| Summary: | This thesis experimentally investigates the statistics (mean and variance) of the fatigue-cycle-dependent evolutions of both the crack tip front penetration's distribution function and the microscopic growth rate's distribution function, as a fatigue crack propagates to final fracture. A novel technique which facilitates striation counting and striation spacing measurements, is developed and used for extracting and analyzing the relevant statistical data for characterizing the stochastic fatigue crack propagation in polycrystalline metals. Two types of pure copper materials are investigated. === The investigation confirms the existence of the mean and variance of both the crack front penetration and its growth rate. Details of the variations of the mean crack penetration with respect to the dispersion of the crack front distribution and the mean growth rate, respectively, are established. Other contributions include the evaluation of the material characteristic associated with the transition intensity of the growth process. === These results are correlated with the predictions of the "Provan-Ghonem" theory in order to ascertain the validity of the linear Markov birth stochastic process, as a viable description of the fatigue crack propagation process in polycrystalline metals. The trend of the experimental results suggest a spatially correlated Markov process which accounts for both the strong nearest-neighbour-interactions between "points" along the crack front, and the boundary effects as a more viable representation of the fatigue crack propagation process.
|
|---|
