The difference in outlook between metallurgist, physicist and designer on the problem of fatigue of metals is due not only to the differences in training and method but also in immediate objective, however closely their ultimate objectives may coincide.

Physicists consider fatigue in terms of crystal structure and composition of the constituents of the crystal, noting in passing that certain types of crystal structures are associated with poor fatigue strength. The metallurgist's prime interest lies in the effect of surface finish, heat treatment, the physical condition of the surface, such as degree of cold work, the effect of carburised and nitrided cases having different hardness from the core. Both think in terms of controlled condition tests, with idealised test conditions such as cylindrical test specimens, close control of changes of section, polished surfaces of only a few micro-inches surface roughness, operating in a controlled atmosphere.

In an interesting Technical Note in the January 1953 Journal, Messrs. Dickinson and Hadley reported some results from tests made to check a proposed extension to the Royal Aeronautical Society Data Sheets on stresses and deflections in flat panels subjected to normal pressure. The test rig was originally too weak and did not provide the edge conditions which were assumed when calculating the curves of Data Sheet 02.09.02. This explains the discrepancy with regard to the deflections, but it is remarkable that the measured membrane stresses f3 were so low. It is a well-known fact that the membrane tension at the centre of a square plate with large deflections is not influenced very much by different edge conditions. This can be seen in Data Sheet 02.09.05 where the two extremes, free edges and edges held apart, are represented for a square plate with edges free to rotate. The membrane tension at the edge is, of course, much more sensitive for an edge displacement, but we suppose that f3, is the membrane stress at the centre of the plate (as in Data Sheet 02.09.05).

There have been many proposed, but few practical, methods of refuelling in flight since the first early experiments in 1923, and until some three years ago the only tried and accepted method was the “ looped hose ” system developed since 1934 by Flight Refuelling Limited, and still in use (with British equipment) by the United States Air Force.

This system consisted, briefly, in carrying a hose unit with a capacity of some 230 ft. of hose in the tanker aircraft, and a hauling line and windlass in the receiver aircraft. Operators were employed for both tanker and receiver aircraft.

To make contact between the aircraft, the receiver operator first trailed his hauling line, with a “ sinker ” weight on the end of it, so that the line curved downwards. The tanker aircraft then flew alongside and a projectile, attached to a “contact line,” was fired from a line–throwing gun, so as to pass the contact line under the hauling line.

Consider a cantilever beam of uniform cross section whose generators are parallel to the z-axis and whose lateral surface is free from surface tractions. The line of centroids of the cross sections in the unstrained state is taken as the z-axis, and the x- and y-axes are the principal axes of the cross section at the centroid of the fixed end z = 0.

The other end of the beam (z = l) is subject to forces which reduce to a single force with components (Wx, Wv, 0), transverse to the z-axis, acting through the load point L of this end section (see Fig. 1). The co-ordinates of L are taken as (p, q, l).

This paper is primarily concerned with structure weight as related to aircraft shape and size, rather than with the saving in structure weight that may be brought about by improved methods of construction. The difference in weight due to differences in methods of construction is a question of structural efficiency, and this will not be discussed here. Structure weight, even when expressed as an overall percentage, is not in itself a measure of structural efficiency. If the depth of an aircraft wing were doubled, without altering the lift distribution on the wing, the bending loads would be unaltered so that the resultant end loads in the top and bottom surfaces would be halved. If the second wing had the same weight as the first the structural efficiency would be less. Conversely, if they both had the same structural efficiency the deeper wing would have the lower structure weight

It has been said that a rocket motor is the simplest “ prime mover.” This is correct in theory, since fundamentally a rocket consists of a chamber containing a gas under pressure, having a nozzle at one end through which the gas expands, forming a propulsive jet. Since the pressure ratio of the rocket is always greater than the critical pressure ratio, the nozzle is invariably of the convergent-divergent or de Laval type in which the velocity is subsonic up to the throat, sonic at the throat and supersonic in the divergent portion.

It is a popular fallacy that a rocket or jet engine obtains a forward thrust by allowing the rearward jet of gas to push against the atmosphere in which it is operating. This is not correct and it can be proved that a rocket will operate successfully, and more efficiently, in a vacuum.

The problem of the stress analysis of circular fuselage frames has been investigated by a number of authors; the analyses are, necessarily, rather mathematical in nature although in some cases the final results have been presented in extremely practical form. The present paper offers, in non-mathematical language, a short account of the fundamentals of the problem and a brief guide to the more important published literature. It also develops an approximate method for the analysis of frames using Lagrange's method for minimising a function of several inter-dependent variables. This method is shown to be particularly suitable for the analysis of frames with large cut-outs, about which little has previously been published. In addition, the underlying parameters which determine the stress distribution are deduced from the theory, and the deductions are compared with previously published work. In general, agreement is good. The paper is written throughout with a view to being of practical use in the actual stressing of frames. Numerical examples and explanations of mathematical methods are included in Appendices.

This paper describes the construction of a simple electronic simulator for the solution of flutter problems in two degrees of freedom. It was intended as a prototype for a much larger machine to solve problems in six degrees of freedom.

Details of the construction and circuits are given, together with some typical solutions obtained on the machine. As a result of the successful operation of this prototype a larger machine in six degrees of freedom has now been built.

