The work reported in this thesis was part of a larger project to evaluate the use thermal plasma processing to synthesise ultrafine TiCx, TiNy, and TiCxNy powders from locally available TiO2, natural gas and NH3. The objectives for the work reported in this thesis were to develop and test the necessary experimental apparatus for a process for synthesis of the ultrafine ceramic powders, and then to investigate the synthesis of ultrafine TiCxOz powders from pigment-grade TiO2 and natural gas using the process that had been developed. The apparatus developed included a powder feeder capable of feeding ultrafine cohesive powders, a RF plasma torch, and powder collection equipment. Feeding cohesive powder at low flow rates with little flow rate variability is recognised as an extremely difficult task. Perhaps as expected, the vibrated elutriated fluidised bed (VEFB) powder feeder that was developed only partially met its design specification as it fed TiO2 agglomerates to the process that were too large to vaporise and react. The poor process repeatability was attributed to the VEFB powder feeder. It was recommended that a Wright system powder feeder be developed to overcome these limitations. The RF power supply required significant modifications before being used for plasma generation, including the addition of silicon-controlled rectifiers for continuously-variable power control and a complete redesign of the grid feedback circuit. The RF plasma torch development took several iterations, and the design used for all experiments performed adequately except for a low tailpiece reactant gas injection velocity which resulted in poor mixing between reactants injected through the tailpiece and the hot flow. The tailpiece injection velocity should be increased to at least 50 m s-1 for future experiments. The powder collection apparatus used for most experiments consisted of a 785 mm long section of 78 mm ID water-cooled stainless steel tube with smaller diameter tubes inserted within the larger tube. The mass collection efficiency of this device was greater than 75%, and was considered sufficient to ensure that statistically representative samples were collected. The experimental programme for the synthesis of ultrafine TiCxOz was centred on a base case set of experimental parameter values, and investigated the process repeatability, the plasma torch power input, the nominal reaction stoichiometry, the natural gas injection location, the tailpiece injection velocity, and the effect of H2 co-injection. Variation of only one parameter in each experiment led to the identification of required process modifications and reaction system parameters for future study. A complex multiphase mixture of four particle types was produced. Ultrafine spherical single-crystal particles ranging from 7 to 25 nm in diameter were almost always observed and were homogeneously nucleated TiO2 or Ti suboxide particles that then grew by liquid droplet coalescence. Ultrafine cuboid or rhomboid single-crystal particles up to 25 nm across parallel faces were almost always observed and were homogenously nucleated TiCxOz particles with x close to 0.99 and z close to 0 that grew by a gas-solid mechanism. Unvaporised single-crystal TiO2 feed particles and largely amorphous free carbon particles ranging from 8 to 40 nm were always observed. The crystallinity of the free carbon particles increased while the extractable content decreased when the natural gas was injected upstream of the plasma. This indicated that the formation mechanism changed from an acetylenic growth mechanism with upstream injection to a polycyclic aromatic hydrocarbon mechanism with downstream injection. The low mass vaporisation efficiency and mixing patterns meant that TiCxOz was formed unless the process power input was too low. TiO2 to TiCxOz, conversion was largely unaffected by either an increase in the reaction stoichiometry or reactant injection location. Little or no conversion of TiO2 to TiCxOz occurred in experiments performed below the minimum power input, which for this process appeared to be that required to bring all streams entering the process to a bulk temperature of around 4000 K. The major enthalpy demands for this process were heating the Ar flow required for torch operation and the natural gas reactant stream. Future experiments should not use upstream injection of reactant gases so that all available enthalpy transferred from the discharge to the cooler central core can be used for particle vaporisation. The TiCxOz synthesis experiments were modelled using free energy minimisation techniques. It was found that it was important to include all inert gases in the system specification and that the thermodynamic properties of graphite should adequately represent the thermodynamic properties of the free carbon, so the prediction of the onset of free carbon formation should be relatively accurate. The system TiCx-TiNy-TiOz forms a solid solution with complete miscibility at elevated temperatures. However, the phase and atom distributions, and the temperatures at which phase transitions occurred, were relatively unaffected by the inclusion of ideal solution TiCxOz(s,l) in the system specification. It was concluded that slight differences did not warrant the added complexity caused by including the solution phases in the system specification. Analyses of the individual experiments indicated that free carbon formation would be a serious problem, and that the high vapour pressure of TiN(g) around the temperatures at which TiC(s) formation was predicted to occur may have led to the formation of TiCxNyOz(s) with significant nitrogen levels. / Whole document restricted, but available by request, use the feedback form to request access.

