The environmental wind tunnel at Nottingham University has been extended so that realistic mean hourly atmospheric boundary layers can be generated at sufficient scale to allow aerodynamic tests of sharp edged vehicles to be undertaken. A moving model rig owned by British Rail Research was installed perpendicular to the flow near the end of the working section. As part of this project an automatic refiring mechanism was developed allowing some 2000 transits of vehicles incorporating an internal balance and data logger to be made across the working section with a realistic mean hourly atmospheric boundary layer present. The quality of the data from the moving model rig was assessed. Moving model rig tests and static model tests of a 1/50th scale lorry and 1/45th railway container vehicles have been conducted and extreme value forces and moments relevant to the gust time that overturn a vehicle were calculated. These are the first measurements to have been made using a realistic mean hourly ABL and modelling the vehicle's movement. This thesis assesses the usefulness of the normalised extreme force parameter in determining the extreme forces that a full scale moving vehicle experiences. It was found that the normalised extreme force parameter remains invariant with model time scale for the range of times considered. Further for both the moving model rig tests and the static tests the value of unity that this parameter takes for yaw angles above 30 degrees implies quasi steady behaviour without additional body induced unsteadiness. At lower yaw angles, however, some body induced unsteadiness is evident. These conclusions are compared with predictions from existing numerical models and previous experimental tests. The measured lift force from the static tests compared with the moving model rig tests at 90 degrees yaw angle, i. e. with the moving model stationary, shows a large difference. This is not understood and two concerns are expressed: the effect of the slot, through which the supports of the moving model travel, beneath the vehicle, may be altering the pressure in this region; or it could be due to a Reynolds number effect caused by the small underbody height above the ground.

Following a description of the general properties of asphaltic materials, a review is presented on fatigue damage. Fatigue element tests have been carried out using two types of procedures; uniaxial tension-compression and a trapezoidal cantilever beam. The data from the Trapezoidal test has been used to develop relationships between dissipated energy and the number of load cycles to crack initiation. A method has been developed which enables the stiffness loss during a fatigue test to be quantified in terms of the initial mixture rheology. In addition, an improved method for defining the crack initiation point, N1, has been developed along with the definition of an energy ratio to enable determination of fatigue life for intermediate modes of loading. An assessment of two tests involving indirect tension has been made. A series of tests were conducted in the Slab Test Facility to determine the performance of various asphaltic mixtures with respect to fatigue. These have been used to validate the results from the element tests and assess the suitability of different shift factors. A 2-dimensional Finite Element visco-elastic analysis method has been used to calculate dissipated energy in pavement structures. This method has been compared to an elastic analysis method. It was observed that the F.E. method is less sensitive to pavement thickness. The F.E. method has some potential for prediction of surface cracking and fatigue life but further work is needed to implement a 3 dimensional model. Finally, based upon an assessment of the results obtained, recommendations have been made for additional work involving materials testing, model development and pavement design.

This thesis investigates the use of Discrete Element Modelling (DEM) to simulate the behaviour of a highly idealised bituminous mixture under uniaxial and triaxial compressive creep tests. The idealised mixture comprises single-sized spherical (sand-sized) particles mixed with bitumen and was chosen so that the packing characteristics are known (dense random packing) and the behaviour of the mixture will e dominated by the bitumen and complex aggregate interlock effects will be minimised. In this type of approach the effect of the bitumen is represented as shear and normal contact stiffnesses. A numerical sample preparation procedure has been developed to ensure that the final specimen is isotropic and has the correct volumetrics. Elastic contact properties have been used to investigate the effect of the shear and normal contact stiffnesses on bulk material properties. The bulk modulus was found to be linearly dependent on the normal contact stiffness and independent of the shear contact stiffness. Poisson's ratio was found to be dependent on only the ratio of the shear contact stiffness to the normal contact stiffness. An elastic contact has been assumed for the compressive normal contact stiffness and a viscoelastic contact for shear and tensile normal contact stiffness to represent the contact behaviour in idealised mixture. The idealised mixture is found to dilate when the ratio of compressive to tensile contact stiffness increases as a function of loading time. Uniaxial and triaxial viscoelastic simulations have been performed to investigate the effect of stress ratio on the rate of dilation with shear strain for the sand asphalt. The numerical results have been validated with experimental data. The geometric factors that influence asphalt dilation are investigated. The level of dilation was found to be dominated by the proportion of frictional contacts in the sample. Simulations have been performed to investigate the effect of particle shape on asphalt dilation. Greater dilation was found in the sample with clumps under loading.

