ten years. The effect of steadily increasing Mach number has been two-fold: on the one hand the aerodynamic derivatives have changed, and in some cases brought new problems, and on the other hand the design for higher Mach numbers has led to thinner aerofoils and more slender fuselages for which the required stiffness is more difficult to provide. Both these aspects are discussed, and various methods of attack on the problems are considered. The relative merits of stiffness, damping and massbalance for the prevention of control surface flutter are discussed. A brief mention is made of the recent problems of damage from jet efflux and of the possible aeroelastic effects of kinetic heating.

The flow separation on a swept-back wing with 44 degrees leading edge sweep at 18 degrees incidence has been investigated by means of detailed pressure distribution measurements over the leading edge, boundary layer flow determination with liquid film technique, and yawmeter traverses. A wool-tuft grid was also used, and a spin detector was developed to search for regions of vorticity. These tests established that the flow separates on the leading edge; over the inboard sections it re-attaches behind a “ short” separation bubble, while outboard it only re-attaches near the trailing edge, thus forming a “ long ” separation bubble, or else fails to attach. The separated flow in what has been commonly called the tip stall does in fact take the form of a “ ram's horn “ vortex with the origin, or “ tip,” located at the junction of the two bubbles on the leading edge. The vortex lies outwards across the wing surface at approximately 20 to 25 degrees to the line-of-flight before curving aft to be shed into the wake, where it extends almost from mid semi-span to the wing tip. This vortex induces considerable changes in flow direction, both on and over the wing, and also in the wake. Thus in the wake a maximum downwash of 23 degrees is induced aft of the mid semi-span, and there is an upwash of 17 degrees at the outer edge of the vortex, almost aft of the tip. A good correlation between yawmeter results and the boundary layer flow direction indications from the liquid film technique was obtained.

At this stage in the Symposium three papers have been presented on the subject of Supersonic Flight. They will have left untouched a field which is much too wide and extensive to cover in a single paper. In fact, any attempt to cover the major problems remaining would result in a superficial and valueless effort. However, the general problem of control, particularly that resulting from the elimination of the pilot, is clearly the next major consideration and this is basically the theme of this paper.

Other major problems may be described as those associated with Electrical and Mechanical Engineering; they include power supplies, services and, in particular, refrigeration. If one looks back some twenty years these problems were not considered of any great importance, although at some stage every designer must have decried their reliability. In fact “ electrics again ” seems to have been heard far too frequently when trouble occurred.

This investigation is concerned with the stresses in a hollow circular cylinder of constant wall thickness, and closed by flexible plates, also of constant thickness. The cylinder is subjected to the combined action of an internal pressure and an axial force, both of which are statically applied. Although the particular case envisaged is that of a shock absorber, there are other engineering elements to which the analysis is applicable. Any closed circular cylinder of constant wall thickness subjected to an internal pressure is within the category. In a paper, Dr. D. M. A. Leggett investigated the case of an oleopneumatic shock absorber with rigid end plates, and what follows is an extension of his paper. As would be expected, the flexibility of the end plates causes a reduction in the flexural stress.

The linearised theory value of the wave drag D at zero lift of a smooth slender body of arbitrary cross-sectional shape was shown by Ward to be given by

where S (x), 0 ⩽ x ⩽ 1, is the cross-sectional area distribution of the body and q is the kinetic pressure. The development of this result has aroused interest in two problems: the derivation of the optimum area distribution for minimum wave drag under certain specified conditions and the numerical evaluation of the wave drag of a specified area distribution. These apparently distinct problems have hitherto been treated separately, but it is shown here how an attempt to solve the first problem has led to a practical method of solving the second.

A solution, in terms of known integrals, is obtained for the motion from rest of a machine, idealised as an undamped linear mass-spring system, when subjected to an exciting force whose frequency varies at a constant rate.

In many installations of modern high speed machinery the running speed of the machine is in excess of the resonant or natural frequency of the system, and consequently starting up or stopping the machine could result in vibrations of large amplitude. The problem of assessing the magnitude and duration of these vibrations is very complicated and has been solved analytically only for the case of a single degree of freedom system excited by an oscillating force whose frequency varies linearly with time. However, even this solution is not easy to evaluate, the integrals involved demanding either graphical construction and numerical integration or summation of series.

