A survey is made of analytic, experimental and numerical techniques in the field of the dynamic initiation, and early propagation, of cracks. As no closed form analytic solutions exist for finite geometries, even in the elastic case, numerical and experimental techniques have to be developed. In instrumented impact tests plasticity often occurs. The specific problem of the Instrumented Charpy Test is discussed in detail by virtue of its technological significance and the extensive literature available for the test. Although a standard for the test has been proposed there are still outstanding questions to be answered, for which the techniques described above can be used. The problem of the dynamic calibration of various notched geometries is addressed in the original work of the thesis. The Charpy, Izod, Slender Cantilever and Double Notched Bar geometries are studied using dynamic photoelasticity and 8000 fps photography. It is shown that the response of the DNB is more straightforward than the Charpy geometry. Further photoelastic study of the latter two geometries, using epoxy model material and 10⁶ fps photography, gives a quantitative measure of the growth of stress intensity factor at the notch tips and hence a dynamic calibration is deduced. An explicit finite difference code is used to supplement photoelastic data. Having achieved progress in the derivation of the dynamic calibration of the two selected geometries, corresponding instrumented impact tests are then undertaken. The Hopkinson Pressure Bar method of loading is used. It is concluded that the proposed standard for the Instrumented Charpy Test is valid within limits but that there is a requirement for a dynamic calibration. Such a calibration is complex in the case of the Charpy geometry whereas a simpler geometry, viz. DNB, could prove to be more amenable to analysis and hence be more practical from the technological point of view.

Recent application of using spheres as carriers of LNG on ships has highlighted a lack of guidance given to the inclusion of imperfections and to the operating con- ditions imposed as a result of the motion of the ship. Spheres with a radius to thickness ratio of between 450 and 1700 were made by electrodeposition of copper on a wax former. Imperfections of three types were introduced: <ul><li>a) A local flat.</li><li>b) A zone of reduced curvature.</li><li>c) A bulge.</li></ul> These shells were tested under partial vacuum and the deformation at selected points monitored by proximity gauges. Buckling pressures were found to range between ⅕ and one-fifteenth of the theoretical value for a perfect shell. The experimental results were analysed on the basis of Koiter's imperfection theory and by a numerical computer solution. The former showed good agreement with types a) and b) and no agreement with type c); the latter showed little agreement with the experimental results. The ASME design code was found to be adequate for all imperfection sizes considered, BS 5500 only partly. Cast epoxy models with a radius to thickness ratio of 900, supported on an equatorial ring and partially filled with water, were subjected to vertical accelerations both normal and inclined to their equatorial rings. Buck- ling was detected by proximity gauges, and occurred as a result of tensile meridional and compressive circumferential stresses. The sphere was found to be imperfection insen- sitive and to be more susceptible to buckling at approximately ⅓ and ⅔ full for respectively the normal and inclined loading cases. Both a numerical com- puter solution and an analytical analysis were found to be in good agreement with the experimental results. Both the ASME and BS5500 design codes were however, found for this mode of failure to be too conservative for design purposes.

Radiation-induced changes in microstructures often lead to significant changes in mechanical properties of alloys used in the construction of nuclear reactors. It is desirable to test small specimens to make efficient use of the small volumes available in test and commercial reactor cores and also because small specimens are less affected by the sometimes steep flux gradients experienced in reactor cores and the sometimes large temperature gradients developed in the specimens from gamma heating. (Continues...).

Weight loss and depletion of fat-free mass commonly occurs in patients with COPD. The objective of the study was to determine the magnitude and duration of protein depletion during an episode of acute exacerbation. Fifteen patients (9 women and 6 men) admitted for an acute exacerbation of COPD participated in a descriptive study that prospectively measured individual nitrogen balance over a 6-week follow-up period using repeated nitrogen balance tests. / The mean nitrogen balance in hospital was -13.20 +/- 11.63 g N/day. Only 2 patients achieved a positive nitrogen balance by 2 weeks post-admission and 4 more patients by 4 weeks post-admission. At 6-weeks post-admission, 7 patients (47%) were still in negative nitrogen balance (-10.75 +/- 9.34 g N/day). Protein and energy intakes were significantly higher in patients who achieved a positive nitrogen balance (1.7 +/- 0.5 g protein/kg/day and 120 +/- 30% of estimated energy expenditure (1.7 x REE)) than patients who remained in a negative nitrogen balance (1.3 +/- 0.6 g protein/kg/day and 70 +/- 20% of estimated energy expenditure). There were no significant changes in weight or handgrip strength over the follow-up period. No effect of cumulative or daily corticosteroid doses on nitrogen balance or changes in handgrip strength were found. / In conclusion, the catabolic stress of an acute exacerbation on nutritional status is remarkable. Patients admitted for an acute exacerbation of COPD are in severe negative nitrogen balance, which improves very slowly post-discharge. A negative nitrogen balance is prolonged in patients who have a decreased protein and energy intake.