As long as there is a certain degree of flexibility i.e. relative motion, the contacting surface of a mechanical component will experience fretting damage, in the form of fretting wear and/or fretting fatigue, depending on the loading conditions and performance requirements. In the present investigations, the fretting wear behaviour of aeroengine materials were studied by both experimental (mainly gross slip regime) and finite element approaches involving the various aspects of fretting variables (applied normal load, strokes, wear duration, material hardness, lubrication and etc. ) that affects the material tribosystem. The aeroengine materials are the Super S/CMV and AerMet®l00 ultrahigh strength steel alloys and Inconel 718 nickel-base superalloys. The present works highlights a number of issues important to fretting damage, which involves studying: (1) the effect of nitriding and lubrication on Super S/CMV steel experimentally and wear data validation and prediction by FEM based on a modified Archard's equation; (2) the effect of moderate temperature on the fretting behaviour of Inconel 718 alloy; (3) the different material combinations (Super S/CMV, AerMet®100, and Inconel 718) on fretting wear. The following experimental tools are utilised to investigate the morphological changes of the contacting surfaces during fretting. A simple crossed round-on-flat arrangement was used to determine the coefficients of friction and the wear coefficients applicable to the contact configuration and loading conditions. Surface topography changes were investigated by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The presence of various phases and elements were analysed by x-ray diffraction (XRD) and energy dispersive x-ray (EDX). The wear depth and wear width were profiled using a surface profilometer for the wear coefficient calculation. The use of FE simulation was to validate and predict the wear data. The evolution of contact geometry and contact variables, including contact pressure and relative slip were studied. The implications of these predicted results are discussed with respect to fretting fatigue prediction, leading to new insight into the experimentally-observed effects of slip regime on crack initiation.

The creep behaviour of welds in service-aged pipes are studied. The aims of the research have been achieved using analytical, numerical and experimental approaches to the relevant subjects. Several features of the work are presented: (i) a systematic parametric study of the creep of two-material test specimens including a stress singularity analysis, (ii) an impression creep testing method using a rectangular indenter, which can be applied to study the creep properties in welds, and (iii) methods used for damage constitutive equation generation involving FE-damage modelling of the rupture tests of cross-weld specimens. General observations on the creep of two-material structures have been made using analytical solutions derived from four simple structures. The effects of geometries and relative creep properties on the creep stress or strain-rate distributions have been investigated using the finite element approach with idealised, two-material, axisymmetric models. The stress variations have been extensively studied on the centre-line and in the singularity regions of the models. An impression creep testing technique with a "long" rectangular indenter has been developed for the study of the creep properties in the narrow zones within weldments. The effect of varying the geometric test parameters has been fully analysed. The application and the possible advantages of the technique for determining the creep properties in welded components have been highlighted. Experimental testing has been performed for the purpose of verifying the impression creep testing technique and to provide the creep and rupture information required for the generation of creep and damage parameters for the exposed materials. The results obtained have further validated the applicability of the impression creep testing technique. Creep continuum damage investigations have been performed in order to generate the material parameters in a continuum damage constitutive equation and to model practical weldments. The methods used for material data generation have been established. The failure analysis and prediction for the service-aged welded pipes using steady-state analysis, damage modelling and extrapolation technique have been addressed.

This thesis investigates the properties of backscattered polarized light from layered scattering media with a view to application in the imaging of in-vivo skin for medical application. The research includes investigation of numerically simulated samples, tissue phantoms and in-vivo tissue. The aim of the early research is to identify the differences in behaviour between initially linearly and circularly polarized illumination concerning the rates of depolarization with scattering. Initial examination is made through Monte Carlo simulations. The analysis yields results which indicate, for forward scattering media, circular polarizations maintain their initial state to greater depths within a scattering medium than linearly polarized light. This result is exploited to show sensitivity of the different polarizations to different layers within a medium and indicates the potential to achieve coarse optical sectioning. These fundamental properties are extended to a full field imaging arrangement, using both simulated and experimental results to illustrate polarization gating to perform subsurface object imaging in a medium composed of uniform scatterers. The concepts are extended to imaging an in-vivo sample. The removal of multiple scatter and surface reflections is performed using a combination of linearly and circularly polarized illumination coupled with image subtraction, to provide a sub-surface, localised tissue image. This provides an improvement on currently applied techniques, which use linearly polarized light and the application of a flat glass plate and matching fluid. To exploit the spectral properties of tissue, the development and construction of a fully automated, multiple wavelength, polarization imaging system, suitable for testing in a clinical setting is presented. Theoretical investigation into the potential to perform the determination of the optical properties of a two layered sample, using simulated results in-keeping with the system's detection regime, is performed. It is illustrated that the variation of top layer thickness, bottom layer scattering, top and bottom layer absorption coefficient provides well-conditioned data when combining polarization and spectral information.

