Failure due to pitting fatigue has been investigated under controlled laboratory conditions. The investigations used both a realistic laboratory test rig using 1/2" face width gears and the geometrically simpler simulation of gears using a disc machine. The results obtained substantiated earlier work of Way and Dawson. The initiation and propagation mechanisms are generally considered to hold true. However, the gear tests showed that failure could occur much more readily than with discs and therefore the application of disc tests to gears must be viewed with caution. The results suggest a fundamental difference between the pitting behaviour of gears and discs. The second part of the thesis is of a more theoretical nature. A theory of surface contact was developed along the lines of that by Greenwood and Williamson using a surface model developed by Whitehouse and Archard. These results show that a distribution of asperity curvatures increases the probability of plastic deformation. The plasticity index has been redefined in terms of a convenient two parameter defirition of surface topography. The theory has been applied to results obtained from a typical ground surface of hardened steel; when the anisotropy, which is part of such surfaces, is taken into account it is shown that only a small proportion of the contacting asperities are plastically deformed. The limitation of this form of model is discussed and a second approach is put forward using digital techniques. Theory has been developed to enable the contact of surface profiles to be simulated in a computer and the interference areas so formed have been related to the real Hertzian deformed areas of two rough surfaces. The approach is equally applicable to run- in surfaces which are not represented by existing models. The implications of this work for future research are discussed; the need for a fuller understanding of partial and micro elastohydrodynamic lubrication by theory and experiment is stressed.

The theory of two-port three-terminal lumped linear RC networks has not yet been fully explored. In particular, there are no general 'necessary and sufficient' conditions for network synthesis, nor standard synthesis techniques for 'non-series-parallel' topologies. By using a digital computer and a technique of coefficient matching, allied with an optimisation routine, 'non-series-parallel' topologies can be synthesised as readily as 'series-parallel' topologies. Coefficient matching involves the construction of individual error functions, to measure the departure of the coefficients required from those currently achieved. The effect of modifying the method of function representation is investigated and it is demonstrated that for the best method the efficiency of the optimisation algorithm is greatly improved. The properties of closed form expressions for the normalising variable are examined. The network topology may require alteration as part of the design process. Criteria for the removal and addition of elements are discussed and their implementation in a computer program for network design without user interaction is described. The effectiveness of the approach is illustrated by the synthesis of a new 'non-series-parallel' network. Only one 'non-series-parallel' network with no 'series-parallel' equivalent has previously been published. It is demonstrated that there are many equivalent realizations with different topologies. The existence of 'non-series-parallel' networks, with no 'series- parallel' equivalents, containing as few as nine elements is verified. Analytical techniques for synthesising 'non-series-parallel.' networks must be developed before 'necessary and sufficient' conditions for all sets of admittance functions can be obtained. The properties of admittance functions with two finite poles are discussed. Several useful topological theorems are derived. Two simple 'non-series-parallel' networks are analysed, and their element values are obtained as expressions in terms of the residues at the poles of the required admittance functions. The existence of many other possible realizations is demonstrated by the use of equivalence transformations.

The primary object of the work was to develop mathematical and other methods for predicting (a) tool shapes to machine defined work shapes, and (b) work shapes machined by defined tool shapes, operating under equilibrium conditions. Existing methods were examined, the essence of which was a "Cos theta" theory. This theory, based on an analysis of machining between plane parallel electrodes normal to the direction of motion, was found to provide a useful approximate method for designing tool surfaces. However, the theory was shown to be inadequate for curved surfaces whose inclination to the direction of motion of the cathode was less than 350, particularly when used to predict work shapes. Equations relating inter-electrode gap and time, obtained from the plane parallel electrode analysis, were shown also to represent the surface produced by a flat cathode inclined to its direction of motion. Work and tool shape prediction procedures making use of these equations are described and a modification to the equations to allow for overpotential is included. The current and potential distribution in the inter-electrode gap was then studied as a field problem. Two boundary conditions at the workpiece surface were identified and analytical solutions were obtained for tool shapes to produce semi-cylindrical and hemispherical work shapes, including modifications to account for overpotential at the work surface. Solutions were also obtained by an electrolytic tank analogue apparatus and by two numerical methods. A finite difference method was used successfully to predict work shapes but was found to be unsuitable for designing tool shapes. A second method was developed specifically for this purpose using a specified work shape and its boundary conditions to build up a field solution step by step. The computations for both methods were performed by digital computer. Experimental work was undertaken to provide shapes to compare with the theoretical predictions and good agreement was obtained. An attempt was made to measure anode overpotential for various metals under actual machining conditions and readings less than 1 volt were recorded. Discussion of the merits of the various methods and their applicability to three dimensional problems, conclusions and suggestions for future studies complete the work.

