A theoretical analysis of the flexural wrinkling of a honeycomb sandwich beam with laminated faces having non-linear stress-strain behaviour is presented. Appropriate tangent moduli are used in the constitutive equations for the compression face and a double iteration technique is used to calculate wrinkling loads. Numerical results are given for sandwich beams with unidirectional and cross-ply carbon fibre reinforced plastic (CFRP) faces. Experimental failure loads of sandwich beams are compared with theoretical predictions, and, in general, the agreement between the two is good.

Uncertainties in the use of the discrete-vortex method in modelling the time development of the wake of a circular cylinder at very high Reynolds numbers are investigated. It is shown that simply introducing vorticity at generally accepted separation positions at a rate of ½Us2, Us being the velocity at separation, gives wholly unrealistic wake predictions. In the base region pressure fields occur which would promote separation in steady flow and so a first approximation for ‘secondary’ separation is incorporated into the model. This brings pressure distributions and vorticity structures at subcritical and supercritical Reynolds numbers into good agreement with experiment. The convection of the vortices is calculated using the cloud-in-cell technique and comparisons are made with direct summation methods.

An inviscid model, originally proposed by Bryson to describe the formation of laterally-symmetric vortices above bodies of revolution at large angles of incidence, is extended to describe flows without lateral symmetry. Each vortex is represented by a single line-vortex and conditions of cross-flow stagnation are applied at the postulated separation lines. The transverse component of the force acting on each vortex is balanced by an equal and opposite force acting on the cut connecting it to the adjacent separation line. It is shown that this model allows asymmetrical vortices to form in the flow over slender circular cones even when the separation lines are disposed symmetrically.

An improved method, based on one strip approximation of the method of integral relations which was reported originally by Belov, Ginzburg and Shub, is presented for the calculation of flow parameters in the impingement region of a supersonic, underexpanded jet striking a normal surface located within the first cell. The results are presented for two impingement conditions and found to be in good agreement with the experimental data.

Measurements are presented of the static pressure, total pressure, swirl and three-dimensional turbulence in the flow through an S-curved rectangular duct mounted in a wind tunnel at different incidences, yaw angles and mass flow ratios. The results show that at high incidence there is a large vortex around an area of flow separation after the first bend and a pair of contra-rotating vortices in the flow after the second bend. The distortion of total pressure at the exit of the S-duct is significantly high, and the corresponding three-dimensional turbulence is up to 16.7%. Appreciation of the existence of this type of flow is important in the field of air intake design for jet aircraft.

A complete unsteady Newton-Busemann flow theory, including centrifugal force corrections, was recently given by Hui and Tobak and successfully applied to study the dynamical stability of oscillating aerofoils (Part I) and bodies of revolution (Part II). The present paper extends these results, which are restricted to the first order in frequency, to general frequencies that may be applicable to flutter analysis as well. Furthermore, the new results are used to investigate the behaviour of the indicial response functions in unsteady flow at very high Mach numbers. In particular, it is found that for a family of body shapes in the Newtonian flow, including the cone, the wedge, the delta wing and slender aerofoils, the aerodynamic response to a step change in angle of attack or pitching velocity contains an impulse at the initial instant. This is followed by a rapid adjustment to the new steady-flow conditions, the adjustment time being equal to that required for a fluid particle to travel a distance of the body length. The impulse component appearing at the infinite Mach number limit (Newtonian flow) is in effect an apparent mass term, analogous to that which occurs initially in the aerodynamic indicial response at the zero Mach number limit (incompressible flow). In the latter case the initial impulse is followed by an asymptotic adjustment to the new steady-flow conditions, theoretically taking infinite time, in direct contrast to the rapid adjustment in Newtonian flow.

The unsteady lift generated on a NACA 0015 aerofoil, a D cylinder (with the flat face down-stream) and an elliptic cylinder was measured when these bodies were exposed to a flow with a two-dimensional sinusoidal upwash at reduced frequencies 0.05 to 0.8. The mean flow Reynolds numbers were in the range 1.2 × 105 to 3 × 105. Unsteady thin aerofoil theory was used in an attempt to predict the unsteady lift on the bluff bodies, as well as the aerofoil section for fequencies in the low range below the vortex shedding frequency. The results were quite accurate for the aerofoil and the D cylinder, but the aerodynamic admittance predicted by this theory for the elliptic cylinder was significantly above that measured experimentally. The movement of the two free separation points was found to play an important role in the characteristic lift behaviour of the elliptic cylinder.

The types of pitch/yaw coupling which may result from roll motion of a conventional tail-controlled cartesian missile have been categorised. The consequences of the couplings upon:

1. (a) the response to a step demand for manoeuvre, and

2. (b) the decay of error in a beam-rider system are presented.

Potentially the most damaging forms of coupling are due to

1. (a) mis-match of the Invariant Frames of airframe and servos,

2. (b) angular mismatch between parts of the system, and

3. (c) gyroscopic coupling.