Recycling has proven not only to address today’s economical, environmental and social issues, but also to be imperative for the sustainability of human technology. The current thesis has investigated the feasibility of a new philosophy for Recycling (Alloying-Recycling) using steel 1020 and aluminum 6061T6. The study was limited to the metallurgical aspects only and has highlighted the potential of recycled alloys made of recycled aluminum and steel to exhibit substantially increased wear resistance and strength-to-weight ratio as compared to initial primary materials. Three alloy-mixtures are considered: TN3 (5wt% 1020 +95wt% 6061T6); TN5 (0.7wt% 1020 + 99.3wt% 6061T6); and TN4 (10wt% 6061T6 + 90wt% 1020). A Tucker induction power supply system (3kW; 135-400 kHz) is used to melt the alloy mixtures for casting in graphite crucibles. Heat treatment of the cast samples is done using a radiation box furnace. Microscopy, Vickers hardness and pin-on-disc abrasive wear tests are performed. Casting destroyed the initial microstructures of the alloys leading to a hardness reduction in the as-cast and solution heat-treated aluminum rich samples to 60 Hv from 140 Hv. Ageing slightly increased the hardness of the cast samples and provided a wear resistance two times higher than that of the initial 6061T6 material. On the steel rich side, the hardness of the as-cast TN4 was 480 Hv, which is more than twice as high as the initial hardness of steel 1020 of 202 Hv; this hints to strong internal and residual stress, probably martensite formation during fast cooling following casting. Solution heat treatment lowered the hardness to the original value of steel 1020, but provided about ten (10) times higher wear resistance; this suggests higher ductility and toughness of normalised TN4 as compared to 1020. In addition, TN4 exhibits about 25% weight reduction as compared to 1020. The actual recycling process and the effect of non-metallic impurities shall be investigated in future works. Also, the casting and heat treatment processes need to be improved.

The microstructure of low alloy Transformation Induced Plasticity (TRIP) assisted steels has been systematically varied through the combination of computational and experimental methodologies in order to enhance the mechanical performance and to fulfill the requirement of the next generation Advanced High Strength Steels (AHSS). The roles of microstructural parameters, such as phase constitutions, phase stability, and volume fractions on the strength-ductility combination have been revealed. Two model alloy compositions (i.e. Fe-1.5Mn-1.5Si-0.3C, and Fe-3Mn-1Si-0.3C in wt%, nominal composition) were studied. Multiphase microstructures including ferrite, bainite, retained austenite and martensite were obtained through conventional two step heat treatment (i.e. intercritical annealing-IA, and bainitic isothermal transformation-BIT). The effect of phase constitution on the mechanical properties was first characterized experimentally via systematically varying the volume fractions of these phases through computational thermodynamics. It was found that martensite was the main phase to deteriorate ductility, meanwhile the C/VA ratio (i.e. carbon content over the volume fraction of austenite) could be another indicator for the ductility of the multiphase microstructure. Following the microstructural characterization of the multiphase alloys, two microstructural design criteria (i.e. maximizing ferrite and austenite, suppressing athermal martensite) were proposed in order to optimize the corresponding mechanical performance. The volume fraction of ferrite was maximized during the IA with the help of computational thermodyanmics. On the other hand, it turned out theoretically that the martensite suppression could not be avoided on the low Mn contained alloy (i.e. Fe-1.5Mn-1.5Si-0.3C). Nevertheless, the achieved combination of strength (~1300MPa true strength) and ductility (~23% uniform elongation) on the low Mn alloy following the proposed design criteria fulfilled the requirement of the next generation AHSS. To further optimize the microstructure such that the designed criteria can be fully satisfied, further efforts have been made on two aspects: heat treatment and alloy addition. A multi-step BIT treatment was designed and successfully reduced the martensite content on the Fe-1.5Mn-1.5Si-0.3C alloy. Microstructure analysis showed a significant reduction on the volume fraction of martensite after the multi-step BIT as compared to the single BIT step. It was also found that, a slow cooling rate between the two BIT treatments resulted in a better combination of strength and ductility than rapid cooling or conventional one step BIT. Moreover, the athermal martensite formation can be fully suppressed by increasing the Mn content (Fe-3Mn-1Si-0.3C) and through carefully designed heat treatments. The athermal martensite-free alloy provided consistently better ductility than the martensite containing alloy. Finally, a microstructure based semi-empirical constitutive model has been developed to predict the monotonic tensile behavior of the multiphase TRIP assisted steels. The stress rule of mixture and isowork assumption for individual phases was presumed. Mecking-Kocks model was utilized to simulate the flow behavior of ferrite, bainitic ferrite and untransformed retained austenite. The kinetics of strain induced martensitic transformation was modeled following the Olson-Cohen method. The developed model has results in good agreements with the experimental results for both TRIP steels studied with same model parameters.

