The generation of electricity to meet an ever-growing demand has become a defining characteristic of the modern world for both developed and developing nations alike. This, coupled with the intensifying concern with pollution and its effects on the environment has put immense pressure on how quickly and efficiently power is produced. Being the most prevalent source of electricity generation, coal fired power plants have been subject to increasing scrutiny and study in an effort to improve the efficiency at which they operate. Hence, coal fired power plants are being run at increased temperatures and pressures such as those observed in Super-critical and Ultra-super-critical plants. This has by extension put excessive demand on materials used in these plants specifically within the boiler and superheater pipe sections where the most extreme thermodynamic conditions are experienced. The most commonly used materials for these applications are in the family of ferritic/martensitic 9-12% Cr steels chosen for their superior material properties especially during long-term exposure as coal fired power plants typically operate for over 20 years before being decommissioned. One of the lesser understood aspects of 9-12%Cr steels is with regard to their long-term material properties specifically that of creep degradation and deformation. This has been partially due to the reliance of creep life predictions in the past being based on accelerated creep testing and empirically based modelling. With the relatively recent revelations of empirically based modelling shown to be inaccurate when extrapolated to the long-term, a need has been identified amongst researchers to develop more accurate models based on physical relationships and material microstructure. Moreover, the insight obtained from modern experimental techniques and technologies as well as ever-expanding computing capabilities provide an opportunity to produce microstructurally based models with a high degree of complexity. Thus motivated, the focus of this dissertation was to develop a physically based dislocation creep model using the Continuum Damage Mechanics (CDM) approach. A dislocation CDM model was developed and implemented in the current work for uniaxial creep loading using the numerical modelling software Matlabᵀᴹ. The CDM approach was built upon fundamental dislocation theory as well as other microstructural considerations pertaining to dislocation creep including subgrain coarsening, M₂₃C₆ precipitate coarsening and stress redistribution. The CDM model was found to require calibration in order to be applied to specific 9- 12% Cr steels which was implemented using a parameter optimisation routine. The results obtained were compared with experimentally obtained, long-term creep-time and microstructural data for the 11% Cr steel CB8 and the 9% Cr steel P92. The CDM creep-time predictions were found to vary in accuracy depending upon the experimental data against which the model was calibrated. Upon further investigation, it was hypothesised that the discrepancy observed was due to the formation of the Modified Z-phase in some of the long term creep data but not in others which was based primarily on the differing creep exposure times of the various samples. The CDM creep-time predictions for P92 were found to be accurate when compared with experimental results regardless of creep exposure times. The apparent difference in the approximation of the creep deformation for the two steels was concluded as being due to the formation of the Modified Z-phase in CB8 but not in P92 as Modified Zphase formation is intrinsically linked with the Cr content of the steel.