Galfan (5 wt% Al-Zn plus 0.05 wt% mischmetal or 0.1 wt% Mg) is a coating which in recent years has undergone many industrial studies and trials. The aim with Galfan is to produce a coating which has no intermetallic phases in the coating but this condition is not always achieved. Though it is important, the nature of the intermetallic phases and the mechanisms of their formation and growth during Galfanizing are far from fully understood. The overall objective of this investigation was to identify the intermetallic phases which form and grow during Galfanizing. At 450°C, the formation of the intermetallic phases was in the form of a local outburst and the intermetallic phases then grew in a breakaway morphology. The previously reported structure and composition of the intermetallic phases were found to be either incorrect or incomplete. The intermetallic local outburst was determined to be mainly Fe2Al5-Znx with the outer part next to the eutectic being FeAl3-Znx. The breakaway phase was FeAl3-Znx and another phase next to the interface was Fe2Al3-Znx. The intermetallic growth is characterized by the growth of Fe2Al3-Znx towards the substrate with a <001> growth direction, following a reaction path of α (substrate) - Fe2Al5-Znx - FeAl3-Znx - Galfan (melt). When the dipping temperature was increased, the rates of intermetallic formation and growth were greatly increased. The intermetallic morphology was shown to change from predominantly breakaway at 450ºC to predominantly layered at temperatures higher than 490ºC resulting from a fast growth of Fe2Al5-Znx. The degree of the preferred orientation of Fe2Al5-Znx, increased at higher dipping temperatures. It was shown that the lattice parameters and therefore the interplanar spacings of Fe2Al5-Znx formed during Galfanizing were dependent on the dipping temperatures. The same reaction path as that described for 450ºC was also followed at higher temperatures. However, when the substrate is dipped at high temperatures and when the Fe2Al5-Znx grows to a considerably thickness, the limited diffusion rate of aluminium across a thick Fe2Al5-Znx layer resulted in the formation of FeZn10-Aly. When silicon was present in the substrate, the time for the intermetallic formation was much lengthened and the growth of Fe2Al5-Znx and therefore the whole intermetallic layer became considerably slower. When silicon was present in the melt, the intermetallic phases did not grow to any noticeable extent, at 450ºC and 470ºC. However, at 490ºC and above though the nucleation time of the intermetallic was also delayed, once nucleated the growth was fast resulting in a highly localized growth. It was found that the existing data on the interplanar spacings of Fe2Al5-znx were incorrect. The two most intense peaks were shown to be due to diffraction from 221 and 311 rather than from 002 and L30. A new list of interplanar spacings was proposed. It was shown that the lattice parameters and therefore interplanar spacings of Fe2Al5-Znx formed during galvanizing were dependent on the aluminium content of the bath. The compositions of intermetallic phases determined in this study were not, consistent with those expected from previously published ternary phase diagrams. At 450ºC, the solubility of zinc was found to be 5.3 wt% in FeAl3, 20.9 wt% in Fe2Al5, and 2.0 wt% in FeAl at 450ºC. Based on these results, a new version of the Fe-Al-Zn ternary phase diagram at 450ºC was proposed.