This thesis relates to the use of knowledge-based systems and fuzzy set theory in electrocardiogram (ECG) interpretation and qualitative ECG simulation. The aims of the research are (1) to design a knowledge-based system for ECG interpretation (KBS-EI) based on morphological features derived from ECGs, and (2) to explore a method for knowledge-based qualitative ECG simulation (KBS-QES) and its potential application in ECG interpretation. In KBS-EI, the architecture of a general blackboard system is used. KBS-EI contains several knowledge sources, each of which has its own inference engine and knowledge base. The knowledge needed for ECG interpretation is encoded in "nearly natural language" propositions which are organized into rules and frames. Fuzzy concepts in the knowledge are expressed in linguistic terms whose meanings are represented by corresponding membership functions. The way the encoded knowledge is used for ECG interpretation is based on modified "approximate analogical reasoning". Each interpretation of a set of 12-lead ECGs is assigned a similarity measure to indicate "confidence" in the interpretation. Currently, the system is being tested using ECGs obtained from medical monographs. Features of interest are extracted by manual measurement. The results so far show that the interpretations made by the system are highly consistent with those made by medical experts. Forty-five out of forty-nine cases were interpreted in a way consistent with the medical experts. KBS-QES is a system that performs qualitative ECG simulation using vectorial analysis. Simulated ECGs in each lead are generated by studying the projections of cardiac vectors in those leads. These projections are expressed in linguistic terms whose meanings are represented by corresponding membership functions. For a particular heart condition, 12-lead ECGs can therefore be simulated. In addition, such a simulation model can also be used to aid ECG interpretation. Case studies that were carried out show that under the various heart conditions considered, simulated ECGs were generated successfully. These studies also confirm that the system can aid ECG interpretation. KBS-QES is based on a multi-level architecture. This benefits the system by facilitating easy extension, modification and other future developments.

Modern railway transportation has imposed severe work conditions on the track. Wear of rails has become an important and costly phenomenon. Recent developments in the manufacture of rail steels have refined the interlamellar spacing to produce harder and more wear resistant pearlitic steels. Despite better nominal properties shown by bainitic and martensitic steels, pearlitic steels have shown lower wear rates. The aim of this study is to explain the mechanisms for the wear performance by observing how the lamellar pearlitic microstructure adapts to the wear loading. Four pearlitic rail steels, with similar chemical composition but with different hardnesses and interlamellar spacings, have been examined. Wear tests have been performed under both pure sliding and rolling-sliding conditions, the latter designed to simulate track conditions. The worn surfaces and the plastically deformed subsurface regions have been examined by optical metallography and scanning electron microscopy. It was observed that the plastic deformation produced considerable fracturing and realignment of the hard cementite lamellae. The effect of these realignments on the surface was to present an increased area fraction of hard cementite lamellae to the surface. Thinner cementite lamellae, associated with low interlamellar spacings, were easier to blend before fracturing. A relationship between the bulk hardness (HV) and the reciprocal root of the mean true interlamellar spacing has been proposed for fully pearlitic steels. Wear rates were found to be a function of the original bulk hardness, rather than the increased hardness of the plastically deformed layers. Also, wear rates were reduced as hardness increased by reducing the interlamellar spacing. Pure sliding and rolling-sliding wear tests ranked the four steels correctly. Furthermore, qualitative comparisons between experimental wear rates and those obtained in-track trials show the same scale in reduction of wear with increased hardness.

A variational principle, and a corresponding finite-element procedure, have been developed to predict steady-state, creep crack growth rates. The cracks are assumed to propagate in a creeping material by the growth, and coalescence of micro-voids on the grain boundary, in the damage zone, ahead of the crack tip. Three mechanisms of void growth have been considered: grain-boundary diffusion, surface diffusion and power-law creep. The power-law creeping material surrounding the crack tip and damage zone has been modelled using well-established finite element methods for incompressible material; and the stiffness of the damage zone elements added to the overall stiffness matrix, to obtain a complete solution. Both stationary cracks, and by an extension of the procedure, growing cracks have been analysed. Only mode-1 type loading has been considered. Loads were applied on a circular boundary around the crack tip in accordance with the appropriate HRR field, with an amplitude given by the C* path-independent integral. Results of steady-state crack propagation rates arising from void growth by grain- boundary diffusion have been compared with published results from analytical models. These comparisons indicate that the developed technique works well, and is efficient in terms of computer resources. The transition between constrained and unconstrained void growth is particularly well defined. For the other two void growth mechanisms results are presented in a similarformat to grain-boundary diffusion, and comparisons made where possible. The relationships between crack velocity and the variables C* and damage-zone size have been established for the three void-growth processes. The methodology and numerical procedures developed have been shown to be viable, and capable of further development. This might include the consideration of more complex void-growth models, or the inclusion of the procedure within an existing finite element code.

