The complicated flow in the tip region of a finite circular cylinder in uniform cross flow has been examined at the Reynolds numbers 0·85 × 105, 1·8 × 105, and 7·7 × 105. Simultaneous measurements of the surface-pressure and wake-velocity fluctuations have revealed the existence of a shedding regime in the tip region that is distinct from the one prevailing on the main body of the cylinder. In particular, this regime can be unstable and intermittent, can have a cellular structure in the wake, or can be subcritical when the main flow is supercritical.

The dynamic contact angle formed when silicone oil displaces air from the surface of glass has been measured. Even though the glass was neither treated in any special way nor cleaned by an elaborate technique, the standard deviation associated with our measurements was approximately 1·5°. A sequence of experiments revealed that the dynamic contact angle depends on both the speed at which the oil spreads across the glass and the size of the characteristic length scale associated with the device within which it is measured. It is shown that the latter implies directly that: (i) the measured and the actual contact angles are not the same; (ii) the usual hydrodynamic model for fluids is inadequate when a moving contact line is present. These conclusions are consistent with recent theoretical studies.

An investigation of the errors arising in pulsed-wire anemometer measurements of the Reynolds stresses in turbulent flows is described. Attention is concentrated first on a theoretical approach, in which an idealized yaw response and an assumed form for the joint velocity probability-density distribution are used to determine the errors in measurements of, principally, $\overline{uv}$ and $\overline{v^2}$ when the probe is used like an ordinary single slanted hot wire. Actual pulsed-wire measurements in a range of turbulent shear flows are then compared with crossed-hot-wire results and with the theoretically simulated pulsed-wire response obtained from calculations and the crossed-wire data. It is shown that whilst pulsed-wire measurements of lateral intensity and shear stress are inevitably rather unsatisfactory in regions of low intensity (less than 10%, say) they agree reasonably well with crossed-wire measurements in flows where the intensities are higher, but do not exceed those for which sensible corrections to cross-wire data are possible (up to, say, 30%). In this medium-intensity range, however, the pulsed-wire errors are found to be critically dependent on the finite limit of the pulsed wire's yaw response; it seems that acceptable measurements can only be made if this exceeds about 75°. Beyond an intensity of about 30% the errors in $\overline{v^2} $ measurements (which are usually much higher than those for $\overline{uv}$) become less dependent on the exact nature of the yaw response and invariably decrease with increasing intensity. They can, with care, be made as low as 15%. It is concluded that pulsed-wire measurements of the Reynolds stresses can be made with an accuracy similar to that of crossed-wire measurements in medium-intensity flows. Such measurements are certainly adequate for many practical purposes in high-intensity flows where hot-wire techniques are useless.

Wall-pressure fluctuations were measured in the plane of symmetry and five off-axis positions in artificially generated turbulent spots in a laminar boundary layer. Ensemble averages of the wall-pressure signatures show that the spot's pressure signature is characterized by two positive-pressure peaks and a central negative-pressure region along the centre line which evolves into one negative- and one positive-pressure perturbation in the wing-tip region. In similarity co-ordinates, based on pressure-signature characteristics, the spatial growth of the pressure signature shows a small Reynolds-number dependence. Contours of the ensemble-averaged wall pressure measured relative to the laminar value show that a turbulent spot produces a leading-edge pressure excess followed by a pressure-defect region having its maximum values at the spot wing tips. The trailing edge of the pressure disturbance was found to consist of a spanwise region of pressure excess, also having its maximum value at the wing tips.

Visualization of successive vortices impinging upon the leading edge of a wedge reveals patterns of deformation of each incident vortex; for certain offsets of the edge with respect to the incident vortex there is pronounced vortex shedding from the leading edge, whereby the shed vortex has a vorticity orientation opposite to that of the incident vortex.

Simultaneous consideration of this visualization interaction and the force induced on the wedge gives the relation between the nature of the interaction mechanism and the relative magnitude and phase of the force exerted on the wedge. The amplitude of the induced force is found to be a strong function of the transverse offset of the leading edge with respect to the incident vortex and the degree of vorticity shedding from the leading edge. Application of Stuart's vortex model to the incident vortices provides a means for approximating the phase and relative amplitude of the induced force as a function of the transverse offset of the leading edge.