In slope stability analysis, as with other areas of geotechnical engineering, there is a high degree of uncertainty associated with design parameters. Engineers rely heavily on historical information and present performance to establish criteria. The observational approach which is essentially taking a Bayesian stance is advocated in most projects. The purpose of this study was to develop a method which will augment current design practice for mine slope stability, with the use of strength parameters derived from the back analysis of failures. A computerised database system for the storage of slope instability cases was developed using a commercial package. This enabled consistent and complete records of failures to be stored and recalled for analysis. To utilize these back analysis results in a probabilistic framework, a Bayesian approach was formulated to calculate the probability distribution of c-φ. Two separate likelihood functions were derived from the observations of slope failures in similar materials for 1) The functional relationship between the c-φ parameters required to meet critical equilibrium conditions. 2) The average normal and shear stress conditions on the failure surface. Uniform prior distributions were used to assess the likelihood function. Strength properties from tests and other sources were collated into informative priors and the corresponding posterior distributions defined. Both likelihood functions produced results in agreement with reported test results. A probabilistic model based on the first-order second moment approach was developed to cope with circular and non-circular slope stability analyses. The spatial variability of the random variables was incorporated into this model and by considering the contributions of the end resistance a quasi 3-D probability of failure was derived. All the case studies were analysed using this model with the posterior strength parameters. It was found that the spatial variability only had a major effect when the range of influence of the variability was small compared to the size of the slope. Both watertable and model error had much greater effect on the variance. The 3-D aspects of failure were investigated and a lateral release factor was introduced to assess the importance of the end resistance on the failure probability. A predictive analysis was presented to demonstrate the use of these methods in slope design. End resistance was shown to have most impact on 3-D probability of failure.

The aim of intensity modulated radiation therapy (IMRT) is to kill tumor cells while at the same time protecting the surrounding tissue and organs from the damaging effect of radiation. To achieve these goals computerized inverse planning systems are used. Given the number of beams and beam directions, beam intensity profiles that yield the best dose distribution under consideration of clinical and physical constraints are calculated. This is called beam intensity optimization problem. In this thesis, we first review existing mathematical models and computation methods for the beam intensity optimization problem. Next, we formulate the beam intensity optimization problem as a multiobjective linear programme (MOLP) with three objectives. For clinical cases this optimization problem involves thousands of variables and tens of thousands of constraints and existing methods such as multiobjective simplex methods can not handle it. The rest of the thesis is dedicated to developing methods to solve this large MOLP efficiently and to the application in the beam intensity optimization problem. Benson (1998c) argues that solving an MOLP in objective space needs less computation time than solving it in decision space if the number of objectives of the MOLP is much smaller than the number of variables. Moreover, the constraint matrix of the problem relies on the calculation of dose deposited in tissue. Since this calculation is always imprecise solving the MOLP exactly is not necessary in practice. This motivates us to develop algorithms for solving an MOLP in objective space approximately. We summarize Benson’s outer approximation algorithm for solving MOLPs in objective space and propose some small changes to improve computational performance. Then in order to approximate the true nondominated set we propose a modification of Benson’s algorithm which is called an approximation version of Benson’s algorithm. Our approximation algorithm computes an inner and an outer approximation of the nondominated set. We prove that the inner approximation provides a set of ε-nondominated points. The geometric duality theory of Heyde and L¨ohne (2006) defines a dual to an MOLP and it assures us to be able to find the nondominated set of the primal MOLP by solving its dual MOLP. Based on this we develop a dual variant of Benson’s outer approximation algorithm to solve the dual MOLP in objective space. We prove that solving the dual provides a weight set decomposition. We compare the primal algorithm and the dual algorithm on small illustrative and on radiotherapy examples. Furthermore, we propose an algorithm to solve the dual MOLP approximately but within specified tolerance. This approximate solution set can be used to calculate an approximation of the nondominated set of the primal MOLP.We show that this set is an ε-nondominated set of the original primal MOLP and provide numerical evidence that this approach can be faster than solving the primal MOLP approximately. Considering that the set of nondominated points is infinite, it is not very useful from the planners’ point of view. We address the problem of finding well distributed nondominated points for an MOLP.We propose a method which combines the global shooting and normal boundary intersection methods. By doing so, we overcome the limitation of normal boundary intersection method that parts of the nondominated set may be missed. Discrepancy analysis of the nondominated points from a geometry point of view shows that this method produces evenly distributed nondominated points. Moreover, the coverage error and the uniformity level can be measured. Finally, we apply the algorithms developed to the beam intensity optimization problem of 3D clinical cases with voxel size of 5mm and 3mm. A technique of reducing the resolution in normal tissue has been used to reduce the computation time. The results clearly illustrate the advantages of our methods.