The effect of fluorine on the solubilities of Mn-columbite (MnNb2O6), Mn-tantalite (MnTa2O6), zircon (ZrSiO4) and hafnon (HfSiO4) were determined in highly fluxed, water-saturated haplogranitic melts at 800 to 1000 °C and 2000 bars. The melt corresponds to the intersection of the granite minimum with the albite-orthoclase tieline (Ab72Or28) in the quartz-albite-orthoclase system (Q-Ab-Or) due to the addition of P2O5 to the melt. The melt content of P2O5 is 1.7 wt. %, and also contains 1.1 and 2.02 wt. % of Li2O and B2O3, respectively. The composition of the starting glass represents the composition of melts from which rare-elements pegmatites crystallized. Up to 6 wt. % fluorine was added as AgF in order to keep the aluminum saturation index (ASI) of the melt constant. In an additional experiment F was added as AlF3 to make the glass peraluminous. The nominal ASI (molar Al/[Na+K]) of the melts is close to 1 and approximately 1.32 in peraluminous glasses, but if Li considered as an alkali, the ASI of the melts are alkaline (0.85) and subaluminous (1.04), respectively. The solubility products [MnO]*[Nb2O5] and [MnO]*[Ta2O5] are nearly independent of the F content of the melt, approximately 18.19 ± 1.2 and 43.65 ± 2.5 x10-4 KSP (mol2/kg2), respectively. By contrast, there is a positive dependence of zircon and hafnon solubilities on the fluorine content, which increases from 2.03 ± 0.03 x10-4 (mol/kg) ZrO2 and 4.04 ± 0.2 x10-4 (mol/kg) HfO2 for melts with 0 wt. % F to 3.81 ± 0.3 x10-4 (mol/kg) ZrO2 and 6.18 ± 0.04 x10-4 (mol/kg) HfO2 for melts with 8 wt. % F. Comparison of the data from this work and previous studies indicates that ASI of the melt seems to have a stronger effect than the contents of fluxing elements in the melt and the overall conclusion is that fluorine is less important (relative to melt compositions) than previously thought for the control on the behavior of high field strength elements in highly evolved granitic melts. Moreover, this study confirms that although Nb, Ta, Zr and Hf are all high field strength elements, Nb-Ta and Zr-Hf are complexed differently.

