To prevent the formation of ice on aircrafts, de-icing systems are incorporated into various structures. These can consist of multi functional composite structures. To maintain control over the production of these systems, the composite must be formed in stages. Although there is a significant body of literature describing the infusion of PEEK into fibre reinforcement, there are many inconsistencies regarding the selection of hold temperature, time and pressure. This work sought to identify a set of optimum process conditions for the infusion of PEEK into a glass fibre fabric. The key considerations were the effect of process conditions on degradation and subsequent crystallisation of PEEK. The repeatability of DSC analysis was determined through the use of thermal cycling through a range of temperatures from above melt (350°C) to above T m 0. The number of runs obtainable, from a sample of PEEK before the effects of degradation occurred, decreased when hold temperature increased. However, a distinct change in degradation rate was observed above 380°C. Results show that multiple runs of PEEK can be obtained at 380°, without the effects of degradation, while ensuring the removal of residual nuclei in the melt. The process of degradation affects the chemical structure of PEEK. Through Fourier-transform infrared spectroscopy (FTIR) of PEEK the effect of degradation can be measured in the variation in the height of peak 1653cm·1 (crystalline phase) and 1648cm·1 (amorphous phase). As the amount of nuclei available for crystallisation decreases the amorphous peak will increase in ratio in comparison to the crystalline peak.

Stress-assisted grain boundary oxidation (SAGBO) is one potential mechanism accounting for the detrimental effect of oxygen on dwell fatigue crack growth behaviour. This mechanism is assumed to weaken the grain boundaries by formation of oxide intrusion along the grain boundary ahead of the crack tip. In this thesis, finite element analyses of oxide formation along grain boundary ahead of a stationary crack have been carried out to study the stress-assisted oxidation behaviour and the corresponding evolution of stress state near the crack tip during oxide formation. A coupled stress-reaction model was established, of which the mechanical constitutive behaviour was represented by a microstructure-explicit constitutive model and the oxide formation by multicomponent diffusion and thermodynamics. Two microstructures with different particle distribution were involved in the simulations to study the effect of microstructure on the oxide formation. A criterion for oxide fracture was proposed. In combination with finite element simulations, it was used for prediction of crack growth rate during dwell period. The predicted results were in the same order of magnitude as experimental data and indicates a same trend of the influence of microstructure on crack growth rate as what was observed in literature.

The aim of the work reported in this thesis is to develop friction stir welding for superplastic titanium alloys. A number of studies have shown that friction stir welding is capable of retaining the fine-grained superplastic microstructure of the base metal and so there has been much commercial interest in combining it with superplastic forming processes. Within this programme friction stir welding was performed with a variety of process parameters and in both the conventional and stationary shoulder configuration. This meant that a number of welds were created with various rates of heat input. The elevated temperature and intense plastic deformation associated with the friction stir welding process substantially alters the base metal microstructure. As a result the influence of the process parameters on the microstructure, texture and residual stress development has been identified within this investigation. Tool wear and deformation also remains a hindrance in the commercialisation of friction stir welding titanium alloys. Wear and deformation of the tool not only changes its shape, but it can have further implications on the structural integrity of the weld. Thus the influence of the process parameters on wear and deformation of the tool have also been identified.

The efficiency of a gas turbine engine is determined, in part, by the dimensional conformity of the turbine blades to the nominal design dimensions. Doncasters Precision Castings in Droitwich UK (DPCD) is involved in the development of next generation designs and as feature complexity increases, so does the dimensional variation of the part. The particle size distribution (P.S.D) of the zircon flour used to make investment slurries was found to significantly impact process capability. P.S.D influenced the rheological properties of slurry, shell build, sintering mechanisms and as a result the annulus length of solid equiaxed tip-shrouded turbine blades. The supplier of zircon flour to DPCD was not capable of controlling the P.S.D during the milling process so a blending technique was developed to reduce the variation of the incoming raw material. Implementation of enhanced process control methods significantly improved the DPCD process. Capabilities of the key process measures; plate weight, viscosity flow time, slurry stability and fired flexural strength were increased. As well as reducing dimensional variation, shell related scrap levels were also improved as a consequence of producing a more consistent ceramic mould system.

Double oxide films (bifilms) are significant defects in light alloy castings which were reported to have detrimental effects on the reliability of the castings. The research reported here was aimed at studying how these defects develop with time. The results suggested that both O and N inside the bifilm would be consumed by reaction with the surrounding melt, and that H might be diffused into the defect. Based on the estimated reactions rates the time required for the consumption of the atmosphere inside a bifilm entrained in pure Al, Al-7wt.%Si-0.3wt.%Mg and Al-5wt.%Mg alloy melts, was determined to be 538, 1509 and 345 seconds respectively. The results also suggested the occurrence of two competing mechanisms during holding of the castings in the liquid state before solidification. The first mechanism was related to the consumption of the bifilm atmosphere, which might reduce the size of bifilms and therefore increase the Weibull moduli the UTS and the % elongation. The other mechanism was the diffusion of H into the bifilms, which would be expected to increase their sizes and reduce the moduli. This research therefore could lead to the development of new techniques by which bifilms might be deactivation in light alloy castings.

