Automated implementation of the Peak Stress Method for the fatigue assessment of complex welded structures

Dealing with the fatigue design of complex welded structures, the Peak Stress Method (PSM) is a rapid approach to the fatigue strength assessment, which is based on the singular, linear-elastic opening, in-plane and out-of-plane shear peak stresses calculated at the weld toe or weld root adopting coarse FE meshes. By adopting the averaged Strain Energy Density (SED) as a fatigue strength criterion, a design stress, the so-called equivalent peak stress, has been defined as a proper combination of the local peak stresses and adopted in conjunction with a reference design fatigue curve to estimate the fatigue lifetime of welded structures. To furtherly enhance the applicability to complex welded structures, the PSM has recently been calibrated by using 10-node tetra elements, which allow to easily discretize complex geometries. Moreover, an interactive analysis tool has been developed within Ansys® Mechanical, in order to automate the application of the PSM to complex both aluminium and steel 3D welded structures subjected to multiaxial fatigue loads. The automated PSM tool has been developed taking advantage of the integrability between Ansys® Mechanical and modern programming languages, enforcing PSM compatibility requirements and automating the implementation tasks and phases of the method. In this perspective, the present work deals with the employment of the PSM tool for the fatigue strength assessment of some complex welded details adopted in modern roller-coaster structures. Due to the geometrical complexity of large-scale structures, private companies usually employ finite element beam models in order to evaluate nominal stresses, that must be compared with appropriate fatigue strength references (FAT classes) in design standards. The assessment procedure proposed in this work aims instead at defining appropriate stress-based FAT classes from the results of automated PSM analyses on the considered joint geometries. Moreover, the proposed approach allows to compensate the limited number of FAT classes and details available in International Standards and Recommendations, when dealing with complex joint geometries