Weld toe fatigue crack initiation is highly dependent on the local weld toe stress-concentrating geometry including any inherent flaws. These flaws are responsible for premature fatigue crack initiation (FCI) and must be minimised to maximise the fatigue life of a welded joint. In this work, a data-rich methodology has been developed to capture the true weld toe geometry and resulting local weld toe stress-field and relate this to the FCI life of a steel arc-welded joint. To obtain FCI lives, interrupted fatigue test was performed on the welded joint monitored by a novel multi-probe array of alternating current potential drop (ACPD) probes across the weld toe. This setup enabled the FCI sites to be located and the FCI life to be determined and gave an indication of early fatigue crack propagation rates. To understand fully the local weld toe stress-field, high-resolution (5 μm) 3D linear-elastic finite element (FE) models were generated from X-ray micro-computed tomography (μ-CT) of each weld toe after fatigue testing. From these models, approximately 202 stress concentration factors (SCFs) were computed for every 1 mm of weld toe. These two novel methodologies successfully link to provide an assessment of the weld quality and this is correlated with the fatigue performance.

The energy efficiency of electric machines can be improved by optimizing their manufacturing process. During the manufacturing of ferromagnetic cores, silicon steel sheets are cut and stacked. This process introduces large stresses near cutting edges. The steel near cutting edges is in a plastically deformed stress state without external mechanical load. The magnetic properties of the steel in this stress state are investigated using a custom magnetomechanical measurement setup, stress strain measurements, electrical resistance measurements, and transmission electron microscopic (TEM) measurements. Analysis of the core energy losses is done by means of the loss separation technique. The silicon steel used in this paper is non-grain oriented (NGO) steel grade M270-35A. Three differently cut sets of M270-35A are investigated, which differ in the direction they are cut with respect to the rolling direction. The effect of sample deformation was measured—both before and after mechanical load release—on the magnetization curve and total core energy losses. It is known that the magnetic properties dramatically degrade with increasing sample deformation under mechanical load. In this paper, it was found that when the mechanical load is released, the magnetic properties degrade even further. Loss separation analysis has shown that the hysteresis loss is the main contributor to the additional core losses due to sample deformation. Releasing the mechanical load increased the hysteresis loss up to 270% at 10.4% pre-release strain. At this level of strain, the relative magnetic permeability decreased up to 45% after mechanical load release. Manufacturing processes that introduce plastic deformation are detrimental to the local magnetic material properties.

No abstract

This work presents the case study of a welded overhead crane runway girder. A full-shell global finite element model was developed and validated using strain measurements. The global model drives solid submodels of different scales. Using load spectra based on real operational data, hot spot stress and XFEM-based fracture mechanics simulations are performed to assess the fatigue properties of two critical joints. For the fracture mechanics-based approach, it is shown that small cracks quickly converge to the same aspect ratio. Hence, the choice of the initial crack aspect ratio was found to be inconsequential. But the aspect ratio must not be assumed to be constant during the fatigue crack growth simulations. Finally, it is shown that the size at which a crack is transferred from one submodel to a subsequent is critical for the accuracy of the fatigue crack growth simulations.