Heat resisting centrifugally cast tubes of HP micro alloy (25/35 Cr/Ni, 0.4C), are extensively used in steam reforming, which is currently the dominate technology for hydrogen generation. High pressure of the reacting gases inside the tubes generates high hoop stress in the tube wall and together with the temperature exposure of 900-1050°C causes the tubes to creep along the circumferential direction. The alloy's ability to successfully withstand the severe operating condition is highly dependent on its high temperature creep-rupture behaviour. In recent years a number of manufactures have introduced higher creep-rupture strength versions of the heat resisting HP micro alloy. As a consequence they are able to offer the centrifugally cast tubes in thinner walls. Therefore, there is a need to enhance the creep-rupture strength of the alloy produced by Doncasters Paralloy. The immediate objective of this research project is to establish the mechanisms that govern the creep process in this alloy system. Once these mechanisms are better understood further alloy development within the present alloy composition can be achieved. The advanced stages of the project specifically involve studying the influence of microstructure in relation to creep resistance as a function of compositional modifications.

In the thesis, four DLC coatings were selected: two hydrogen free DLC coatings (GCr and GCrSi) and two hydrogenated DLC coatings (DCr and DCrSi), which are the two main groups in the DLC family. Three annealing conditions (argon annealing, air annealing and HIPping/high pressure argon annealing) were designed to evaluate the stability and performance of the four DLC coatings. The microstructure of the DLC coatings was characterised by Raman spectrum, GIXRD, SEM/Due-beam SEM, EDX and TEM. Scratch tests were performed to assess the adhesion of the coatings. Both the mechanical and tribological properties of the DLC coatings were evaluated including nanohardness, coefficient of friction and wear factor. The influence of high pressure and oxidation was discussed, and the comparison of four different coatings was made.

Inertia weld process variables arc characterised using empirical relations that arc functions of the weld input parameters, allowing the variables to be predicted for any future production welds. The techniques for establishing the empirical relations can be applied to future alloy system for inertia welding by using significantly refined test matrices, reducing the development costs of new welds. Power loss in the bearings for two inertia welding machines is characterised by using hydrostatic bearing theory and several experimental techniques. This allows energy-based empirical relations for new alloy systems to be developed using sub-scale welds only, reducing the cost of implementing future production welds. Average temperature and now stress of the plasticised zone of welds is used to characterise the average shear zone thickness. Two separate models arc developed and arc found to correlate well with each other and with experimental observations. This work helps to improve the mechanistic understanding of inertia weld interfaces and can be used in computational fluid dynamics models to characterise the bond-line cleaning mechanism.

The joining of the titanium alloy Ti-6Al-4V by LFW is analysed by FE and analytic modelling, in both 2D and 3D. For the purposes of model validation and testing, experimental work is carried out using pilot-scale apparatus. Thermocouple wires were inserted in to several welds, to measure thermal values at varying locations. The sensitivity of the measured temperatures and upset rates to the critical process variables; amplitude, frequency and the applied pressure - are shown to be consistent with the predictions of the 2D and 3D modelling. The flash produced is found to be dependent upon the ratio of oscillation amplitude to applied load; when this is large a rippled morphology is produced. The 2D models replicated the flash formation behaviour well. Results for residual stress fields in three principal axes were calculated from the FE model, and compared to experimentally determined residual stress values. The modelled predictions were found to be reasonable. Further, a simple analytical model of the process is proposed, in which mechanical work is balanced against flash formation; at steady-state temperature is predicted to decrease exponentially with distance in the HAZ, and the temperature gradient in the HAZ is predicted to increase as the upset rate increases, as observed.

Two distinct techniques for “optimum” fibre spreading of 2400 Tex E-glass and 12K carbon fibre bundles were designed and developed in this research project. These techniques enabled increases in fibre bundle width of 956% and 1121% for E-glass and carbon fibre bundles respectively. These values surpass anything reported in the literature. Tensile testing was conducted on dry fibre bundles in the as-received and spread state. A reduction in the tensile failure stress and Weibull survival probability was observed with an increase in gauge length. The effect of spreading was determined to be statistically significant at a gauge length of 100 mm; however it was not significant at a gauge length of 50 mm. Composites were manufactured using as-received or spread fibre bundles and their properties analysed and compared. The composites manufactured using spread fibre bundles were significantly thinner than those manufactured using as-received fibre bundles. In hybrid composites the degree of mixing of the two reinforcing fibre types was greater than has been reported in the literature. The overall conclusion from this study is that the utilisation of spread E-glass and carbon fibre bundles can be used as a means for enhancing the apparent failure strain of carbon fibre composites.