Following Parts 1 and 2 of this experimental study the effects of rows of spherical roughness elements upon transition of the boundary layer are now described. Data presented includes that for the start of the end of transition, the effect of spacing of the roughness elements, the effective origin of the turbulent boundary layer and its development and the spanwise variations across the turbulent boundary layer.

It is probably not necessary for an economist to justify his temerity in addressing the Air Law Group of this Society on Licensing since I am here by invitation. But it is worth noting that a great deal of air law is, in fact, concerned with the economic regulation of the industry: the law is endeavouring to give effect to conclusions about how best to arrive at certain economic results.

I shall not discuss any details of the law relating to air transport licensing: I shall concern myself with the broad economic issues which underlie the law. I shall start from first principles and ask “Why should we regulate air transport?” and, if there are good reasons for so doing, “what are the essentials of a regulatory system?”.

Two recent papers by Bradshaw have drawn attention to instabilities in the flow through wind tunnel screens. In Plate 1 of ref. 2 a smoke tunnel photograph shows the manner in which the jets coalesce behind a high blockage, two-dimensional cascade. In the present note, three points relating to this flow will be discussed.

It has long been accepted within the aircraft industry and others, that the number of man-hours needed to produce successive units or batches of units decreases. This effect is called ‘learning’; and a plot of the man-hours required per unit set against the production number is called a ‘learning curve’. Similarly a plot of unit cost (at constant economic conditions) against production number is also called a learning curve, since cost is related to the number of man-hours required.

When the production costs of future aircraft are estimated, learning has to be simulated. This is generally done using curves chosen from a standard family of ‘learning curves’ generated by a simple parametric equation; a curve is chosen on the basis that it is the best-fit by the standard equation to the out-turn data from previous similar projects.

Maskell has developed a theory for the blockage effects of bluff bodies in a closed wind tunnel and has shown that it gives good results for flat plates normal to a stream. For streamline bodies there are well established methods of calculating the corrections, but for long bodies with blunt bases it is not obvious what method should be used.

Now that major space flight operations are under way in Russia and America there is a natural interest in such topics in this country among those working on aeronautics and guided missiles. At present one can only speculate on how this country will eventually contribute to space flight. As the Duke of Edinburgh has remarked, the difficulty is not due to lack of scientific talent but to the absence of surplus funds. Even so, we can make effective contributions in many ways and we should not lack boldness in seeking out possible ways of doing this. In the meantime, however, there seems to be a place for a review of the present technical situation and prospects for the near future, as seen from the standpoint of one in the British Aircraft Industry. As the majority of interplanetary flights will begin through the Earth’s atmosphere and many will finish by way of the same medium, the Royal Aeronautical Society can claim a special and legitimate interest in these matters. Ever since its inception the Society has encouraged the discussion of new technical advances, although its ability to do this has sometimes been hampered by the requirements of security. The same restrictions apply to this paper with the result that I have chosen to draw upon published material throughout, and to review the subject on a broad basis. I hope the work will not be less acceptable on these grounds.

In the past few years, two problems in applied aerodynamics have received much attention. The first is the problem of low drag in which it is desired to reduce considerably the drag of aerofoils over a certain range of lift coefficient by suitable design and construction. The second is the problem of increasing the lift in which it is desired principally to obtain a higher maximum lift coefficient.

It is Always rather risky to ask old people to talk about what things were like when they were young. Having turned on the tap you may not see how to turn it off again. Fortunately there is a circumstance which keeps this trouble under control. Old people remember a lot of things when there is no call for them to do so but seldom when there is. That is the reason why I have written down most of what I can say for this lecture.

I cannot attempt to give a complete or an objective picture of aeronautical science when I started thinking about it. The best I can do is to describe a few of the things which happened to me in those days to try to convey some picture of our primitive ideas.

In this paper, the results of computational studies on the unsteady flow features in three-dimensional empty cavities and cavities with a representative store are presented. Flow simulations with a turbulence model based on a hybrid method, which behaves as a standard Reynolds-averaged Navier-Stokes (RANS) model within the attached boundary layer and as a Large-Eddy Simulation LES sub-grid scale model in the rest of the flow (commonly known as Detached-Eddy Simulation (DES)) are used in this study. The time-mean flow study showed the presence of three-dimensional effects inside the cavities. The mean flowfield visualisation also clearly showed the presence of a pair of ‘tornado-like’ vortices in the upstream half of the cavity which merge to a single, large recirculation further downstream. Visualisation for the cavity-with-store case revealed that the mean flowfield was effectively divided into two halves with significant reduction of the spanwise flow across the cavity width. In the unsteady flow study, near-field acoustic spectra were computed for the empty cavity and cavity-with-store cases. Study of unsteady pressure spectra for the cavity-with-store case found the presence of many peaks and the corresponding mode frequencies were found to agree well with the Rossiter modes. The blockage effect of store and strut on the spanwise flow is thought to have reduced the interaction, and subsequent non-linear coupling, between the Rossiter modes. This may be the reason for the co-existence of multiple modes without the coupling among them.

