Continuum approach to fatigue life prediction based on defect size

Fatigue analysis is essential in the design of machine components because fatigue is one of the biggest problems in component failure. Many models have been developed for this problem, for example, cycle counting methods and continuum-based models. However, most fatigue models are based on the assumption of constant amplitude loading. This assumption is not always sufficient to describe the multiaxial loadings that commonly occur in machine components. In this thesis, we use a continuum model. The aim of this thesis was to investigate how we take small defects into account in a continuum-based high cycle fatigue model. This fatigue model was developed by Ottosen, Stenström and Ristinmaa in 2008. Continuum high cycle fatigue is based on a stress space endurance surface and the accumulation of damage is described by an evolution equation. We made changes to the original endurance function to better fit the experimen tal data for a material containing small defects. Murakami-Endo equation was used to consider the effect of small defects on the endurance surface. This introduced the property that the fatigue strength decreases with the growth of the damage. This reduction in the endurance surface also makes the damage growth nonlinear. Including the Lode angle in the fatigue model introduced other properties. It allows us to redefine the fatigue limit ratio and better fit the experimental data. We mainly used steels S35C and S45C with different initial defect lengths as materials in the tests. This is because they provided the most experimental data and quite a lot of research. We performed a few tests on the model to see that it predicts approximately correctly. The tests included uniaxial, pure shear and multiaxial loading cases. We can state that the model is able to describe several fatigue cases well.