With the close relationship which exists between the airline industry and aerospace, it is appropriate to begin an aerospace technology symposium with a short review of what the airline industry has achieved and where it is now going. The timing is similarly appropriate because it is the 50th anniversary of the true beginnings of the airline industry. Boeing introduced its B247 into service at the end of 1933, and Douglas followed with its DC2 in 1934. Both aircraft were major steps from previous technology, and created the equipment base from which air transportation could develop. At the same time, the Air Mail Act of 1934 created the organisational environment in which the industry could flourish and, inter alia, compelled Boeing and United Aircraft to divest themselves of airline interests and marked the true beginning of United Air Lines. TWA and American had preceded it by a few months, while General Motors were persuaded to sell off Eastern Air Linesonly a few weeks later.

New wind tunnels appear fairly infrequently and are usually of individual design to meet specific requirements and constraints. Detailed information on existing tunnels can be of great benefit to the designer, enabling him to avoid expensive pitfalls and to capitalise on the successful features that have evolved over the years. For this reason, and to advance wind tunnel design generally, new designs should be reported. The following account is presented in this light, even though the low speed tunnel described is unremarkable, being of conventional design and modest proportions. Indeed, its design follows closely the guide-lines laid down for such tunnels by Bradshaw and Pankhurst, guide-lines which themselves draw on the cumulative experience gained from previous low speed wind tunnels. The satisfactory performance of the tunnel is in no small measure the result of this approach.

A duct design algorithm is presented by which duct wall shapes are produced directly from prescribed wall friction velocity distributions. The method is based on matched axisymmetric core flow and integral boundary layer solutions in which the flow is assumed to be in viscid in the core, turbulent in the boundary layer and incompressible throughout. An inverse boundary layer analysis is based on Coles profiles coupled with momentum and entrainment equations. Wall friction velocity is used as the major boundary condition which leads to the calculation of the required core velocity at the boundary layer edge. This velocity distribution is simultaneously employed as the major boundary condition in an inverse core flow calculation which is based on an inverted form of the Stokes-Beltrami equation. The displacement surface method is used to match boundary layer and core solutions which together lead to a predicted duct shape. Numerical results are presented, some of which include separated boundary layer flows. An assessment of the influence of wall friction velocity distribution on duct shape and wall pressure field is included.

A method for determining stress intensity factors for cracked plates is proposed. The method uses a generalisation of Westergaard's stress functions to describe the stress intensity at the crack. Eigenfunctions of angular regions are used to satisfy the boundary conditions.

An analysis of results from reports on droplet evaporation indicated a marked difference in behaviour depending upon the suspension technique used. Few results were obtained with freely falling droplets, the condition which is of interest in aerial spraying operations. Test results obtained at Cranfield by dropping droplets down a climatic tower showed two unexpected results.

The diameter of water droplets reduced linearly with time until they reached a size of about 150 micrometers and then the rate of reduction in diameter increased by about 27%. It is shown that this change occurs when the fall Reynolds number of the droplet becomes less than about 4.

Above this Reynolds number the airflow is separated from the base of the droplet and the trapped air in the toroidal vortex becomes fully saturated and no evaporation occurs over the base region. Below a Reynolds number of 4 the flow is attached everywhere and evaporation occurs from the whole surface area.

When small amounts of molasses are added to a spray formulation the droplets evaporated initially in a way similar to water droplets. As a result of falling, a toroidal vortex motion is induced within the droplet. This causes the same small fraction of the volume of the droplet to cycle close to the surface.

Hence, this fraction of the volume looses all its water, becomes far more viscous as a result and the consequence is the toroidal motion almost stops. In effect the molasses forms a skin over the droplet surface inhibiting further evaporation of the remaining volume of water.

A technique for the prediction of buffeting response in flight from wind-tunnel tests on models of conventional construction is described and assessed. Results are presented from tests on models of a Gnat T Mk 2 trainer aircraft, and predictions are compared with flight measurements of buffeting made during high incidence manoeuvres. Flight/tunnel comparisons of buffeting response measurements on the TACT F-111 and SAAB 105 aircraft are also discussed. In general, good agreement is demonstrated between measured and predicted responses. Some remarks are also made on the determination of damping ratios from accelerometer or strain gauge readings recorded under buffeting conditions.

A sequence of high Reynolds number two-dimensional entry flow problems is considered by varying the form of the entry profile from one representing uniform shear to one representing different parts of a Poiseuille profile. The detailed development of the entry flow is illustrated in a few cases. The entry length, pressure drop, and pressure loss in the entry region are given in each case.

The ‘sliding block’ asymmetric nozzle, which is mechanically simple and robust, is ideal for a small, variable Mach number supersonic wind tunnel installation, due the fact that it is affordable and relatively easy to manufacture. Until now there has been no rational procedure for selecting the correct nozzle profiles to satisfy certain predetermined nozzle performance criteria, ie, a range of operating Mach numbers and an acceptable test section flow non-uniformity over the given range. The present paper considers the minimum requirements which are sufficient and necessary to design an asymmetric nozzle and shows that these can be met by nozzles defined in terms of a shape parameter, an angular parameter and a length parameter. The lower end of the range of operating Mach numbers is determined by the choice of the overall flow turning angle. For a family of nozzles defined by a particular shape parameter, an heuristic analysis indicates how the test section flow non-uniformity must vary as a function of a variable which is a particular grouping of the angular and length parameters, and the range of operating Mach numbers. The dependence of the nozzle performance on the parameters is investigated numerically, assuming inviscid flow and utilising the method of characteristics. The result shows that the best nozzle shape is where the flow is expanded very rapidly immediately downstream of the throat. For a particular family of nozzles, the numerical analysis shows the complex relationship between nozzle performance and the angular and length parameters, which is best described in a graphical representation. As a result, the task of selecting correct profiles for sliding block nozzles is reduced to simply selecting the appropriate values of nozzle parameters from the curves supplied.

Transverse curvature effects in axi-symmetric flow on a circulai cylinder are discussed and it is deduced that for small δ/a, where δ is the boundary layer thickness and a is the cylinder radius, transverse curvature effects are small, for δ/a = 0(1) the effect is felt mainly in the outer region and for δ/a » 1, both the inner and the outer regions are affected. It is argued that the last case can only be achieved experimentally in the laboratory if the radius Reynolds number Ra= U1a/v is small and the streamwise Reynolds number Rx=U1xlv is large, implying x/a large. New experimental data are presented for this case. The data show that both the mean and turbulent flow field scale on outer variables throughout the layer and, for Ra≥455, exhibit Reynolds number independence. No logarithmic region is found in the mean profiles, a result which is echoed by the absence of a shoulder in the streamwise turbulence profiles which clearly evidence that the flow in the layer is fully turbulent. A flow model is suggested, in physical terms, to explain the very high turbulence levels found close to the cylinder at values of Ra which should, according to stability theory, be bordering on laminar flow.