A turbulent boundary layer, subjected to a freestream velocity which changes sinusoidally with time around a spatially uniform mean value has been studied experimentally. The experiments were conducted in a specially designed unsteady-flow water tunnel. A two-component laser Doppler anemometer was used to measure the instantaneous velocities, and hence the turbulent stresses, in the periodic boundary layer. A surface-mounted heat-flux gauge was used to measure the unsteady wall shear stress. The experiments were conducted at relatively high amplitudes of oscillation (≈ 40% of the mean) and over a wide range of reduced frequencies (1.7-28). While the present data on the oscillatory flow properties generally confirm previously known trends the information on the turbulent shear stress and skin friction obtained in the present studies is more complete than previously available. The experimental data, all of which has been archived on magnetic tape, form a comprehensive base for the development and verification of predictive models for unsteady turbulent boundary layers.

The development of a new calculation method for compressible 3D boundary layers is described. The method involves a finite-difference discretisation of the governing mean-flow equations. In particular, the differencing scheme used to discretise spanwise derivatives adapts automatically to the sign of the local crossflow within the boundary layer. A plane-by-plane solution procedure in the spanwise direction enables second-order accuracy to be maintained throughout the whole flowfield. A normal coordinate scaling with the local total momentum thickness removes most of the boundary layer growth in computational space. The Cebeci-Smith algebraic turbulence model is used for the initial validation of the calculation method. A simple modification to this model is tested, involving an explicit dependence of the outer eddy viscosity on the crossflow within the boundary layer. There results a significantly improved prediction of the NLR infinite swept wing flow experiment.

The effect of trailing edge extensions was tested for the twodimensional Göttingen 797 and Wortmann FX 63-137 aerofoil sections with smooth surfaces at chord Reynolds numbers from 3 x 105 to 1 x 106 in the 8ft x 6ft low speed wind-tunnel at the Cranfield Institute of Technology. Force measurements were made on the basic aerofoil sections and with extension plates of 10% and 20% chord set tangential to the upper surface at the trailing edge. It was verified that the addition of an extension plate effectively brings about an increase of maximum lift coefficient and an improved lift-to-drag ratio for the Göttingen aerofoil section. However, the effect of the extension plate was small on the Wortmann aerofoil section.

The motion of an aircraft along a straight flight path inclined to the vertical is considered from an arbitrary initial velocity which may be far from steady flight speeds. The velocity and incidence, as a function of distance or time, provide a solution to the ‘inverse phugoid problem’ of finding the control laws that keep an aircraft on a straight flight path by exactly cancelling the phugoid instability (the short period mode is omitted by neglecting the rotational inertia of the aircraft). This is a non-linear problem, namely finding the motion of an aircraft from an initial condition far removed from a stable steady state, or finding the large amplitude deviations from an unstable equilibrium. The stability and instability problems are solved analytically using the following assumptions:

1. (i) the aircraft flies subsonically so that only form and induced drag need be considered;

2. (ii) the thrust consists of a constant term, minus a term specifying a dependence on velocity similar to that of the total drag; and

3. (iii) the aircraft incidence is away from stall so that lift is a linear function of incidence.

The response curves show that:

1. (i) flight at velocities below the lower steady flight speed, which is unstable, is difficult to maintain due to fast divergence towards the stall; and

2. (ii) for an initial velocity above the lower steady flight speed, there is convergence towards the higher steady flight speed.

The convergence, in the latter case (ii), may be slow along a straight flight path, although it can be hastened by suitable manoeuvres.

This is the second part of a paper on aerospace transporters (i.e. fully reusable aeroplane-like piloted launch vehicles) and describes a design concept intended to meet short-term European requirements at minimum development cost. It is generally similar to the European aerospace transporters studied in the 1960s but uses technology developed since then for such projects as Concorde, Spacelab, Ariane and the Space Shuttle.

It is concluded that such a vehicle could be built with more or less existing technology, and, using the costing approach described in the first part of the paper, that the development cost should be less than that of the semi-reusable Hermes/ Ariane 5 combination. A full feasibility study is recommended to check these conclusions.

The effects on the stress intensity factor of bonding repair patches over a crack in a large sheet under uniaxial loading are studied using two simple numerical models. The variation of the stress intensity factor with increasing crack length is investigated and results for repairs involving large elliptical patches are found to agree with a closed-form approximation. Further cases are considered where the patch is either rectangular in shape or less than a certain critical size. For such repairs the closed-form approximation for the stress intensity factor is shown to be unsuitable.

Manned spacecraft cost typically ten times as much to develop as advanced technology prototype aeroplanes of similar weight. For example, the Hermes spaceplane, which is designed to be launched into orbit by the expendable Ariane 5, has an estimated development cost of $4 billion, which is about ten times greater than the cost of the Experimental Aircraft Programme (EAP) demonstrator of the European Fighter Aircraft ($350 million). The Hermes spaceplane and the EAP are of comparable size and weight and both use very advanced technology. The EAP is already more mature than Hermes would become because, although the EAP is only a prototype, it has made more flights than are planned for Hermes over its entire operational career.

