Present day structural forms and methods of fabrication have considerably increased the importance of wind as a design consideration. For estimations of the overall stability of a structure and of the local pressure distribution on the cladding, a knowledge of the maximum steady or time-averaged wind loads is usually sufficient. Mind tunnel tests to determine the wind loading coefficients are often made in smooth uniform flow, but for more accurate data account must be taken of the effects of the vertical gradient of wind speed and the turbulence of natural winds. Further research is needed into these effects and also into methods of obtaining a sufficient representation of the natural wind in the wind tunnel.

There are a number of ways by which wind excites structures into oscillation; among these are vortex excitation, galloping, proximity effects including buffeting, stalling flutter and classical flutter. The vortex and galloping excitation might be expected to be especially sensitive to the turbulence properties of the air flow. Also, in the absence of any mechanism for instability, atmospheric turbulence may directly excite oscillations through the random forcing by the pressure fluctuations which it produces. Further understanding of this problem must come through research into the effects of turbulence (and to the extent to which these effects may be disregarded), but the range of the conditions is so vast and complicated that it seems unlikely that sufficient aerodynamic and wind data will be accumulated to permit the response of a proposed structure to be calculated with reasonable certainty, and for major projects it is anticipated that comprehensive tests on aeroelastic models in wind tunnels with appropriate turbulent air flow will continue to offer the more reliable predictions.

The air flow around buildings is of concern inasmuch as it influences the dispersal of combustion and other gases from the smokestack and also in its effect on the speeds and turbulence of the wind over areas used by pedestrians. The erection of a tall building may cause an increase in wind speeds and gustiness at ground level. These problems of the external flow over buildings are readily examined in wind tunnels. For this purpose tunnels with large working sections are desirable to permit a sufficiently wide area of the neighbourhood to be represented.

Since some aspect of fluid dynamics is involved in the design or operation of many items of plant in the electrical supply industry, this survey is confined to those items which are peculiar to the industry and whose performances are directly related to their aerodynamic behaviour. Three main problems will be discussed namely, the internal flows in boilers and natural draught cooling towers and the aerodynamic behaviour of overhead lines. In addition, the aerodynamic aspects of current and recently initiated field experiments will be considered.

This paper defines Internal Aerodynamics and via similarity generalises the concept to include internal flows of fluids other than air.

Variable area ducted flows are considered distinguishing one, two and three-dimensional treatments and the difference between accelerating and diffusing flows is stressed. The design of variable area ducts using ideal inviscid fluid theory, ideal fluid plus boundary layer theory and inviscid but rotational fluid theory is discussed. The success and limitations of these theories with reference to contractions and diffusers is surveyed The particular difficulty of a diffusing flow as affected by entry velocity profile and turbulence is considered.

From the problem of simulating entry velocity profiles arises the need for the generation of required velocity profiles in duct systems. Examples of methods of generation and difficulties are discussed.

Next follows a series of examples of internal flow problems.

The radial flow diffuser is considered as an example of the interaction of viscous and inertia effects. An approximate theoretical treatment is given and the real flow, with the occurrence of reverse transition, is described.

The deliberate production of a reversed flow regime for flame stabilisation is described with reference to a marine boiler register Aspects of pressure drop and stability are considered.

Secondary flow is defined and the origin in the distortion of the upstream vorticity field is described. The case of a strut wall junction is given with details of the downstream flow where the secondary flow produces apparent multiple wakes from a single strut.

Air flow excited vibration of heat exchanger tubes is described. Different mechanisms resulting in transverse or streamwise oscillations and the coupled modes of tube rows are discussed.

Finally the paper emphasises the large available body of information in internal aerodynamics, largely of aeronautical origin, and considers its usefulness and limitations for the problems discussed.

I am going to restrict myself to consideration of those problems which arise from the natural wind, and I am going to begin—and perhaps also to continue—with what may be, to paraphrase one of our Chairman’s expressions, some offensively simple . considerations. I do this because, in our studies of industrial aerodynamics, we are increasingly applying the sophisticated techniques developed from years of aircraft aerodynamics; and there is always a danger that, unless we go back to fundamentals, we may invoke a theorem or technique from aircraft practice which is inappropriate for an industrial problem. First of all, therefore, I shall go back to square one.

In a general survey of the Guided Weapon scene over the past twenty years, in the space available, it is not possible to be comprehensive, nor may it be desirable. It is necessary, however, in order to set the background, to go back to early work before the period under review.

Aviation in its early days was as beset as any new science with the ravages of that elusive creature the creeping error. They are rare today but a particularly daring one has crept into the Centenary review of the Fleet Air Arm (January, p. 208). We did not have catapults to launch our Sopwith Pups, Camels or Strutters from the turrets of battleships during the 1914-18 War. Catapults came later with the amphibians designed for service with cruisers.

I have always been an enthusiast for engines, even before 1913 when after a flight with Geoffrey de Havilland in a BE2a, I took my ticket in a Caudron biplane at Hendon, enjoying a little thrust, and much castor oil from the “Y” Anzani engine. When I completed my figure of eight, nose well down on the turns, the only engine control—the switch—failed, so I foreshadowed my technical future by kicking off the petrol tap! Before that I Was acquainted with engines in the motor trade. I was mechanic in the Maudslay car in the first Tourist Trophy Race of 1905; our engine ran well till it became a bit short of petrol and ate its butterfly throttle. But, in 1907, I drove their first live axle car in the Scottish Reliability Trials, obtaining a gold medal for a non-stop run.

This note outlines a method for the prediction of drag due to lift of plane wings at Mach numbers below drag divergence and Reynolds numbers above 106. The method is based on the correlation of a number of wind tunnel measurements in terms of the effect of viscosity on lift curve slope. A comparison is made of the accuracy of estimating the induced drag factor, k, using this method, with the method of ret. 1, and it is shown that considerable improvement has been made, and that, in general, the predicted value of k is within 10% of experiment.

A further element for the Matrix Displacement method discussed in the author’s Main Lecture was a parallelogram or arbitrary anisotropy under plane stress. As for the other elements analysed in the preceding technical notes of this series, we assign to the parallelogram a linear variation of the strain along the edges. This is a more general specification of the strain pattern than for the parallelogram in ref. 2. Consequently, it is necessary to introduce additional nodal points in order to account for the increased number of kinematic freedoms. In the present case eight nodal points are provided as against four in the original model. The procedure is best understood in