In this thesis the structural analysis of two basic types of bridge deck systems are discussed: 1. the nmltibeam bridge deck II. the skewed anisotropic bridge deck. The major difficulty in the analysis of I, the multibeam deck, arises from its lack of transverse bending stiffness; load distribution occurs by shear transference at interlocking shear keys. An analysis method, developed from transfer matrix theory is proposed and shown to be satisfactory for such a structure. Model studies on a quarter scale multibeam bridge deck are described together with field tests on the prototype decks - the southern motorway bridges crossing Slippery Creek. Agreement between theory, model studies and field tests is illustrated. The satisfactory analysis of II, the skewed anisotropic deck, is complicated by its anistropic elastic properties and skewed geometry. An analysis procedure is introduced which is an extension of the finite element technique already established in other plate bending and plane stress problems. Using therefore the matrix displacement method and finite element discretization, the method has been programmed for solution by digital computer. Comparison of the computed displacements with those obtained by experiment on skewed isotropic and anisotropic steel plates is given. The finite element method is seen to be a powerful analytical tool, particularly because of its ability to handle elastic anisotropy and arbitrary geometric shapes.

The transverse vibration of a bluff body in a steady fluid flow is a phenomenon that has been observed and discussed through the ages. Named after the Greek God of Wind Aeolius, the Aeolian Tones, such as are emitted by the wind in telephone or power lines, were known to the Greeks, who produced musical sounds from an Aeolian Harp by hanging it in a current of wind. The first recorded association of the transverse vibration with a periodic wake pattern was made by Leonardo Da Vinci, who observed and sketched the process of alternate periodic vortex shedding from a bluff body to form the staggered vortex trail in its wake. It is unlikely that prior to the nineteenth century many structural failures occurred due to hydrodynamic excitation. Wood, stone and brick were the main construction materials, and the elementary design methods used were very conservative ensuring that the structure had high frequency, large mass and a large damping factor. A very high flow velocity was therefore needed to initiate structural oscillations by vortex shedding and, if initiated these structural oscillations would usually be quickly damped out. During the nineteenth century the rapid advancement in the art of civil engineering design and the introduction and development of concrete and steel as a construction material led to the design of streamlined structures with more economical dimensions and consequently lower frequencies, smaller masses and smaller damping factors. Although design codes made a reasonable allowance for static loadings, dynanic loadings caused by earthquakes and hydrodynamic excitation were eirther ignored or underestimated.

Specimens of high-density low-density polyethylene were uniaxially compressed at nominally constant strain-rates ranging from 0.000066 to 0.66 sec-1 at temperatures from 77° to 350°K and hydrostatic pressures from 0.001(latm) to 4 K bar. It was shown that there were significant effects of temperature, pressure and strain rate on both polyethylenes. At small strains (O.O1), the time-temperature and time-pressure superpositions were used successfully to construct the master curves of stress versus reduced strain-rate. These master curves, which were found to have been taken in the range of linear-viscoelastic behaviour of the polymers, were employed to derive the stress-relaxation moduli and the relaxation spectra. The viscoelastic response to temperature and pressure of high-density polyethylene in the transition zone between the γ and α relaxations, and that of low-density polyethylene in the transition zone between the γ and β relaxations have been found to be controlled by the same molecular motion. This motion, an Arrhenius-type activated process, was identified as involving molecular segments of six to eight carbon atoms with an activation energy of 22 Kcal/mole and an activation volume of 110 cm3/mole.

Research in radio astronomy at the University of Aucklan School of Engineering has been in progress for a period of six years. The program was initiated primarily to provide a variety of topics for post-graduate research in the Department of Electrical Engineering, and at the same time to develop a useful radio telescope installation. Initial work was concentrated on the development of antenna arrays (Lim, 1968) and low noise receivers at a frequency of 42MHz, utilising both phase-switched and correlation interferometer principles (Irving, 1966; Yarrall, 1968; Saunders, 1968). At the beginning of 1969, the Department moved from its former country site into the centre of Auckland city. As no alternative field site was available, the telescope was also moved, but space limitations meant that the city site was inadequate for any useful observations at 42MHz. It became clear that to utilise the site more fully a higher observation frequency would be necessary, and a frequency of 200MHz was chosen, as this was the upper limit of most of the available laboratory equipment, yet still within the range of semiconductor devices. In view of the inadequacies of the site, it was obvious that the performance of the system could be significantly improved by processing the output in a digital computer.