This part of the paper attempts to identify the factors which explain this difference in cost and then to determine which of them would apply to fully reusable, piloted, aeroplane-like launchers, i.e. aerospace transporters.

A fundamental study is presented of the three-dimensional aspects of the wake/boundary layer interaction as it occurs on high-lift wings, particularly the slat-wake/wing boundary layer interaction. A digital system for measuring and recording all the components of the mean velocity and the Reynolds stress tensor using a triple hot-wire probe has been developed. This system has been tested and has been used to measure the interaction between the wakes of swept aerofoils and the fully turbulent boundary layer developing on a flat plate. An important and interesting result of all the tests is the rapid decay of the three dimensionality in the wake, so that in many cases the actual interaction is almost two-dimensional. Two cases of three-dimensional interaction have been simulated. In one case the interaction is fully three-dimensional while in the other case it can be assumed to be quasi three-dimensional. For the quasi three-dimensional confluent flow, the measured results have been compared with the predictions of the two-dimensional form of K-ε and algebraic stress turbulence models. The overall agreement between the measured and predicted results is good. However, for the turbulence flow field the agreement is less satisfactory in the initial mixing region where the three-dimensionality is strong and it is also poor in the wall boundary layer.

This, the third part of the paper on aerospace transporters, considers their applications. It is concluded that a small nearterm vehicle would be useful for launching small satellites, for servicing satellites in orbit, for supplying space stations and as a lead-in to larger and more advanced vehicles. The latter have a potential cost per flight about 1000 times less than that of the Space Shuttle, given traffic levels sufficient to enable design and operational maturity to be achieved. Space tourism, (i.e. the making of short flights to Earth orbit by fare-paying passengers), would appear to be the first application likely to provide these traffic levels. Given high priority and an immediate start, the first space tourism flights could possibly take place this century.

The equations of motion of an aircraft flying at a constant glide slope (Fig. 3), in the presence of arbitrary head- or tailwinds, and up- or downflows is considered. The equations are integrated analytically, in the case of an aircraft initially on a steady flight, perturbed by winds of ‘moderate’ strength, in the sense that the wind velocity is not negligible compared to the aircraft's steady speed, but the ratio of their squares is much smaller than unity. The case of an approach through a downburst (Fig. 1), leads to winds which can be simplified to a one-period sinusoidal wind along the flight path, changing from head- to tailwind, at the peak of a superimposed downflow (Fig. 2), of half-period sinusoidal shape, and this is discussed in some detail. Data sheets are presented for three combinations of amplitudes of the head-to-tailwind and downflow; each contains plots of the scheduling of incidence that exactly cancels windshear effects, and of the groundspeed a'nd airspeed profiles which will keep the aircraft flying along the original glide slope. Each plot is given for a range of values of the windshear susceptibility parameter, representing aircraft with small or large inertia, with high or low approach speeds, including light aircraft, jet fighters and large transports; the cancellation of windshear effects isachievable only if the incidence schedule lies wholly below the stall limit. Since the rotational inertia of the aircraft is neglected, the short period mode is absent, and the pitch control acts to cancel the ‘phugoid’ instability induced on the aircraft by the wind profiles typical of a microburst.

The aerodynamic interference between typical tanker and large receiver aircraft during air-to-air refuelling was modelled using a simple horseshoe vortex representation of the tanker wake whilst the aerodynamic forces and moments acting on the receiver were calculated using the vortex lattice method. For each combination of tanker and receiver, the aerodynamic loads depend mainly on the vertical separation, whilst the receiver tailplane position on the fin strongly effects the pitching moment acting on the receiver due to the tanker downwash. Particular attention was given to a high T-tail receiver aircraft. The predicted increase in the receiver drag compares favourably with available flight test data. Changes in trim in pitch of the receiver were analysed and related to the static stability of the aircraft. Aerodynamic derivatives due to the position and attitude of the receiver aircraft within the tanker downwash field were calculated and solutions of the linearised equations of motion obtained for the high T-tail receiver refuelling from different tanker aircraft at different flight conditions. Typically the receiver exhibited a divergent oscillation at the normal refuelling position. The dynamic stability characteristics are consistent with the trends observed in flight tests on the tendency of the receiver to oscillate in pitch.

An economic numerical method has been developed to calculate the incompressible potential flow about complete aircraft configurations by taking advantage of the computational efficiency of the internal distribution of singularities. For wing and wing-like components, source and vorticity singularities are distributed on the respective mean camber surfaces, while the fuselage carries a source distribution on its wetted surface. The singularity strengths are obtained by satisfying the flow tangency condition at selected points on all the wetted surfaces of the configuration. No attempt has been made to consider wing or body vortex shedding or separation. The results of the present method were compared with other theoretical and experimental data.

