Fatigue life prediction of welded steel components based on geometric scan data

Abstract The life cycle assessment of welded steel components subjected to high cyclic loads is usually carried out based on the usual concepts (nominal, structural, and notch stress concept), which only provide rough geometric minimum specifications in the form of weld angles or excess weld metal. These specifications are usually assigned to the corresponding fatigue details and empirically derived. This type of estimation allows for a safe design, but on the other hand, it also gives away potential and can lead to a significant underestimation of the actual service life. In addition, these concepts are practically based on decades‐old tests that do not reflect the advances in welding technology and/or the improvements in welding skills. In order to better exploit the actual potential of safe and economical steel constructions, it is essential to consider geometry as the most significant influencing factor when dealing with welded joints. The introduction of efficient scanning systems and powerful finite element methods has increased the interest in considering geometry more closely in recent years. In fact, numerous research projects attempt to use the real geometry of welded joints as the basis for life cycle assessment. However, new difficulties arise. This article discusses the state of the art by considering real geometry, the problem of notch sensitivity in the context of scan and mesh resolution, and the corresponding approaches.