| dc.description.abstract | Fatigue is a phenomenon having local nature; it is obvious that an accurate fatigue assessment of a component must take into account the local parameters as close as possible to the reality. Fatigue in welded components is even a much more complex phenomenon because of the welding process itself that involves heating and cooling processes, different filler material with consequent inhomogeneity of the joints, residual stress, undercuts and many other defects and imperfections. Despite this, the current standards and recommendations consider, for the fatigue assessment of welded details, mostly global methods, like the nominal stress approach, even if the nature of the fatigue phenomenon is local.
A more precise evaluation of the fatigue life of a component can be performed through the socalled local approaches although they also neglect, like the global approaches, some welding peculiarities or deal with them in a statistical way. This is due to the fact that it is not possible to take into account every aspect that could affect the fatigue life of a component due to the high computational effort required leading in some cases to a not appreciable difference in the assessed fatigue life. Each of these methods consider a local parameter whose critical value has the important advantage to be independent on the global geometry of the component considered. However, also the local approaches have some limitations. First of all, they require a certain expertise considering that each method has its own theory and procedure to be applied and, second, they usually require the use of finite element (FE) software with, for some approaches, high computational time.
The thesis investigates the application of the local approaches for the fatigue assessment of common welded components highlighting the advantages that their use brings to the fatigue assessment of welded components and the discrepancies that can be detected between the different codes and recommendation available nowadays for the fatigue assessment of welded components. The application of the local approaches to complex welded components not enlisted in current standards is also considered in order to prove the wider range of applicability of these methods over the more common global approaches. Finally, a procedure to simplify the application of a local approach, the strain energy density (SED) method, is proposed to overcome one of the major drawbacks of this method and to allow its use as a post-processing tool to be integrated in the currently available software for FE analysis. | en_US |
