Representative structure elements for the fatigue assessment of additively manufactured components

Abstract

The fatigue life estimation of additively manufactured structures can be a very challenging task, because the component behaviour will be influenced by many parameters, such as surface roughness, imperfections and inhomogeneous properties. Furthermore, the loading conditions and the component geometry have to be taken into account. The problem of considering the singular influences adequately is intensified by their interactions, which invokes a simultaneous treatment of all relevant influencing factors. Without predefinition of the fatigue approach, properties to describe the cyclic aspects of component behaviour and the fatigue life are required. Even in the case of using small sized specimens, it is not possible to produce a defect-free material for studying the behaviour of sound material in order to derive cyclic material properties as a requirement for the local strain-based fatigue concept, or in order to derive a reference SN-curve and knock-down factors for load-based concepts.Analysing the microstructure offers additional information about the local material state, because, depending on the lightening strategy, the material can be divided into different areas (up-skin, down-skin, contour, core, etc.) with characteristic distributions of imperfections. In order to use a conventional fatigue approach method, this information is not sufficient, because the interaction of statistically distributed imperfections is superimposed on the effect of singular defects. Due to these reasons and after introducing a new interpretation of the measured stress and strain, a fatigue approach will be discussed, which is based on a combined experimental and numerical derivation of the required local properties of the so-called representative structure elements.In addition to the reduced numerical effort compared to the conventional approaches, a combined experimental and numerical derivation of RSE properties is enabled.