This paper describes an experimental study of the flow associated with the intrusion of an air cavity into a long horizontal duct as water was allowed to drain from one end. Flows of this nature were discussed by Benjamin (1968), who showed that throttling of the flow of water from the end of the duct would cause both the celerity and the depth of the cavity to reduce. However, the experiments described in this paper revealed that the celerity of the cavity was not reduced from its unthrottled value until the water depth beneath the cavity was 0·78 of the duct depth. For values of this depth ratio between 0·5 and 0·78, the flow as a whole was unsteady. It is shown that Benjamin's model can be modified to allow for the unsteady nature of the flow. Benjamin's original model was found to describe accurately the form and behaviour of the cavity in the case of unthrottled flow, when the flow was steady, and also when the depth beneath the cavity exceeded 0·78 of the duct height, when the flow was again steady. Surface-tension effects were found to reduce the celerity of the cavity and to modify its shape as described by Gardner & Crow (1970).

The transient natural convection of a mass of water confined within a closed cavity with wall temperature decreasing at a steady rate is considered. For situations where a linear density-temperature relationship can be assumed, a quasi-steady state following an initial transient may be reached, provided that the cooling rate applied to the wall is held constant long enough. Steady-state flow characteristics in the case of a specific geometry are functions of a single parameter, the Rayleigh number, in which a dimensionless temperature, based on the cooling rate, is used. For the particular case of water cooled through 4 °C, the temperature at which maximum density occurs, a linear variation of density with respect to temperature is no more acceptable. However, it can be assumed that a linear relationship between the water thermal-expansion coefficient and the temperature is valid in the neighbourhood of 4 °C. With such an assumption it is still possible to characterize the cooling process that follows the initial transient by a single parameter. Detailed numerical results are presented for the particular case of a square cavity. Existing experimental and numerical results for the case of a horizontal circular pipe are also discussed.

Extensive measurements were made of the response of a high-Reynolds-number turbulent boundary layer to a short length (about three boundary-layer thicknesses) of surface roughness. The measurements include the mean velocity, all the Reynolds streses as well as all the triple products occurring in the Reynolds-stress transport equations. At the last measurement station, 55 boundary-layer thicknesses from the roughness, the boundary layer still had not relaxed to the universal smooth-wall structure. Comparison of the measurements with a calculation of the flow suggests the major requirement for improving the agreement is a more sophisticated treatment of the turbulent-diffusion process.

The effect of increasing disturbance size on the stability of a laminar streaming flow is considered theoretically at high Reynolds numbers Re. The theory has a rational basis that allows detailed understanding of the delicate physical balances controlling stability, and is presented with an accelerating boundary layer taken as the basic flow. The theory predicts that the scales and properties required to produce the Rayleigh situation (where the disturbances have wave speed and wavelength comparable to the typical speed and thickness respectively of the basic flow) in neutral stability are very different from those predicted by a classical approach, involving a relative disturbance size O(Re−⅙) rather than the classical suggestion $O(Re^{-\frac{1}{3}})$. Before then, however, the disturbances undergo an abrupt alteration in scale and character as they pass through the just slightly smaller size $O(Re^{-\frac{7}{36}})$, with the stability structure changing from the relatively large-scale form of linear theory to the more condensed Rayleigh form by means of a nonlinear interaction within the critical layer. Strong higher harmonics of the fundamental disturbance are induced throughout the flow field by the velocity jump across the critical layer, but the phase jump remains the most significant property. Solutions for the nonlinear critical layer are recalculated and reanalysed. Also, the mean-flow correction produced by the nonlinear critical layer is shown to be smaller than the main part of the fundamental, owing to the regularity of the latter. As the Rayleigh stage is approached, the lateral variation of the induced pressure force through the critical layer begins to exert a considerable influence. Similar characteristics also arise in other fundamental streaming flows, and the implied Rayleigh stage is the subject of a subsequent investigation.

A new technique using surface-film activators has been developed to induce and control laminar-instability waves by periodic heating. A flat plate was instrumented and installed in the GALCIT High-Speed Water Tunnel with flush-mounted surface heaters and probes. Extremely two-dimensional naturally occurring Tolmien-Schlichting (TS) waves were observed along with the subsequent formation of turbulent spots. Laminar-instability waves were then excited in a controlled fashion using the surface-mounted heaters. A preliminary experiment on cancellation of excited laminar-instability waves was carried out. Finally, turbulent spots were produced using amplitude-modulated bursts to form Gaussian TS wave packets. Flow visualization, along with wall shear measurements, was used to infer the velocity and vorticity field near the wall.

Instability waves, commonly called T-S waves, can be introduced in a laminar boundary layer by periodic heating of flush-mounted heating elements. Experiments have demonstrated that nearly complete cancellation of a T-S wave excited in this way can be achieved by using a second downstream heating element with a suitable phase shift. As one application of the technique, a single element together with a feedback loop activated by measured wall shear stress has been used to reduce the amplitude of naturally occurring laminar instability waves. A significant increase in the transition Reynolds number has been achieved.