In A. H. Stratford's interesting note in the August Journal, he assumes that an airline is presented with an aircraft of given performance, and must decide what is the best range for it to fly. Surely the practical case is the exact opposite: that an operator wishes to carry passengers or freight between two given points, and must choose between perhaps three or four available types of aircraft. Moreover, it is up to the operator to tell the manufacturers which feature he wants improved (e.g. fuel consumption, take-off distance, and so on), when he is ready to order replacements.

Secondly, Mr. Stratford wishes to replace the usual concept of direct operating cost per payload ton mile by that of per cent, profit, as a measure of the efficiency of operation over a certain route. While this may be a more realistic approach to the problem of a private airline company, operating over a single route, its merit is not so obvious for the large national or subsidised airlines, which are often required to fly unprofitable routes.

The carpet and lattice methods of plotting the relationship between three and four variables respectively, developed by R. F. Sargent, of the National Physical Laboratory, have been described previously by A. H. Yates, who gave examples related to the presentation of aerodynamic and engine performance data. Outside these fields there appears to be still some lack of appreciation of the usefulness of this form of graphical representation, although it is capable of wide application. The writer has found both carpet and lattice plotting of great value in connection with the preparation of structures data sheets, and is convinced that interpolation is generally rendered easier and more accurate if graphical data are presented in this way. In one particular instance it was found possible to develop a nomogram in the form of an “ extended lattice ” plot, which gave in graphical form a direct solution to a set of equations which could be solved otherwise only by continuous approximation.

The problems discussed in this note refer to the simplest case of the use of linear accelerometers, in which the flight path is in the vertical plane of symmetry of the aeroplane, and the acceleration normal to the path is to be determined.

Generally two facts are neglected:–

1. (i) That the accelerometer senses along an arbitrary body axis ZB (Fig. 1), instead of along the wind axis Z (normal to the path).

2. (ii) That the accelerometer is not placed exactly at the e.g. of the aeroplane but at some point A, having co-ordinates xA and zA in the body axes.

The problem of fatigue in aircraft design is a comparatively new one so far as the airframe designer is concerned. There are three main reasons for its prominence. Firstly, the utilisation rate has increased enormously, particularly with transport types; indeed, the hours flown by a civil aeroplane in a single year nowadays far exceed those in the whole life of a bomber up to the period of the 1939-45 War. Secondly, the increased cruising speeds of modern aircraft have aggravated the loading conditions which give rise to fatigue; and thirdly, the development of modern high strength light alloys has enabled higher ultimate stresses to be used, although the fatigue resistance of the material has not kept pace with its ultimate strength.

In Mr. J. M. Stephenson's Technical Note, “ The Elimination of Wall Effects in Axial-Flow Compressor Stages,” in the April 1953 issue of the Journal, the author suggests that the blade rows of an axial flow compressor are so closely spaced as to ensure that the axial velocity profile is unchanged across the rows. Whether this statement is correct or not such an assumption regarding the axial velocity profile is a basic design condition and when made it will not leave any flexibility in the choice of the function f(r).

The problem of aileron effectiveness when the elastic properties of the wing structure are taken into account has already been adequately covered in the subsonic regime by A. R. Collar and E. G. Broadbent. The method used depends upon strip theory which pre-supposes a knowledge of the relevant aerodynamic derivatives. A unified method has been considered by the present author which employs either the semi-empirical theory of D. Küchemann or the lifting surface theory of H. Multhopp in such a manner that the effective values of the relevant derivatives are found, together with the rolling efficiency.

There has been much discussion of late on the nature of the noise from aircraft travelling with velocities greater than that of sound. It is possible that these phenomena can largely be explained in terms of the Doppler effect. For a source of sound travelling with a velocity greater than that of sound relative to the observer, the Doppler effect takes a very interesting form. It is discussed mathematically in Ref. 1.

There is at present no thoroughly satisfactory method of measuring accurately the true air speed of an aircraft, particularly at high speeds.

Ground radar can accurately follow a high-speed aircraft but, without the development described here, can only measure the speed of the aircraft relative to the ground. The method developed by the authors makes immediate allowance for wind speed and is capable of giving the true air speed of an aircraft to within one per cent. There is no apparent reason why the method should not be applicable even at very high speeds and there is no limit in aircraft altitude at which the measurement may be made.

The commonly accepted practice of resolving a velocity V at incidence α into components V sin Λ parallel to the leading edge and V cos Λ normal to the leading edge at incidence α sec Λ is only approximately true and the exact method of resolution is developed here.

In dealing with the aerodynamics of swept-back wings the free stream velocity may be resolved into a constant component parallel to the leading edge and a variable component normal to the leading edge. This is only possible if the constant percentage chord lines are all parallel to the leading edge, that is there is no taper, thickness taper or twist on the wing. This note gives the exact form and the small angle approximations of the velocity components and of the incidence of the component normal to the leading edge.

Frame analysis problems in general involve the determination of the values of three redundancies, which can become somewhat laborious. By applying the redundancies at the elastic centre of the frame the amount of work is much reduced and the solving of simultaneous equations is avoided. The method does not seem to be very widely used and it is considered that this may be due partly to the way in which the proof has usually been presented. The following proof is given in the hope that it may assist in clarifying the method.