Detailed measurements of atmospheric-turbulence velocity have been made by a specially instrumented aircraft at altitudes below one thousand feet over a variety of terrains. These measurements are analysed in terms of power-spectral-density and a statistical-discrete-gust model. To a good approximation, spatial increments in turbulence velocity are shown to satisfy scaling laws implying self-similarity for amplitudes up to about four times their root-mean-square value and over a wide range of scales. A three-parameter relationship is established between two parameters from the statistical-discrete-gust analysis, corresponding respectively to intermittency and gust intensity, and one from the power-spectral representation, related to the average rate of energy dissipation.

A comparison is presented of three methods for the generation of numerical, modal approximating functions for use in modal Lagrangian analysis of rotating flexible blades. The methods considered are those based on the use of uniform beam/bar eigenfunctions, smooth bending moment or torque modes, and modes generated by recourse to one stage of the Stodola method. For blades which are highly non-uniform in their specific stiffness and inertial properties, and where the objective is to use a small number of approximating functions, consistent with accurate determination of eigen-solutions in the fundamental spectrum, it is demonstrated (as is well known) that direct use of uniform system eigenfunctions is unsatisfactory. On the other hand, it is demonstrated that the use of smooth bending moment modes, even in cases where the variations in sectional inertia properties are very large, can produce excellent results. The use of ‘Stodola modes', however, is shown to offer the all-important advantage of enhanced convergence rate in all cases considered.

This fourth and final part of the paper on aerospace transporters attempts to explain why, if aerospace transporters are as technically and commercially viable as has been suggested in the previous parts of the paper, they were not developed several years ago.

A brief historical survey of space transportation reveals two unique features compared with other forms of transportation. First, technically straightforward methods of greatly reducing cost have persistently failed to be adopted. The feasibility of winged flyback boosters was demonstrated as long ago as 1945 when the Germans flew two experimental V-2s fitted with wings to increase range. Winged flyback boosters have never been built. Second, all significant development funding has come from government agencies - there has been very little private investment in launch vehicle development. It is suggested that the link between these two features is that space policy was initially dominated by considerations of defence and national prestige, and that the commercial uses of space have not become significant enough to give priority to cost reduction.

In this situation, it is proposed that developing a small aerospace transporter would assist the permanent manned space stations now being planned by government agencies, by providing safe and low-cost transport for their crews; and would accelerate the commercial exploitation of space by greatly reducing launch costs and thereby permitting the start of a space tourism industry. However, at present there are no firm plans for aerospace transporters, and passenger transportation is not being considered by government agencies.

The main conclusion of the paper is that feasibility studies of aerospace transporters should be undertaken soon to include particular emphasis on the prospects for commercial space passenger transportation.

Tests were made to investigate the effects of Mach number on the oscillatory lateral derivatives of a three-surface, high-incidence research model (HIRM 1). Effects of Reynolds number and frequency parameter were also investigated at low speed. A new flexible sting and hydraulic actuator system were designed for the tests in the 2.4 m x 2.4 m Pressurised Wind Tunnel.

Results showed that the effects of Reynolds number and frequency parameter were not very significant at low speeds but there were quite large effects of Mach number on some derivatives, notably yawing and rolling moments due to sideslip and rate of yaw.

This paper presents a method for assessing two-dimensional aerofoil lift and pitching moment characteristics including trailing edge and gross laminar separation. The model used is a direct viscid-inviscid interaction scheme based on a vortex panel method with boundary-layer corrections and an inviscidly modelled wake. The integral boundary-layer methods adopted behave well in the region of separation and thus, good comparisons with measured separation characteristics are obtained. Generally the predictions of lift and pitching moment may be considered to be within the experimental error, but where this is not the case, the applicability of the modelling technique is discussed.

An asymptotic theory valid for large Reynolds number is developed for describing the oscillatory properties of a turbulent boundary layer in a free stream whose velocity varies harmonically with time around a non-zero mean value. The theory identifies two frequency parameters in terms of which four different regimes of oscillation frequency are defined. Similarity laws are identified for each of these frequency regimes and for the different layers of the periodic boundary layer. The theory seems to bring together the available data on unsteady wall bounded flows into a general framework.

Experimental and theoretical, small and large amplitude stability data are presented for a pointed and a 0.2 bluntness ratio, 10° semi-angle cone performing pitching oscillations in hypersonic flow at a Mach number of 6.85. Analysis identifies that large amplitude model motion time histories cannot be predicted from a knowledge of small amplitude oscillation stability derivatives data. At the Reynolds numbers of the experiments the pointed and blunted cone are subject to significant hypersonic flow viscous phenomena, which are proposed as the cause of the small to large amplitude stability prediction being invalid.