Solid particle erosion tests of glass, stainless steel, WC-Co and sintered alumina, have been performed with seven erodents using a range of particle diameters D (63 μm - 1000 μm), velocities V (33 m.s⁻¹ - 131 m.s⁻¹ ) and impact angles α (30° - 90°). The seven erodents are steel shot, glass beads, silica, alumina, tungsten carbide, silicon carbide and diamond· particles. In addition, the target materials have been subjected to cavitation erosion using a conventional ultrasonic horn in distilled water. Systematic studies of the influence of the impact variables on the erosion rate have been made. Scanning electron microscopy of the eroded surfaces and the erodents after impact has been performed. Empirical correlations between erosion rate and the parameters of erosion and the erodents were obtained and are discussed in terms of the modes and mechanisms of erosion. A semi-quantitative theoretical model has been developed to explain the empirical correlations for brittle and ductile materials. The mode of erosion of glass impacted by irregularly shaped particles is associated with the formation and interaction of lateral cracks over all impact velocities and angles used in this study. The erosion of glass by spherical particles is determined by particle size, impingement velocity and angle. An erosion map, in which the erosion of glass is mapped against velocity and particle size, has been constructed to categorise the types of damage observed in glass for impingement angles between 90° and 30°. The erosion. behaviour of 304 stainless steel is associated with cutting or ploughing and plastic accumulation processes. The erosion of WC-Co is associated with a combination of ductile and brittle modes of erosion. The erosion of alumina is brittle and associated with intergranular spallation and grain-crushing. An analysis of the results reveals that for the brittle materials, glass and alumina, the erosion rate is determined by kinetic energy, particle size and the relative hardness and toughness of the erodents. However, for ductile materials, the shape and kinetic energy of erodents are the most important factors determining the erosion rate. There is no significant effect of hardness and toughness of erodents on erosion. Surprisingly, the erosion resistance of the softer 304 stainless steel is better than that of alumina and WC-Co when hard erodents are used at impact angle greater than 40°. On the other hand the erosion resistance of the harder WC-Co and alumina is better than that of 304 stainless steel for softer erodents like silica erodents. Glass always exhibits poor erosion resistance. In cavitation erosion, stainless steel exhibits better cavitation erosion resistance than glass, alumina and WC-Co. The cavitation erosion resistance of WC-Co is dependent upon the cobalt content. An attempt to rationalise the results in terms of mechanisms has been made. Both solid particle and cavitation erosion rate for the as received glass is higher than that for the tempered glass due to introduction of residual compressive stresses into the surface by the tempering process. Particularly, it reveals that compressive stresses are more efficient in preventing the formation and propagation of Hertzian cracks. These findings will assist in the choice and design of materials that undergo both particle and cavitation erosion under specified conditions.

Uniaxial tensile testing has been undertaken on a range of aluminium base alloys. Material investigated included commercial binary Al-Mg (5182), ternary Al-Mg-Si (6061) and quaternary Al-Cu-Mn-Si (2014) as well as experimental alloys containing 2at.% additions of Ag, Mg and Zn to commercially pure AI (1070). In addition, composite materials based on both alloys 2014 and 6061, containing 10%, 15% and 20% additions of Ah03 particulate, as well as 20% SiC particulate in the case of 6061, were also tested. Microstructures of materials were varied by prior heat treatments but, for comparison, all materials, were initially tested in the solution treated and quenched condition. Mechanical testing was undertaken at room temperature throughout the course of the work, and at strain rates such that serrated tensile test curves were manifest. The evolution of microstructural features of the deformation was evaluated utilising both optical and electron microscopy. Surface deformation features, including the formation of both type A and type B deformation markings, was examined on pre-polished specimen gauge lengths at various levels of tensile strain. The planarity of slip line traces was correlated with the evolution of related deformation structures in dynamic experiments in a high voltage transmission electron microscope (HVEM). In addition, the formation of slip lines on the surface of the HVEM microtensile specimens compared favourably with those formed on the surfaces of macroscopic tensile specimens. Microscale heterogeneities in the deformation observed during in-situ dynamic HVEM experiments on poly crystalline material correlated with the extent of serrated flow manifest in bulk specimens. All materials deformed in the HVEM displayed inhomogeneous dislocation motion consistent with the macroscopically observable discontinuities. The alloys tested were microstructurally distinguishable during dynamic experiments depending primarily on whether or not they had been deliberately alloyed with magnesium. The alloys containing Mg exhibited the activation of parallel slip traces together with minimal cross-slip in any single micro-yield event. In contrast, the alloys which did not contain Mg exhibited the simultaneous activation of various intersecting slip systems and were characterised by extensive cross-slip during similar yield events. On the basis of these observations, the magnitude of serrations and extent of serrated flow in the alloys has been discussed. The extent to which the different alloys were able to undergo dynamic recovery affected both the evolution of the dislocation structure observed in the conventional transmission electron microscope ( CTEM) as well as the final fracture mode. The existence of a characteristic shear fracture mode was consistently observed to follow tensile deformation which had been dominated by unstable plastic flow. The ready occurrence of dynamic recovery and the associated formation of dislocation cell structures allowed for more fully developed plastic instability during the final stages of tensile deformation and a lower likelihood of final failure by premature shear. Finally, the addition of particulate reinforcement to 2014 and 6061 had different effects that were accounted for by the difference in strength between the two monolithic materials. In the case of the weaker 6061, all particulate additions had a strengthening effect whereas in 2014, increasing the volume percent of reinforcement progressively weakened the composite. Serrated flow properties of both alloys were affected by the addition of the particulate reinforcement. The homogeneity of particle distribution as well as the size of the particulate inclusions affected both the tensile properties and final fracture of the composites.