The work presented in this thesis is aimed at investigating the interaction between wear and rolling contact fatigue, which are two of the most serious forms of deterioration caused by the wheel on rail contact stresses. Wheel-rail contact conditions were simulated by a two disc contact using the LEROS (LEicester university ROlling-Sliding wear testing machine). Investigation of the wear behaviour of BS11 pearlitic rail steel showed that steady state wear behaviour is established after a certain number of rolling-sliding cycles. Contact surface failure by ratchetting (accumulation of unidirectional plastic strain) was found to be the dominant failure mechanism during the period leading to the steady state. This mechanism was confirmed by the drop in the wear rates when the direction of rolling-sliding; i.e. strain in the surface layer, was reversed at predetermined numbers of cycles. The effect of repeated rolling direction reversals on crack morphology, propagation and rolling contact fatigue (RCF) life of BS11 rail steel was also investigated. It was established that rolling direction reversal has a beneficial effect on RCF life. A new mechanism, the "variable crack face friction mechanism", was proposed to explain this effect. Interaction between wear and RCF fatigue was investigated through rolling-sliding experiments where specimens were run dry for certain number of cycles, to induce different levels of wear damage, before the fatigue performance was investigated. It was shown that initial dry cycles above a critical number causes sudden and significant deterioration in RCF life. This deterioration has been explained in terms of the role of the accumulation unidirectional plastic strain (ratchetting) in initiating and propagating the early cracks during the dry phase. A strong correlation was found between the total ratchetting strain induced during the dry phase and the deterioration in RCF life. An empirical relationship to estimate this deterioration was concluded.

Dunlop Suspensions and Components manufacture microvon, a flexible polyurethane foam, which is used extensively in the automotive industry as spring aids for car suspension systems. The material is used because of its desirable non-linear elastic stress/stain characteristics, its quick recovery behaviour, and because it tends to produce little lateral expansion during compression. The design of such spring aid components, however, remains something of a black art, since the behaviour of the material is not fully understood. Complications arise because the mechanical properties of the material are controlled by a large number of physical, chemical and processing effects. The aim of the research has been to gain further understanding of the material, and its response under load, in order to be able to predict the compressive behaviour of the material. This has been achieved by combining microstructural observations of the deforming material with information obtained from detailed mechanical tests. There have been many attempts by researchers to describe the behaviour of cellular materials. This has been done in a variety of ways, and approaches have included developing complicated strain energy functions or utilising simple models based on repeating cell units. However, a number of difficulties are encountered when applying these material models to microvon. The microstructural observations and mechanical tests undertaken in this research have led to the development of new material models for cellular materials. Two types of model have been developed; a physical model and a series of phenomenological models. The physical model is a bi-linear relationship between stress, strain and density, in which the mechanisms of deformation are described. The phenomenological models have been developed using curve- fitting, to accurately predict the axial stress/strain and lateral/axial strain behaviour of microvon over a wide range of strain, density, temperature and strain rate.

The simultaneous stabilization of a collection of systems has received considerable attention over a number of years. The practical motivation for a solution to the simultaneous stabilization problem (SSP) stems from the stability requirements of multimode systems in practical engineering. For example, a real plant may be subjected to several modes due to the failure of sensors and nonlinear systems are often represented by a set of linear models for design purposes. To examine these problems, it is necessary to establish a simultaneous stabilization theory. This dissertation considers the problem of simultaneously stabilizing a set of linear multivariable time-invariant systems. Three methodologies are presented. The first method is based on finding new approaches to solving the strong stabilization problem (i.e. stabilization by a stable controller) which can then be used in the SSP of two plants. New sufficient conditions and algorithms are derived for the solution to this problem. The second method utilizes robust stability theory applied to a "central" plant obtained from a given set of plants. A generalized two-block L-optimization problem is formulated and solved to find the central plant. The third method utilizes the parametrization of all stabilizing controllers. Sufficient conditions for the existence of a solution are derived and in the case of two plants a formula is derived for finding a simultaneously stabilizing controller. The work advances the theory of the SSP (and the Strong Stabilization Problem) by introducing and investigating several new approaches, and deriving new sufficient conditions. The work is less successful in deriving practical algorithms for the SSP except in the second method where a reliable algorithm is given for finding a central plant on which existing robust stabilization methods can be applied. This method is illustrated by its application to helicopter control.

The current hydraulic torque converter has many drawbacks. It requires a significant amount of oil to drive the impeller, and the turbine, which may cause a loss of energy. In addition, the direct contact of the mechanical parts of the conventional hydraulic torque converter increases the possibility of parts failure and mechanical noise. This thesis proposes a new and novel Electromagnetic Torque Converter (EMTC) system, which replaces the conventional hydraulic torque converter. The new EMTC system comprises of two parts, a Differential Machine (DM) part, which is a double rotor single stator machine, and a Variable Frequency Rotary Transformer (VFRT) part, which is responsible for transferring the slip power from between the DM rotors to the DM stator. One of the purposes the VFRT is to eliminate the use of slip rings. In addition, an electronic power converter is used in the EMTC system to convert the frequency of the slip power for re-injection into the DM. Finite Element Modelling (FEM) simulations were primarily used to define the design specifications of an experimental prototype EMTC system. These specifications include the minimum cogging torque, the ripple torque, and the high electromagnetic torque of the DM. FEM simulations were also used to select the stator slot configuration and for predicting the dynamic behaviour of each part of the EMTC system, and a mathematical model was developed. Each part of the EMTC system was characterised experimentally. The complete EMTC system was then tested in both open and closed loop configurations. The FEM simulations and the mathematical model showed good agreement with the experimental results. The new EMTC system resolves many of the drawbacks of the conventional hydraulic torque converter. The outcome of this thesis is expected to provide a step change in electromagnetic converter technology.