ANALISI STATISTICA DI TEST DI FATICA E REGIONI DI TOLLERANZA

In materials science, fatigue is the progressive and localized structural damage that occurs when a material is subjected to cyclic loading. Fatigue occurs when a material is subjected to repeated loading and unloading. If the loads are above a certain threshold, microscopic cracks will begin to form at the stress concentrators such as surface roughness, micro-cracks or notch or vacancy diffusion in microstructure of material. Eventually a crack will reach a critical size, and the structure will suddenly fracture. It is known that the results of fatigue tests display significant variations even when the test is controlled very accurately. In part, these variations are attributable to non-uniformity of test specimens. Examples of such non-uniformity include slight differences in chemical composition, heat treatment, surface finish, etc. The remaining part is related to the stochastic process of fatigue failure itself that is intrinsic to metallic engineering materials. Adequate quantification of this inherent variation is necessary to evaluate the fatigue property of a material for the design of machines and structures. The objective of this dissertation is to analyse fatigue data that requires techniques from statistics, especially linear regression and tolerance regions. In order to study the statistical analysis of fatigue tests, we initially followed procedures suggested by some norms that have been written about this field. In particularly, we have discussed norm ISO 12107 and norm ASTM E739-10. Only by a critical sense for the interpreting of these norms could we pinpoint some passages that in our opinion need a better explanation. This dissertation comes as a result of an internship with "Centro Ricerche Fiat" (C.R.F.) of Torino, and in particular its department of Materials and Anticorrosion. The C.R.F. was interested in a statistical analysis of data coming from fatigue tests, and thus they provided us with different sets of data that comes directly from the real fatigue tests of the department of the C.R.F. The numerical results shown in this work has been implemented in MATLAB.