Stress corrosion cracking (SCC) is a corrosion phenomenon which continues to plague the power generating industry especially in low pressure (LP) steam turbine blades operating in the phase transition zone. An investigation has therefore been conducted to examine the effect of heat treatment condition on the microstructure, mechanical properties and SCC properties of one such LP turbine blade material, FV520B, used in the steam turbines of coal-fired power stations in South Africa. The three stage heat treatment cycle of the FV520B turbine blades consists of homogenisation at 1020°C for 30 minutes, solution treatment at 790°C for two hours and precipitation hardening at 545°C for six hours. In this study, the precipitation hardening temperature was varied in the range 430-600°C to investigate how this variation would affect the material and SCC properties. Hardness and tensile testing were performed to obtain mechanical properties while the investigative techniques used to characterise the microstructures were light microscopy, dilatometry, X-Ray Diffraction (XRD), Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM). Stress corrosion susceptibility for the different heat treatment conditions was quantified using U-bend specimens while crack growth rates and threshold stress intensities for SCC (KISCC) were measured using fatigue precracked wedge open loaded (WOL) specimens. Both SCC tests were conducted in a 3.5% NaCl environment maintained at 90°C. XRD results revealed the presence of reverted austenite in the higher tempered specimens due to the precipitation hardening temperature being close to the Ac1 temperature for the material. The presence of reverted austenite was shown to adversely affect mechanical strength and hardness which decreased with increasing precipitation hardening temperature. Light and electron microscopy (SEM and TEM) revealed the presence of Cr-rich precipitates along the prior austenite grain boundaries in all tested heat treatment conditions. The propensity, quantity and size of the Cr-rich precipitates increased as the specimen temper temperature increased. SCC susceptibility was shown to be dependent upon yield strength and decreased as precipitation hardening temperature increased with specimens in the overaged condition showing no cracking after more than 5000 hours in the test environment. WOL testing only produced cracking in the three highest strength specimens after 2000 hours. Crack growth rates and threshold stress intensities were found to be dependent on yield strength and decreased with increasing precipitation hardening temperature. Analysis of fracture surfaces revealed crack propagation along prior austenite grain boundaries in all test heat treatment conditions indicating intergranular stress corrosion cracking (IGSCC) as the dominant cracking mechanism.

The nickel-base superalloy MarM-002 is a high strength precipitation hardening material used in structural applications in the gas turbine field. The microstructure of MarM-002 consists of fine y' precipitates, a y matrix, carbides and a eutectic y- y' where the y' can be composed of coarse lamellae or blocky y' precipitates. Increasing the volume fraction of fine y' by dissolving the coarse eutectic y' during solution treatment can raise the alloy strength. In practice the solution heat treatment temperature does not usually exceed 1220°C because of the danger of incipient melting. At 1220°c the eutectic y' does not dissolve and persists in the alloy structure. In the current project an alternative solution treatment technique, the varied rate solution heat treatment (VRSHT), was determined specifically for MarM-002, and was used to achieve higher solution treatment temperatures up to 1260°C without incipient melting. The microstructural response of MarM-002 to the conventional 1220°C solution heat treatment, and to solution treatment at temperatures above 1220°C, was studied extensively. Microstructural features such as carbide decomposition and the degree of y' and eutectic y' dissolution were noted. The tantalum and titanium rich carbides which form during casting were seen to decompose during solution treatment, while hafnium rich carbides precipitated in the interdendritic regions during heat treatment. The y' solvus and eutectic y' solvus was measured to be1260°C and 1280°C respectively. A 1050°C/12h +_ 870°C/16h ageing heat treatment was applied to the alloy following solution treatment and the effect on the y' characteristics was also studied.

