This thesis describes a series of fretting fatigue experiments carried out under closely controlled conditions of partial slip. These experiments confirm the existence of a size effect whereby the fretting fatigue life of an aluminium alloy is shown to vary with contact size. The configuration chosen, of cylindrical fretting pads contacting a plane specimen is amenable to classical stress analysis and the surface tractions between the contacting bodies are derived. The effects of tension in the specimen, finite specimen thickness, differing elastic constants, and surface roughness are all investigated and incorporated into the analysis where appropriate. A technique is then developed to calculate stress intensity factors for plane cracks growing under the contact load at an arbitrary angle to the free surface. The analysis is then applied to the experimental results and three possible explanations for the size effect are proposed, based on statistical effects, crack arrest, and crack initiation. These are examined in the light of the experimental evidence and it is proposed that the variation of fatigue life with contact size is due to an increase in the amount of fretting damage above a threshold level for crack initiation. A composite parameter is chosen to characterise the severity of fretting conditions and this is shown to describe the experimental results accurately. Finally, the use of this parameter in design calculations is discussed.

The main purpose of this research project was to determine the fatigue-life behaviour of AISI 1008 sheet steel which has been mechanically formed to a radius of curvature of 120mm and then to correlate the fatigue-life behaviour to that of the parent or “as manufactured” material. During the forming process it was felt important to induce plastic strain through stretch-bending by clamping the sides of a plate sample’s (width) edges in the bending fixture before being bent by a single acting mechanical press. It was determined through actual testing that there was a decrease in fatigue-life when the mechanical formed data was compared to fatigue data of the parent material. Standard fatigue mathematical models were used to relate the actual fatigue data. Due to the material being cold formed to a radius of curvature of 120mm, residual stresses induced during the forming process played an essential role during the fatigue-life prediction calculations. The maximum relieved stress in the parent material was compressive in nature having a magnitude of 11percent of the “as manufactured” yield strength (265 MPa). For the mechanical formed material compressive residual stresses were measured on the outer surface while tensile stresses were measured on the inner surface. The difference between actual number of cycles to failure to that calculated using the standard mathematical models for the parent material, ranged between 48 percent and 18 percent and for the mechanical formed samples between 35 percent and 1percent, depending on the strain amplitude used. An important aspect of this study was to determine the criteria required for mathematical modelling of the parent material as testing occurred between the limit of proportionality and yield point. Although this aspect requires further investigation the mathematical results obtained during this study were considered to be acceptable.

No description available.

Electron BackScattered Diffraction (EBSD) systems can be considered as a tool providing three kinds of responses: EBSD patterns (EBSPs) themselves, an indexing response (orientation data), and an intensity response (also called Band Contrast "BC", Image Quality "IQ", or Pattern Quality "PQ or P"). This work focused on the characterization and development of the intensity response. For now, the intensity response cannot be used for quantitative microstructural analyses, including strain analyses of materials. Indeed, this response is affected by several material and experimental conditions. Moreover, properties of the intensity response (strain sensitivity, reproducibility, exact relation with EBSP quality or diffraction band intensities) are not well-known and understood. This project constitutes an exploratory study on the development of the intensity response for quantitative strain analyses of materials. A new modelling and statistical approach is presented and assessed for transforming the raw (current) intensity response (values and grey-tones of intensity images) of commercial EBSD systems into an accurate and reproducible parameter allowing objective visualization and measurements of strain. / Les systèmes de diffraction des électrons rétro-diffusés (EBSD) peuvent être considérés comme un outil offrant trois types de réponses: les patrons EBSD (EBSPs), la réponse d'indexation (données d'orientation) et la réponse d'intensité (aussi appelée Contraste de Bande «BC », Qualité d'Image « IQ », ou Qualité de Patron « PQ ou P »). Le présent projet s'est concentré sur la caractérisation et le développement de la réponse d'intensité. En effet, pour l'instant, la réponse d'intensité ne peut être utilisée pour réaliser des analyses microstructurales quantitatives des matériaux. En effet, cette réponse est influencée par de nombreux facteurs relatifs au matériau analysé et aux conditions expérimentales utilisées. De plus, les propriétés de la réponse d'intensité (sensibilité à la déformation, reproductibilité, relation exacte avec la qualité des EBSPs ou les intensités des bandes de diffraction) ne sont pour l'instant pas bien connues ni même compnses. Le présent projet représente donc une étude exploratoire visant le développement de la réponse d'intensité des systèmes EBSD commerciaux pour l'analyse quantitative de la déformation des matériaux. Dans ce projet, une approche de modélisation statistique est présentée et évaluée afin de transformer la réponse d'intensité actuelle (valeurs et tons de gris des images d'intensité) des systèmes EBSD commerciaux en une réponse précise et reproductible permettant de visualiser et de mesurer objectivement la déformation.

An effective way to describe changes in the microstructure of a material or to assess fatigue damage at an early stage in fatigue life is by measuring the acoustic nonlinearity parameter beta. The nonlinearity parameter is defined for harmonic longitudinal plane waves and it depends on the ratio of the amplitudes of the first harmonic of the exciting signal and the second harmonic. A reliable measurement of the amplitudes of these harmonics is crucial since their amplitude of the second (higher) harmonic is much smaller than the amplitude of the first harmonic. This research investigates the influence of the apparent nonlinearity that can occur due to the signal processing and shows how this influence can be quantified and minimized to enable a more accurate evaluation of the acoustic nonlinearity parameter. Furthermore the concept of the nonlinear parameter is extended to Rayleigh surface waves by developing a connection between the harmonic amplitudes and the third order elastic constants, using the approximate model by Zabolotskaya. Finally the problem of modeling the influence of fatigue damage on the nonlinear parameter and the elastic constants is discussed. The reduction of the processing nonlinearity combined with Rayleigh surface waves - Rayleigh surface waves are more efficient in the detection of fatigue damage initiated and concentrated at the surface - helps improve the prediction of fatigue damage and the remaining life of a sample.