The investigation reported in this thesis is an extension of an earlier investigation of the stress/strain behaviour of β-brass single crystals under static and dynamic loading. It was aimed at elucidating the mechanisms of deformation, using optical microscopy to examine slip habits and transmission electron microscopy to examine dislocation substructures. Marked differences were observed in the stress/strain response, slip behaviour, and the substructural rearrangements of single crystals of β-brass when widely different rates of uninterrupted compressive loading were used. Under static loading (strain rates = 10-4 sec-1) the crystals workhardened in a parabolic manner, slip being confined to either of the two (101) [111] or (211) [111] systems. Cross-slip in the statically deformed crystals was evidenced by the wavy appearance of the slip traces. Under dynamic loading (strain rates > 103 sec-1) the crystals exhibited large yield drops followed by a linear rate of work-hardening which was lower than the static work-hardening rate. Slip on the four (-101) [111], (-110) [111], (101) [-11l] and (110) [-111] systems occurred simultaneously. The traces on any one dynamically deformed specimen were both coarser and finer than those observed on specimens statically deformed. The coarse traces belonging to a given {110} <111> slip system always bounded the finer traces of another {110} <111> system. No wavy slip was observed on the dynamically deformed specimens. The substructures present in the statically deformed crystals were consistent with a single mode of slip and exhibited features indicative of high rates of hardening on the active slip planes. The important features of the statically induced substructures were as follows:- (a) There was a predominance of dislocations with Burgers vectors of a[111]. Many of these lay in non-screw high-energy orientations, consistent with the occurrence of cross-slip and the observed high hardening rates of the crystals. (b) Strong tangling developed on the slip planes as a result of a[111] edge segments interacting with other a[111] segments and with loop and dipole debris produced on the slip planes. (c) A partially developed cell structure resulted from this tangling. The dynamically induced substructures were consistent with the simultaneous operation of the four {110} <111> slip systems. It was inferred from the substructural investigations that the lower rates of hardening in the dynamically deformed crystals arose from the distribution of slip over the four systems with a consequent reduction in hardening on each individual system. The interactions between dislocations on intersecting slip planes appeared to be less important in contributing to hardening than the influence of the four systems in reducing the hardening on any one system. A series of interrupted tests indicated that behaviour under staticstatic, dynamic-dynamic and dynamic-static loading sequences was consistent with the uninterrupted static and dynamic behaviour, but that the dynamic reload stress/strain and slip behaviour, and the associated substructural changes were greatly modified by static preloading. The crystals deformed by a "static" type slip mechanism, but exhibited "dynamic" type stress/strain behaviour during the dynamic reload in this sequence.

An investigation is made into the coupled lateral-torsional response of torsionally unbalanced “shear” buildings to horizontally directed earthquake excitation. Attention is confined to analytical models that have linearly elastic, viscously damped responses. The investigation involves three separate analyses: Firstly, in a preliminary study, the earthquake response of an asymmetric single storey building model is analyzed and general expressions are derived for the location of the centre of stiffness and the orientation of the principal axes. Secondly, an analysis is made of the coupled lateral-torsional response of a partially symmetric single storey building model to a single component of earthquake excitation. A modal solution of the two equations of motion is developed and a general criterion is derived for the existence of full modal coupling. By employing the design spectrum concept, together with conservative rules for the combination of modal maxima, analytical results in dimensionless form are evaluated for an equivalent static shear and an equivalent static torque. The combination expressions are then modified to include an allowance for modal coupling before the final results are computed and tabulated. The results substantiate previous findings which have pointed to a possible link between strong modal coupling and severely coupled lateral and torsional responses. In particular, they indicate that those nominally symmetric buildings which exhibit strong modal coupling are liable to respond more strongly in torsion than has hitherto been recognised by most building codes. This effect has not in the past been quantified in analytical terms. Although the results have practical applications in design, the analysis concerns itself primarily with the determination of realistic estimates for the dimensionless response quantities and no attempt is made to derive design rules. Finally, the partially symmetric single storey model is extended to a special class of partially symmetric multistorey "shear" buildings. The importance of this final analysis derives from the similarity between the results for the single storey model and those for the continuous multistorey model.