Ti-6Al-4V is a two phase alloy used in the aerospace, military and biomedical industries. The thermomechanical processing (TMP) of Ti-6A-4V involves the breakdown of the ingot microstructure at temperatures above the beta transus (Tβ) into fine lamellar colonies. Subsequent hot deformation is carried out at temperatures within the two phase region to produce a microstructure that is either equiaxed or bimodal. However, to avoid the formation of shear bands, voids or cracks within the final microstructure, this secondary fabrication is carried out at slow speeds, and these render the TMP of titanium a time consuming and expensive process. This project aims to investigate the possibility of using a steckel mill for the TMP of Ti6Al4V. A steckel mill is a single stand reversing mill with 2 coiler furnaces on either side. These furnaces allow for the annealing of strip metal in between passes thus keeping the strip at a high temperature throughout processing. In this way, reversing passes can occur indefinitely until a desired gauge thickness is achieved. A steckel mill therefore represents a cheaper and faster method to produce metal sheet or plate. Ti-6Al-4V cylinders with a diameter of 10mm and length of 15mm were uniaxially compressed in the Gleeble 3800. Hot compression was carried out isothermally in a 3 pass schedule at temperatures of 850°C, 950°C and 1050°C. The strain was kept constant at 0.48 and strain rates of 1/s and 10/s employed. Interpass times corresponding to a strain rate of 10/s were 44s and 77s, whilst those corresponding to a strain rate of 1/s were 324s and 712s. Statistical analysis in the form of analysis of variance (ANOVA) was used to determine the parameter most influential on the microstructural evolution of Ti-6Al-4V and the Taguchi method used to identify the optimum parameters suitable for the TMP of Ti-6Al-4V using a steckel mill. Three successive passes at 850°C, 1/s resulted in a microstructure consisting of coarse, deformed grains and some finer recrystallised grains. The influence of a low strain rate was such that it promoted recrystallisation at this temperature, while that of interpass time brought about recrystallisation and grain growth. At the same temperature and a strain rate of 10/s, less recrystallisation, together with a heavily deformed microstructure was observed. This was due to the heterogeneous distribution of strain which was a consequence of the high strain rate used. At 950°C, at both 1/s and 10/s, 1 pass resulted in a bimodal microstructure. With subsequent passes, the amount of equiaxed alpha was observed to increase. This increase was a result of a strain induced phase transformation (SIT) from beta to alpha at high temperatures. The extent of this transformation increased with an increase in strain rate. Therefore, after 3 consecutive passes at 10/s, a fully alpha (though heavily deformed) microstructure was formed. A subsequent post deformation heat treatment would lead to recrystallisation of these grains and a microstructure consisting of refined equiaxed grains the result. After 1 pass at 1050°C, at either 1/s or 10/s, a Widmanstätten microstructure was formed. However, after 3 consecutive passes at 1/s, the microstructure remained mostly Widmanstätten whilst at 10/s, a bimodal microstructure was formed. The combination of a high strain rate, low interpass times, sequential strain imparted on the sample as well as the high temperature at which the compressions were carried out, elevated the extent to which a strain induced phase transformation from alpha to beta proceeded. Statistical analysis using ANOVA and the Taguchi method revealed that a schedule of 3 passes performed at 1050°C, 1/s and the corresponding interpass times as being the optimum parameters for the TMP of Ti-6Al-4V during steckel mill rolling. Analysis of the microstructural evolution across all 3 temperatures, however, showed that 3 passes carried at 950°C, 10/s, with interpass times of 44s and 77s, as being the optimum parameters. Steckel mill rolling of Ti-6Al-4V has thus been confirmed as a feasible process for the production of Ti-6Al-4V sheet material.

