Although kites have been around for hundreds of years and put to many uses, there has so far been no systematic study of their performance. This research attempts to fill this need, and considers particularly the performance of kite anemometers. An instrumented kite tether was designed and built to study kite performance. It measures line tension, inclination and azimuth at the ground, :sampling each variable at 5 or 10 Hz. The results are transmitted as a digital code and stored by microcomputer. Accurate anemometers are used simultaneously to measure the wind local to the kite, and the results are stored parallel with the tether data. As a necessary background to the experiments and analysis, existing kite information is collated, and simple models of the kite system are presented, along with a more detailed study of the kiteline and its influence on the kite system. A representative selection of single line kites has been flown from the tether in a variety of wind conditions. The results from these experiments are analysed to obtain general performance measurements for these kites in real, turbulent, winds. The analysis is taken a stage further to evaluate the kites for anemometry, and to study the dynamics of the kite system. The most suitable kites for anemometry are identified, together with the wind conditions in which they may be used and the wind information available from measurements at the tether. The study ends with a review, and a discussion of useful areas of further work. [This digital (pdf ) version of the thesis was created in August 2005. The content is identical to that of the original paper copy of 1986, although the layout and pagination differs in detail. Readers should note that anemometer calibrations described in Hobbs (1994) suggest that the windspeed values given in this thesis are incorrect by a few percent: however, the general conclusions about kite performance stated in the thesis are unaffected by these changes.]

The usefulness of the energy equation integrated over the thickness of the boundary layer, in predicing heat transfer rates to smooth body surfaces in investigated. It is found that on assuming very simple closure relations, similar to those often used with the momentum equation, highly accurate predictions are made. It is shown further that the usefulness of these predictions extend into areas where the momentum equation-skin friction predictions, which have proved so popular, break down such as regions of reverse flow and shock/boundary layer interactions. The technique is has been tested in laminar transitional and turbulent flows with both experiment and other more complex theories. The technique is extended to three-dimensional laminar flows with the inclusion of a crossflow model.

The problem of mathematically modelling the symmetric vortex pair formed on the lee side of a body of revolution at moderate incidence to the freestream direction is addressed with a view to predicting its aerodynamic characteristics. The objective being to develop a model simple enough to enable rapid calculation whilst maintaining acceptable levels of accuracy. Existing techniques are reviewed and their strengths and weaknesses evaluated. The physics of the flow are explored with an emphasis on its three dimensional aspects. Experimental results, including surface pressure tapping data and flow visualisation, are used to investigate the nature of the flow. To gain a deeper insight into the flow processes less measurable experimentally, Navier-Stokes solutions are examined in considerable detail. The mechanisms of vorticity generation and propagation are explored. A simple mathematical model is presented, based upon an extension to slender body theory, which predicts aerodynamic characteristics that compare well with experiment.

This thesis is divided into two distinct parts. Part I describes the design and development of an intermittent cryogenic wind-tunnel, in which the cold conditions are generated by the expansion of high pressure gas. The device uses a light piston moving in a tube and conditions during the running time are maintained constant by 'tuning' the piston motion, i.e. by matching the volumetric flow rate entering and leaving the tube. The results of the pilot tunnel (running time 0.3 secs. ) show that gas temperatures of about 110K can be obtained with a pressure ratio of 35. Part II describes the flow at, transonic speeds on five aerofoil sections (thickness-chord ratios 6 to 14%). The aerofoil sections were 6 & 14% biconvex, an NACA 0012, a supercritical aerofoil CAST 7 and an 11.8% Joukowski profile. The tests were made in an intermittent perforated wall wind-tunnel which was developed from an existing supersonic wind-tunnel. A periodic flow due to shock-induced separation occurred over a narrow range of Mach numbers (from 0.82 to 0.90) on the 14% biconvex and 0012 sections at zero incidence, for both laminar and turbulent boundary layers. The frequency parameter(uc/Um) was about 1. Tests were also made with the aerofoils at incidence. Periodic flows occurred on the 14% biconvex, 0012 and 3oukowski profiles., The instability Mach numbers ranged from 0.84 to 0.90 and the frequency parameters from 0.40 to 1.36. A detailed study was made to determine the frequency spectrum of the tunnel noise and its influence on the periodic flow. Also, experiments were made to determine the influence of the aerofail geometry on the periodic flow. The experimental results on the 14% biconvex aerofoil have been compared with the numerical computations of the full Navier-Stokes equations performed at NASA, Ames.

A two-dimensional model of the aerodynamics of rotor blades in forward flight is proposed in which the motion of the blade is represented by periodical variations of the freestrearn velocity and incidence. A novel implicit methodology for the solution of the compressible Reynolds averaged Navier-Stokes equations and a twoequation model of turbulence is developed. The spatial discretisation is based upon Osher's approximate Riernann solver, while time integration is performed using a Newton-Krylov method. The method is employed to calculate the steady transonic aerodynamics of two supercritical aerofoils and the unsteady aerodynamics of pitching aerofoils. Comparison with experiment and independent calculations for these test cases is satisfactory. Further calculations are performed for the self-excited periodic flow around a biconvex aerofoil. Comparison of quasi-steady and unsteady calculations suggests that the flow instability responsible for the self-excited flow is due to the presence of a shock induced separation bubble in the corresponding steady flow. Finally the method is used to predict the aerodynamics of aerofoils performing inplane and combined inplane-pitching motions. Results show that quasi-steady aerodynamic models are unsuitable at conditions representative of high-speed forward flight. For shock free flows, the unsteady effects of freestrearn oscillations can be represented by a simple phase lag. For transonic flows the influence of unsteadiness on shock wave dynamics is shown to be complex. Calculations for indicial motion show that the unsteady behaviour of the flow is related to the finite time taken by disturbance waves to travel to the shock wave from the leading and trailing edges of the aerofoil.