This research study examined the effect of corrosion of web reinforcement (stirrups) on the shear behaviour of slender reinforced concrete (RC) beams. The experimental program consisted of seventeen slender shear-critical RC beams: five uncorroded and twelve corroded beams. The test variables included: 1) corrosion level (0%, 7.5% and 15%); 2) type of stirrups (smooth and deformed); 3) stirrup diameter (D6, D12 and 10M); 4) stirrups spacing (100mm and 200mm); and 5) the presence of CFRP repair. The corroded beams had their stirrups subjected to corrosion using an accelerated corrosion technique and the mass loss in the stirrups was estimated based on Faraday’s law. All of the beams were monotonically tested to failure in three point bending. The corrosion cracks formed were parallel to the locations of stirrups as evidence of the corrosion damage in the corroded beams. The maximum decrease in the ultimate shear strength ranged from 11% to 14.4% for beams with high corrosion level of 15.6% mass loss. At a low corrosion level (4.39% mass loss), the shear strength of beams with smooth stirrups increased up to 35% due to the enhancement of shear friction at the concrete-corroded stirrups interface. The stiffness of the corroded beams was enhanced in comparison to the control beams. The ultimate deflection of the corroded beams was decreased up to 25% in comparison to the control beams. The CFRP repair increased the shear strength by 36% and improved the overall stiffness by 39% in comparison to the corroded unrepaired beams. All of the unrepaired beams failed in diagonal tension splitting, while the CFRP repaired corroded beams failed in diagonal tension splitting in addition to debonding of the FRP or concrete cover delamination. The actual corrosion mass loss results were in good correlation with Faraday’s law for the D12 and 10M stirrups. Poor correlation between actual and estimated mass loss was obtained for D6 smooth stirrups, possibly due to errors in the impressed corrosion. iv The analytical model used the modified compression field theory (MCFT) to predict the shear strength of uncorroded and corroded slender RC beams. In the corroded beams, two reduction factors were added to the MCFT model including the mass loss factor and the effective web width. Predictions based on the model revealed that the control beams gave a very good correlation with the ratio of experimental to predicted values that ranged from 0.94 to 1.02. On other hand, the ratio of experimental to predicted strength in the corroded beams ranged between1.06 to 1.4. The poor correlations were obtained for the beams with the D6 smooth stirrups. This study demonstrates that corrosion of web reinforcement can have a detrimental effect on the shear strength and ductility of slender shear-critical RC beams. The experimental results and analytical approach will be very useful for practicing engineers and researchers dealing with corrosion damage in slender RC members.

Constant strain rate compression and torsion tests were carried out isothermally at temperatures of 875 to 1075(DEGREES)C on a series of six steels. The base steel had a composition of 0.06% C and 1.43% Mn and the others contained one of the following sets of additions: (i) 0.035% Nb; (ii) 0.035% Nb + 0.115% V; (iii) 0.035% Nb + 0.30% Mo; (iv) 0.035% Nb + 1.90% Mn; (v) 0.115% V. The tests were conducted to determine the effects of Mn, Mo, Nb and V, singly and in combination, on the high temperature flow and recrystallization behavior of the materials. The dynamic precipitation kinetics for Nb(CN) and VN were determined by the Weiss method. The two PTT curves were similar, with the nose of the VN curve being situated at a slightly lower temperature (885 vs. 900(DEGREES)C) and at a somewhat longer time (26 vs. 16 s), in agreement with the lower equilibrium solution temperature of VN. The dynamic precipitation kinetics of Nb(CN) were retarded by the addition of Mn, V or Mo. This retardation is attributed to the increased carbonitride solubility that follows the addition of these elements because of the way in which they decrease the C and N activity coefficients. / RTT curves were constructed for dynamic recrystallization in the six steels investigated. These were derived from the peak strains of the compression flow curves, as determined at a strain rate of 3.7 x 10('-2)s('-1). Recrystallization occurred earliest in the plain C steel followed fairly quickly by the 0.115% V steel. All of the Nb bearing steels recrystallized considerably later, with the greatest retardation being noted in the 0.3% Mo steel, where it was nearly twice that due to Nb addition alone. This very large effect, and the retardation due to each of the transition elements, is explained in terms of the electronic differences between iron and the particular element. The effect of the atomic size differences with respect to iron is also considered. / The strengthening due to the presence of Mn, Mo, Nb and V in solution was determined from the yield strengths of these steels. The increment in yield strength over that of the plain C steel was determined as 70% and 7% per 0.1 at.% of Nb and V when each is added singly. The strengthening increased to 80% and 8% respectively for these elements when present jointly in austenite. The strength increments were 9% for Mo and 1.3% for Mn per 0.1 at.% when added to a 0.035% Nb steel. The rank order of these effects is also explained in terms of the electronic and atomic size differences, and a possible reason for the synergistic effect (e.g. in the case of Nb and V in a Nb-V steel) is proposed.

Graduation date: 1993