The twist that arises from bending stresses in straight cantilevers of thin-walled section asymmetrical about any vertical plane, is well known and eventually leads to the concept of a shear centre. If an analysis is made along the same lines as that used for investigating straight beams, of cantilevers curved in plan, it is found that the bending moments transmitted are again responsible for shear stresses in the flanges of the beam and cause twisting.

The following analyses refer, principally, to cases in which the cantilever carries a concentrated load at its end and are confined to the relatively simple forms of the channel and I-section. Each cantilever is perfectly built-in at one end and, for simplicity, it is considered that the web of a section offers no resistance to bending and that the beams are of constant curvature in plan.

The simplest case of asymmetric heating of a fluid flowing between parallel plates is considered. This occurs when heat is transferred to the fluid through one plate, the other being adiathermal. For a theoretical analysis, it is convenient to assume fully developed flow in a parallel channel of infinite aspect ratio, this condition being not widely different to that met with in practical applications.

The problems associated with the fatigue life of aeroplane structures have been so thoroughly discussed by experts of international repute that it is with some trepidation that the present speaker ventures into this difficult and controversial field. However, the need for more experimental fatigue data on actual structural components (as distinct from idealised test specimens) is undeniable; further, it is hoped that the present paper may stimulate discussion. Opinions expressed in this paper are the author's personal views, and do not necessarily reflect the policy of the Ministry of Supply.

A comprehensive bibliography on this subject has been given recently by Lundberg in the 18th Wright Brothers’ Lecture.

The part played by strength testing in establishing structural integrity and efficiency is well known to aeronautical engineers. In principle, the great advances made in structural theory and stressing methods should lessen the pressure on strength testing facilities. In practice, things do not seem to work out this way. For this there appear to be two reasons. The first is the continuous change in aircraft structural design to meet new or more exacting operational roles. The second is the demand for still lighter structures without prejudice to safety. Strength testing, therefore, is as necessary as ever it was if the best results are to be obtained.

For a great many years structural strength testing meant just static strength testing. This is no longer true, since for many types of aircraft fatigue testing has become at least as important as static testing.

A procedure was discussed in Ref. 1 for dealing with structural discontinuities by making the strain energy of the structure a minimum, subject to the conditions that the members to be cut out are entirely free from load, and the proof was based on the well-known method of Lagrangian Multipliers. Here is another proof.

A new proof is given of a method whereby records of the transient vibrations of linear systems can be analysed to give the damping factor and frequency of each mode of vibration associated with the motion. The method is applicable to acceleration, velocity or displacement records, and as the amplitude of each mode can be found, by using a suitable number of pickups, the normal modes of continuous systems may also be determined. Obvious applications lie in the analysis of flight flutter test data, and of the transient responses of servomechanisms.

The object of this paper is to present an analytical method of investigating the flexure of a non-uniform beam under transverse loading. A method due to Strandhagen for a uniform beam is extended to the case of a non-uniform beam, the deflection appearing in the form of a Fourier series, the coefficients of which are functions of the loading, the end-conditions, and parameters which define the non-uniform flexural rigidity of the beam.

More than two years ago Pollard in a paper given to the Society made out the case for the use of fibre-reinforced plastics in the structural field. He and others have shown that materials based on strong, inert fibres, such as glass or asbestos, have mechanical properties which, when stabilised by suitable resins, permit them to function with a high degree of efficiency within certain well defined parameters of loading.

Four distinct groups of resin are used in combination with such fibres, phenolic, polyester, epoxide and silicone. With the present exception of silicone resins, each can be used at low or even contact pressures during moulding and thus do not impose serious processing limitation upon the size, quantity or complexity of the structures which can be contemplated.

Strong plastics consist of two-phase arrangements of pairs of brittle solids which in combination may be very tough. Improvement now depends upon an understanding of brittle solid theory. The strength of brittle solids is generally very low but may occasionally approach the very high theoretical values. This is due to the high inherent strength being generally marred by tiny defects. When these defects are removed or contained high strength may be achieved. The various mechanisms by which this may be done are exemplified in mica, common salt, glass, metal whiskers and so on. The relevance of these substances and mechanisms to the manufacture of plastics is discussed.