Multiaxial fatigue assessment of a structural steel joint according to the peak stress method

In fatigue design of welded joints, the local approach based on the notch stress intensity factor (NSIF) assumes that the weld toe profile is a sharp V-notch having a tip radius equal to zero, while the root side is a pre-crack in the structure. The Peak Stress Method (PSM) is an approximate, FE-oriented application of the NSIF approach to fatigue design of welded joints, which is based on the singular linear elastic peak stresses calculated from FE analyses carried out by using a proper mesh pattern. The element type is kept constant while the average element size can be chosen arbitrarily within a given range of applicability. The PSM allows rather coarse FE meshes to be adopted if compared to those required for the NSIFs evaluation from the local stress fields. The PSM was originally validated for pure axial or bending loadings as well as pure torsion loadings. Recently, the PSM has been extended to analyse the fatigue strength of welded joints subjected to multiaxial fatigue loadings. By adopting the averaged Strain Energy Density (SED) as a fatigue strength criterion, a design stress, the so-called equivalent peak stress, can be defined and adopted in conjunction with a reference design fatigue curve to assess multiaxial fatigue failures. In the present contribution, new fatigue test results relevant to weld toe failures have been generated by testing full-penetration welded steel joints under in-phase bending-torsion fatigue loadings, generated by an external fatigue axial loading (F in the figure). Joints, having the geometry reported in the figure, have been tested under both as-welded and stress relieved conditions. First, the experimental fatigue test results have been compared with the theoretical estimations based on the nominal stress approach as proposed in International Standards and Recommendations. Then, experimental data have been re-analysed using the equivalent peak stress in conjunction with the multiaxial design scatter band according to PSM. All experimental data fall definitely in the safe side as compared to the design scatter band. To better explain this result further investigations have been carried out by analyzing in more detail both the root radius and the residual stress field at the weld toe. It has been shown that the notch tip radius at the weld toe is about equal to 4.5 mm, so that the sharp V-notch hypothesis was not consistent. Moreover, high values of compressive residual stresses have been measured close to the weld toe of both as-welded and stress-relieved joints by means of X-ray diffraction technique. Finally, by considering a blunt V-notch at the weld toe, the deviation between experimental results, re-converted in terms of equivalent peak stress, and the PSM-based design curve has been significantly reduced, even if results fall slightly on the safe side. However, this could be explained on the basis of the compressive residual stress state, which allows to protect the weld toe against fatigue crack initiation and propagation, increasing the fatigue life.