The use of glued laminated timber (glulam) in building and construction provides a designer with an attractive, environmentally friendly alternative to steel and concrete. However, along with other engineering materials, glulam is subject to creep and in order to use it efficiently without the risk of unacceptable long-term deflections occurring, a reliable method of estimating creep in glulam is desirable. With this mind, the Structural Timber Research Unit (STRU) at the University of Brighton embarked, in 1988, on an extensive glulam research programme. Since then, the deflection of prepared glulam specimens has been measured in the laboratory under controlled conditions and in 1992, the programme was broadened to include measurements to determine the behaviour of glulam in full-scale structures. A number of test sites have been established in a variety of buildings and the method chosen to monitor glulam beams in these is theodolite intersection using a Leica Electronic Coordinate Determination System (ECDS3). Borrowed from industrial and engineering surveying, the use of the theodolite intersection technique to monitor glulam is unique and it has, despite the practical difficulties and size of structure involved, enabled the movement of beams to be monitored with an accuracy of 0.1 mm. By processing three-dimensional intersection coordinates, vertical deflections and creep have been determined and results show that the creep response of the glulam beams monitored in full-scale structures can be represented by a seasonally modulated exponential function. This correlates well to results from the laboratory tests and demonstrates that measurements taken in the laboratory can predict the behaviour of glulam in-situ. The creep factors obtained also agree well with the values given for these in Eurocode 5: Design of Timber Structures.

This thesis presents an advanced quadratic formulation of the boundary element (BE) method for two-dimensional elasto-plastic analysis in which 3-node isoparametric quadratic elements are used to model the boundary and 8-node isoparametric quadrilateral quadratic elements are used to model the interior domain. The main objectives of the research are to present a comprehensive review of the many different BE approaches in elasto-plasticity to investigate the potential accuracy, robustness and reliability of each approach, and to implement the favoured approach in a comprehensive computer program for use by engineers. Full details of the elasto-plastic analytical formulations and numerical implementations are presented without ambiguity or omission of details. A brief review of the basic principles of plasticity is presented followed by the expressions for elasto-plastic flow rules and the numerical implementations which treat mixed hardening material behaviour. The analytical BE formulation in linear elastic applications is presented. Full details of its expansion to elasto-plastic problems are shown. Two main BE approaches in elasto-plasticity are presented in detail in this work; the initial strain displacement gradient approach with its compulsory modelling of the partial or full interior domain, and the particular integral approach which can be applied exclusively to the surface avoiding any modelling of the interior. It was decided that the initial strain displacement gradient approach is more robust than the particular integral approach and is more likely to be favoured by an inexperienced user of a BE program, despite its main disadvantage of interior modelling. The initial strain displacement gradient formulation as well as other alternative formulations are presented. The values of stress and strain rates at interior points are calculated via the numerical differentiation of the displacement rates in an element-wise manner; an approach similar to that used in Finite Element (FE) formulations. Full details of the numerical implementation algorithm which uses incremental and/or iterative procedures are presented. The details of the particular integral approach which circumvents the strongly singular integrals arising in domain integrals are also discussed in detail. A computer program for the particular integral approach was written, but, due to the constraints of time and the added complexity of this approach, it was not possible to fully test the program on practical elasto-plastic cases within this project. A full computer program, in Fortran, based on the initial strain displacement gradient elasto-plastic BE formulation is written and applied to several practical test problems. The program is written with emphasis on clarity at the expense of efficiency in order to provide a foundation for extension to three-dimensional applications and more complex plastic behaviour. The BE solutions are compared with the corresponding FE solutions provided by the commercially available FE package, ABAQUS, and, where appropriate, exact analytical solutions. The BE solutions are shown to be in very good agreement with other analytical and numerical solutions. It is concluded that the numerical differentiation of displacement rates in an element-wise manner is an accurate and numerically efficient technique which enables the strongly singular integrals to be performed.