Results of direct numerical simulations (DNS) for the decay of an initially Gaussian field of turbulence convecting a passive scalar are compared with equivalent results for the direct-interaction approximation (DIA) and the test-field model (TFM). The Taylor microscale Reynolds number Rλ and the equivalent Péclet number Pλ of the comparison ranged from 20–8 and 10–4, respectively. The Prandtl number Pr equals 0·5. Our results show a satisfactory agreement of both theories and numerical simulations, with the DIA giving better overall agreement, especially at small scales. This improved small-scale agreement - which appears to hold up to Rλ ≃ 30 - is related to the relatively long coherence times of the small scales, and to the fact that the TFM, containing as it does a built-in compliance to the fluctuation dissipation theorem, cannot properly cope with this fact. We also give a comparison of results for the velocity skewness with the experiments of Tavoularis, Bennett & Corrsin (1978).

This paper examines the problem of the survival of Salter's duck wave-energy device in extreme waves, both experimentally in a narrow tank at Edinburgh University and theoretically, using programs originally designed to study nonlinear ship motions (Vinje & Brevig 1981). Both approaches are essentially two-dimensional, corresponding to waves normally incident upon the duck string, and comparison of the two sets of results for a duck on a fixed axis shows good agreement.

A series of measurements, obtained using arrays of hot-wire anemometers, has been performed in the fully developed region of a plane turbulent jet. The anemometer output signals were simultaneously sampled, digitized with sufficient resolution for numerical linearization, and recorded on magnetic tape for subsequent analysis.

The data was processed to extract information about the structure of the large eddies within the flow. Firstly, a selection of the two-point velocity correlation functions was evaluated, and diagrams of the contours of constant correlation for the stream wise velocity component were constructed. Secondly, an iterative procedure similar to the techniques generally described as ‘pattern recognition and image enhancement’ was used to form ‘ensemble’ averages of the two-dimensional patterns of the streamwise velocity component associated with the large eddies.

The results indicate that the large eddies in the fully turbulent region of the flow are roller-like structures with axes aligned approximately either with the direction of the strain associated with the mean-velocity gradient or with the direction of homogeneity (spanwise). It was found that these basic structures tended to occur in various preferred combinations. Estimates were obtained for the total intensity contribution from the large eddies, the ranges of sizes and intensities of the individual structures, and their effective packing density.

Oscillatory flow in ∪-shaped pipes has been investigated experimentally. The velocity field is mapped by observing the motion of tracer particles either photographically or by laser-Doppler anemometry. The most important result seems to be the observation of regions with strong secondary streaming at the inlet of the curved section, which, to the best of our knowledge, has not been reported before.

The Korteweg-de Vries (KdV) equation and the finite-depth equation of Joseph (1977) and Kubota, Ko & Dobbs (1978) both describe the evolution of long internal waves of small but finite amplitude, propagating in one direction. In this paper, both theories are tested experimentally by comparing measured and theoretical soliton shapes. The KdV equation predicts the shapes of our measured solitons with remarkable accuracy, much better than does the finite-depth equation. When carried to second-order, the finite-depth theory becomes about as accurate as (first-order) KdV theory for our experiments. However, second-order corrections to the finite-depth theory also identify a bound on the range of validity of that entire expansion. This range turns out to be rather small; it includes only about half of the experiments reported by Koop & Butler (1981).

Moffatt's (1978) model for instability, by magnetic buoyancy, of a rotating layer of compressible fluid is extended to demonstrate that local stability analysis, theoretically valid for only a shallow layer, can still be useful for predicting instability when the layer is deep.

The principle of large-scale anisotropy due to small-scale layering is applied to thermal convection. The motion takes place in a bounded porous medium heated from below. The medium is periodically layered with respect to permeability and thermal conductivity. The onset of convection as well as slightly supercritical convection are investigated. Anisotropic modelling proves useful even for small numbers of layers as long as the motion is of ‘large-scale convection’ type (Masuoka et al. 1978). The modelling always fails for motion of ‘local convection’ type.

Fully developed turbulent channel flow has been simulated numerically at Reynolds number 13800, based on centre-line velocity and channel half-width. The large-scale flow field has been obtained by directly integrating the filtered, three-dimensional, time-dependent Navier-Stokes equations. The small-scale field motions were simulated through an eddy-viscosity model. The calculations were carried out on the ILLIACIV computer with up to 516096 grid points.

The computed flow field was used to study the statistical properties of the flow as well as its time-dependent features. The agreement of the computed mean-velocity profile, turbulence statistics, and detailed flow structures with experimental data is good. The resolvable portion of the statistical correlations appearing in the Reynolds-stress equations are calculated. Particular attention is given to the examination of the flow structure in the vicinity of the wall.

An analysis is provided for the free development of slender jets of water, in which the cross-sections are of small thickness-to-width ratio.