An experimental investigation of the hypersonic flow over (i) a wing-body configuration, (ii) a hemi-spherically blunted cone-cylinder body and (iii) a one-half- power-law body has been conducted for M = 8.2 and Re = 9.35x104 per cm. The tests were performed at model incidences, a=0,5 and 10° for flap deflection angles, (3 = 0,5,15, and 25° for the wing-body. The incidence ranged from -3 to 10° for the cone- cylinder and -5 to 15° for the power-law body. (i) The schlieren pictures showing top and side views of the model indicate that the body nose shock does not intersect the wing throughout the range of a under investigation. Detailed pressure measurements on the lower surface of the wing and flap along with the liquid crystal pictures suggest that the body nose shock does not strike the flap surfaces either. The wing leading edge shock is found to be attached at a=0 and 5° but detached at a= 10°. The liquid crystal pictures and surface pressure measurements indicated attached flow on the lower surface of the wing and flap for 13 =0 and 5° at all values of a under test. However at a= 0°, as the flap angle is increased to 15° the flow separates ahead of the hinge line. As incidence is increased the boundary layer becomes transitional giving rise to complex separation patterns around the flap hinge line. The spherically blunted body nose causes strong entropy layer effects over the wing and the trailing edge flap. A Navier-Stokes solution indicated a thick entropy layer of approximately constant thickness all around the cylindrical section of the body at zero incidence. However, at an incidence of 10° the layer tapers and becomes thinner under the body. The surface pressure over the wing and the plateau pressure for separated flow was found to increase from the root to the tip. This is partly because of the decrease in local Reynolds number across the span, however in the present case, entropy layer effects also affected separation. The entropy layer effects were found to reduce the peak pressures obtainable on the flap. The peak pressures, over the portion of the flap unaffected by entropy layer effects, could be estimated assuming quasi two dimensional flow. (ii) Force measurements were made for the blunted cone-cylinder alone as well as with the delta wing, with trailing-edge flap, attached to it. The lift, drag, and pitching moment characteristics for the cone-cylinder agree reasonably well with the modified Newtonian theory and the N-S results. The addition of a wing to the cone-cylinder body increases the lift as weil as the drag coefficient but there is an overall increase in the lift/drag ratio. The deflection of a flap from 0° to 25° increases the lift and drag coefficients at all the incidences tested. However, the lift/drag ratio is reduced showing the affects of separation over the wing. The experimental results on the wing-body are compared with the theoretical estimates based upon two-dimensional shock-expansion theory. (iii) The lift, and drag characteristics of a one-half-power-law body are compared with other existing results. The addition of strakes to the power-law body are found to improve its aerodynamic efficiency without any significant change in its pitching moment characteristics.

No description available.

No description available.

As a parachute descends, its axis of symmetry oscillates about a vertical axis. This oscillation implies angular acceleration, with consequential linear acceleration developing. The significance of the acceleration terms in evaluating parachute performance has been long appreciated and commonly allowed for in dynamic stability analysis through introduction of appropriate apparent mass and apparent moment and inertia terms. This research, while dealing specifically with the aerodynamic performance of the parachute, considers the experimental technique to measure the total fluid resistance and apparent mass components of the parachute canopy which are related to the behaviour of any bluff body moving unsteadily through a fluid. Total fluid resistance and apparent mass components were evaluated by measuring forces and moments with strain gauges during the relative motion of parachute models submerged in water in a ship tank. While being towed by the motion of the carriage, a slider crank mechanism caused the sting-mounted canopies to be harmonically oscillated at a low frequency along any required line which was parallel to the tank bed. Results show that the apparent mass components depend on the shape of the canopy, its angle of attack and the acceleration modulus (the product of the acceleration of the canopy and its diameter divided by velocity squared) and, except at high values of the latter, can be considerably in excess of potential flow evaluations. A set of differential equations which describe the three-dimensional motion of the parachute canopy-store system during descent were developed. The non-linear equation of motions were solved numerically. The effect on dynamic stability due to the variation of system parameters was studied and appropriate stability criteria were developed. Results show that the resultant dynamic performance is highly sensitive to the chosen values of the apparent mass components.

The technique of the boundary element method consists of subdividing the boundary of the field of a function into a series of discrete elements, over which the function can vary. This technique offers important advantages over domain type solutions such as finite elements and finite differences. One of the most important features of the method is the much smaller system of equations and the considerable reduction in data required to run a program. Furthermore, the method is well-suited to problems with an infinite domain. Boundary element methods can be formulated using two different approaches called the ‘direct' and the ‘indirect' methods.