Includes bibliographical references. / This dissertation describes an investigation into the application of fracture mechanics to brittle materials, with particular emphasis on the fracture toughness testing techniques used on these materials.

Includes bibliographical references. / This work is an investigation into the phenomenon of three-body abrasive wear. A specially designed three body abrasive wear apparatus has been built, modified and evaluated as part of this overall study. Further, a series of commercially available candidate materials has been evaluated for wear resistance using silica sand as the abrasive on this purpose made rig. The effect of normal load, abrasive particle size, abrasive feed rate and the type of abrasive on three body wear resistance has also been examined. It has been shown that there is little increase in wear with an increase in particle size in the size range from 50µm to 180µm and that above an abrasive particle size of approximately 200µm there is a sharp decrease in the wear with increasing particle size, followed by a levelling off in the wear. The wear was found to increase linearly with increasing load. Varying the abrasive feed rate showed that at lower feed rates the abrasive particles were more efficient at removing materials, so the wear was higher than at higher abrasive feed rates. It has also been shown that although the use of ash from coal-fired power stations as an abrasive produces wear of materials, the volume losses were much smaller than those obtained using silica sand and thus it is considered that the tests using silica gave results which were more reliable. The volume losses of alumina ceramics abraded against ash were insufficient to give reliable wear test data and it was concluded' that ash could not be used to rank materials of high hardness. A number of materials were ranked for wear resistance using silica sand abrasive particles. The alumina ceramics and tungsten carbide composite materials showed the best wear performance.

Contains published articles. / Bibliography: p. 131-138. / Experiments on the behaviour of brittle materials, particularly quartz, during the initial period of loading, at the onset of fracture and during the stage of rapid crack propagation, have been performed in order to gain an insight into some of the fundamental processes of brittle fracture which might prove useful in the solution of practical problems in mining and process operations. Studies have also been conducted into the nature of the damage produced by a diamond stylus sliding on a quartz surface. The results of these studies have been correlated with ancilliary experiments in which the effects of the deformation produced by a sharp indenter and sliding diamond polishing particles on quartz were examined. The geometry of cracks in glass and quartz during the loading stage has been observed by scanning electron microscopy and was found to be approximately elliptical. By making some simple assumptions, an equation has been derived that enables the fracture surface energy to be deduced from shape of the crack and the loading conditions. The brittleness of a material is indicated by the difference between the fracture surface energy and the thermodynamic surface energy of the material. In a perfectly brittle body they are equal. In order to make an accurate comparison with the true surface energy in quartz as calculated from atomic bond energy data, anisotropic elastic theory was used in the evaluation of the fracture surface energy.

An investigation was performed on alloys of a 12wt% Cr steel (3CR12) that had undergone different thermomechanical processing routes prior to the annealing step of production. The aim of this research was to identify any changes in the annealing response of 3CR12 steel due to the different thermomechanical processing routes to which the alloys had been subjected. Three alloys of 3CR12 were subjected to various cooling treatments immediately following hot rolling. This was accomplished by water quenching, air cooling and insulatory cooling of the hot-rolled plates. Characterisation of the alloys in the hot-rolled and annealed conditions included optical and electron microscopy, macro- and microhardness tests, tensile tests and impact energy tests, dilatometry, differential thermal analysis and volume fraction analysis. Alloys in the as-rolled state exhibited an increase in martensite content related to an increase in the cooling rate following hot rolling. The increase in martensite content influenced the alloys' mechanical properties by raising their bulk hardness, UTS and yield strength. This was coupled with a decrease in the elongation and impact toughness of the alloys. Furthermore, the Ac1 was found to increase with higher cooling rates following hot rotting. Variations in the as-rolled microstructures influenced the alloys' subsequent annealing response, with there being a greater and more immediate response to the annealing treatment by alloys subjected to higher cooling rates after hot rolling. The fully annealed alloys showed little variation in their microstructures or mechanical properties, despite clear variations in their as-rolled microstructures.