Conventional force-balances, “accelerometer balances” and photographic monitoring of freely-flying models are not suitable for use in the Q.M.C. shock tunnel where the force levels are of order 10N for flow durations of about 1 ms. An interferometric method has therefore been developed for following the trajectory of a weakly-restrained model. The prototype system consists of two simple Michelson interferometers, a single He-Ne laser being employed to provide two measurement and two reference beams so that the motion of two points on the model can be followed. The measurement beams are returned by corner-cube retroreflectors carried on the model which ensures that for each measurement arm the reference and measurement beams recombine at the surface of a photo-detector. As the model moves, interference fringes are produced at the detectors, the cycle dark-light-dark corresponding to a model displacement along the measurement beam of ½λ, about 0.3μm. The frequency-modulated wave-trains produced are recorded using two transient recorders, the data being subsequently played back to a two-channel pen recorder giving a record 500 mm in length corresponding to the test time. The fringe number as a function of time is read manually, and the data analysed by curve fitting to a parabola which yields the accelerations of the measurement points. A knowledge of the inertial characteristics of the model then gives the forces on it. By suitably aligning the beams, lift and pitching moment for a ridge-delta of aspect-ratio 1 were obtained. Two models of the same geometric size but of different inertia were tested. All the data were obtained for model displacements less than 1 mm and pitch rotations less than 0.1°.

A general procedure for the analysis of thermal stresses in a polygonal elastic plate due to heating over a small circular or rectangular area is presented. The method employs eigenfunctions of angular regions together with a least squares procedure to establish continuity. Numerical results are given.

The unsteady breakdown or stall of streamlined flow near the rounded leading edge of an aerofoil, as the small angle of attack is raised above its critical stall value, is studied theoretically for large Reynolds number motion. The unsteady developments take place first over a relatively slow time scale but then the corresponding solution breaks down with a singularity, forcing a switch to a faster and more nonlinear process. The latter involves a very pronounced local bulge appearing in the flow displacement, accompanied by reversed flow at the aerofoil surface, and comparisons with experimental observations of dynamic stall are noted.

The problem considered is that of Poiseuille flow in a two dimensional channel impinging on a flat plate placed symmetrically between the two channel walls so as to bifurcate the channel into two equal branches. The results of a theoretical study of the problem are presented and analysed to portray in full details the characteristics of the flow before and after the bifurcation point. In particular the effects of the flow divider on the flow near the channel walls and the effects of the channel walls on the flow near the flow divider are examined.

Simulations of vortex sheet dynamics by the discrete vortex method have often shown a growing randomness of the discrete vortices. This paper describes two techniques which attempt to mitigate this tendency whilst preserving the overall structure of the vortex sheet. Clusters of vortices near the points of concentration are amalgamated into a single equivalent vortex and a process of repeated rediscretization of the remaining part of the sheet maintains the integrity of its representation. Use of these techniques allow the flow model to be followed beyond previous expectations, without undue randomisation of the vortex positions. The effects of using these techniques on the overall vortex sheet development are fully investigated in order to validate their use in future investigations.

A theoretical method is used to predict the surface pressure distribution and drag-rise characteristics of blunted ogive forebodies. The results confirm the observation that a moderate degree of nose blunting can reduce wave drag relative to that of a pointed body of the same fineness ratio. An analysis of the theoretical data provides a simple explanation of this phenomenon.

An exact analysis is first given, within the spirit of elementary theory, of the torsional modes of a uniformly tapered solid wing whose section shape - as distinct from size - is constant. Numerical results are presented for the first five torsional modes for various values of tip chord/root chord. It is shown that the frequencies are fairly insensitive to the magnitude of the taper and a corollary to this, which is also investigated, is that an analysis based on a partially stepped representation of the taper yields close estimates of the frequencies but can yield massive differences in the higher mode shapes. An analysis is also made of warping restraint and distributed inertia effects.

Detailed measurements have been made of the boundary-layer development on a small two-dimensional aerofoil with supercritical flow and a weak shock wave, together with similar measurements on the tunnel side wall opposite the aerofoil surface. The Reynolds number of the test is similar to that found in the turbines of jet engines and there is a strong favourable pressure gradient ahead of the interaction of the shock with the boundary layer as often occurs in turbine blade passages. However, whereas the boundary layer on the aerofoil is thin and of the same thickness as that on a turbine blade, the thicker boundary layer on the wall is more typical of that on the hub or casing. The experimental results are compared with results from a wide range of calculation methods. One interesting conclusion from these comparisons is the fact that prediction methods which perform well for the thin boundary layers on the aerofoil do not necessarily perform as well for the thicker boundary layers on the wall.

A comparison is made between three models of the two-dimensional flow of an inviscid fluid over a bluff body. The models are the Roshko free-streamline model, the Parkinson and Jandali wake-source model and the Bearman and Fackrell vortex-lattice model. The models are used to predict the surface pressure distribution on the front face of a right-angled wedge for the two symmetrical positions of the wedge, and a comparison is made with the experimental results of Slater. A further comparison is then made of Slater’s results for the surface pressure distributions on a wedge of finite thickness with the distributions predicted by the vortex-lattice model, for the wedge at a number of angular positions to the flow.

The flow around a normal flat plate close to a large plane surface has been investigated for a range of boundary-layer thickness on the plane surface of 0.72 to 2.53 times the plate height, h. As the gap between the plate and the plane surface is reduced below about 0.55h vortex shedding from the plate is inhibited by the presence of the plane surface. Boundary layer thickness has little effect on the gap at which this occurs, but does affect the strength and frequency of vortex shedding at greater gaps. As the gap between the plate lower edge and the plane surface is decreased towards the critical value of 0.55h, both vortex strength and Strouhal number are more significantly affected when the approaching boundary layer is thin. It is not clear whether this is a result of shear or increased turbulent intensity.