Rotating-Drum (or mirror) cameras have often been used in conjunction with schlieren apparatus to investigate the performance of shock tubes. In the conventional arrangement the working section is viewed through a slit parallel to the axis of the tube, which is also parallel to the drum axis. The photographic record obtained on a film attached to the drum represents the distance-time (x, t) diagram of the propagation of (density) disturbances in the working section. The quantitative information obtainable from such records is usually limited to the speed of propagation of such disturbances.

The complicated time-dependent flow in the wake of a blunt based body at low speeds is a well known phenomenon, but drastic changes in the wake structure occur as speeds increase. Somewhere in the transonic regime, the periodic vortex shedding ceases, and at supersonic Mach numbers a steady flow pattern is established. The boundary layer on the body separates to form a free shear layer, the two layers from opposite sides of the body meeting at the so-called neck of the wake. Downstream of this point, the viscous wake persists for many hundreds of body diameters, and the term “far wake” is usually taken to refer to downstream distances greater than about ten body diameters. The neck of a hypersonic wake occurs typically one or two diameters from the base, and marks the downstream end of what is often known as the base flow region; this region includes a low speed, recirculating flow bounded by the shear layers and the base wall. Between the base and somewhere a few body diameters beyond the neck, the flow field is known as the near wake, and forms the subject of this survey.

This is not intended to be a treatise on the way to investigate air accidents. That story has been told many times and there is much authoritative literature on the subject already—besides I am thirteen years out of date. My aim is to tell something about the way it began and how, in little more than 30 years, it has become a really scientific art, always keeping step with the advance of aircraft and powerplant design and yet never gaining a sufficient lead to put a stop to accidents.

A new similarity parameter has been used for analyzing the unsteady aerodynamic behaviour of vehicles undergoing sinusoidal pitching motion. If this parameter is identical for two unsteady manoeuvres with different reduced frequencies and oscillation amplitudes, the corresponding hysteresis loops of the force and moment collapse on each other. To support and verify this, extensive unsteady wind tunnel tests have been conducted on a standard model, which is a well known fighter type configuration. The acquired data were used to train a certain type of neural network, called the Generalised Regression Neural Network (GRNN), to reduce the number of wind tunnel runs. The scheme, once proved to give the correct results for various conditions, was applied to extend the experimental data to other conditions that have not been tested in the tunnel. Both the predicted and acquired tunnel data were used to show the performance of the similarity parameter.

A methodology is described that determines the statistical confidence bounds on the results from flight flutter tests: modal parameter estimates, flutter margin values and flutter speed estimates, without the need for Monte-Carlo simulation. The approach is based on least squares statistics and eigenvalue perturbation theory applied to the various stages of the analysis process, starting with frequency and damping estimation through to the flutter margin calculations. The technique is demonstrated upon a number of data sets from aeroelastic simulations of flight flutter tests.

The prospect of reducing the induced drag of an aircraft by using both a canard and tailplane for trimming is investigated and results are compared with those for conventional tail-aft and canard arrangements. It is concluded that improvements of approximately 20% in lift-drag ratio are theoretically possible at high lift coefficients by the use of an additional trimming surface.

Particle image velocimetry (PIV) measurements of the flow around a wing section are employed as a basis for non-intrusive aerodynamic mean loads characterisation, providing sectional lift, drag and pitching moment. The technique relies upon the application of control-volume approaches in combination with the deduction of the pressure from the PIV experimental data through application of the momentum equation. The treatment can also be applied when the flow is unsteady; in that case time-mean loads are obtained from velocity statistics, through the use of Reynolds-averaged formulation of the governing equations. The procedure was applied in the experimental investigation of a NACA 642A015 aerofoil, in which the PIV approach is validated against standard pressure-based methods (surface pressure distribution and wake rake). The chord Reynolds number considered in the investigation ranges between 1 – 7 × 105. In addition, the consistency and potential performance of the method was assessed by means of synthetic velocity field data obtained from a numerical flow simulation.

In 1961 I assisted Mr. G. M. Lilley in writing a survey of the design and performance of man powered aircraft. Our conclusions were that man powered flight is just possible if the aircraft is carefully designed to give excellent aerodynamic characteristics, with a stiff, light structure and flown by a pilot whose power output is comparable with that of a National Amateur Cycling Champion. Such a machine would not fly more than fifteen to twenty feet above the ground and could operate only in still air conditions. The type of aircraft which emerged from the calculations was similar in general principles to those being developed currently at Hatfield and Southampton.

In this lecture I propose to investigate some of the basic relationships which are involved in the hope that possible ways of obtaining further improvements may emerge. Since the conclusion of our survey was that flight was marginally possible, slight improvements might well go a long way towards making man powered flight possible.

Formation flight could greatly assist the air transport industry in tackling the challenges of environmental impact, excessive reliance on fuel and overcapacity. Previous studies have shown drag reductions leading to significant fuel savings for aircraft in formation relative to their solo flight. Safety is guaranteed with the use of extended formation distances, and practical implementation issues could be solved in the near future. Since studies so far have focused on existing aircraft configurations and technology, a case study using a strut-braced wing airliner was carried out to ascertain its applicability to less conventional craft. The present results did not indicate such clear-cut benefits. If formation flight is to be successful and beneficial for the next generations of aircraft, it will be vital to consider its interaction with new technologies developed for highly efficient operation, in particular those aimed at reduction of aircraft drag such as laminar flow, and to do so early in the design of aerospace vehicles and wider systems.