The aim of this project is to consider the effect of damage on the energy absorption potential of continuous filament random mat (CoFRM) E-glass / polyester composite tubes. Composite materials have been shown to absorb significantly higher specific energy levels than metals under axial crushing conditions. This property can be exploited in automotive crashworthiness applications. Replacing steel crash structures with composites can lead to significant weight reductions. However, damage in composite structures can be difficult to assess and may not be visible by casual inspection. There is a concern that damage may accumulate in the crash structures, as a result of in-service wear and tear or due to operator negligence. It is important to understand how much accidental damage the crash structures can sustain before they are no longer able to fulfil their requirements. Two wall thicknesses of circular and square tube geometries were tested, with over 650 samples crushed either quasi-statically at 5mm/min or dynamically at 5m/s. Damage was induced in three ways: drilled holes, delamination in the form of Melinex® inserts moulded into the samples, and out-of-plane impact damage of various energy levels. Cylindrical samples made from this low cost composite are able to absorb up to 87 kJ/kg when tested quasi-statically. Dynamic testing was carried out as it provides a better representation of the loading conditions the parts will see in operation. Dynamically tested samples absorbed less energy than the quasi-statically tested samples (up to 18%). This was due to the viscoelastic nature of the matrix causing a greater degree of fragmentation at the higher test speed, leaving the load bearing fibres less well constrained and therefore reducing their load bearing capability. Square section tubes absorb less energy (up to 31 %) than a circular section of the same cross sectional area and fibre volume fraction. This is due to geometric stress raisers at the comers causing intralaminar failure. This splitting at the comers leaves the fronds less constrained and allows them to splay at a lower load. A threshold level was found for each type of damage. Below the threshold level the damage zone had no effect on the progressive failure mode or the specific energy absorption (SEA). Above the threshold level unstable compressive failure occurs in the form of a crack initiating at the damage zone and then propagating around the tube. In this situation a portion of the tube breaks off uncrushed and therefore reduces the energy absorption capability of the structure. For this material tested, relatively small hole sizes (5mm) and relatively low impact energy levels (l.5J - 3J) can cause unstable failure to occur at quasi-static test speeds. However, it has been shown that the damage tolerance of the material increases (to 10mm and 3J - 9J) at higher test rates (5m/s). Having observed the failure modes and damage tolerance of the tubes under various testing parameters it was important to look at ways of improving the damage tolerance of the samples. Moulding a thermoplastic interleaf into the sample to increase the interlaminar fracture toughness increases the damage tolerance of the tubes. Increasing the wall thickness and adding an interleaf increases the damage tolerance by up to a factor of 9. However, the increased damage tolerance of samples with interleaf was offset by a reduction in SEA by up to 48% due to a reduction in coefficient of friction in the crush zone from 0.36 to 0.22. The ultimate compressive stress (UCS) increases at dynamic test speeds and the mean crush load observed decreases. Therefore the crushing stress of the dynamically tested samples is a far lower percentage of the UCS of the material than under quasi-static loading. A greater stress concentration is therefore required to cause unstable failure at higher rates. Improved damage tolerance is also seen by increasing the wall thickness of the sample, testing square rather than circular section samples, and moulding interleaves into the samples. In all of these cases the changes that lead to improved damage tolerance lead to a reduction in the crush load of samples as a percentage of the ultimate crush load. Understanding the work in this thesis will enable the design of damage tolerant composite crash structures for the automotive industry. Such a part will, even with the inclusion of accidental damage, be able to absorb the energy required in the event of a collision.

This thesis describes work undertaken to investigate the effects of damage on the energy absorption potential of composite tubes. Tubes of various geometries and manufactured from either continuous filament random mat (CoFRM) or glass braid and polyester resin were subjected to various types of damage before testing. Damage types consisted of drilled holes, to simulate the use of drilling components for the need of assembly, impacts, to simulate damage that may occur through tool drops or items being kicked up during use and PET inserts to simulate delamination. Large glass CoFRM/polyester tubes with an outer diameter of 89.1mm and varying wall thicknesses were crushed quasi-statically at a speed of 5mmlmin. Small CoFRM and braided glass/polyester tubes with an outer diameter of 38.1mm and a 2mm wall thickness were tested quasi statically and dynamically at a speed of 5m1s. Tubes were tested undamaged and containing various sizes of holes, simulated delamination and impacts. Specific energy absorptions (SEA) and failure modes were compared. Threshold values of damage size have been found for each tube and test type, above which unstable failures and subsequent unpredictable reductions in energy absorptions occur. The small CoFRM tubes showed a decrease in SEA as the test rate increased and this was attributed to the rate dependency of the resin, causing greater fragmentation allowing fibres to bend more easily and without fracturing. The braided small tubes showed an increase in SEA as the test rate increased due to a change in the mode of failure attributed to a higher compressive strength at the increased rate. Relatively small hole sizes and impacts, of 5mm and 1.5J-3J, were seen to reduce the energy absorption of the materials tested at quasi-static test speeds. However, an increase in damage tolerance was identified as test rate increased and this was attributed to an increase in compressive strength and fracture toughness, and reduction in crush load, as the speed of test increased.