Influence of the optical measurement technique and evaluation approach on the determination of local weld geometry parameters for different weld types

The joining by welding processes of components is usually related to the creation of additional notches and geometrical peculiarities. Multiple investigations have shown that a clear correlation between the local weld geometry and the fatigue life of welded joints exist. Thereby, the local increase of the local stress can be expressed by a stress concentration factor at the transition from the base material to the filler material, the so-called weld toe. The stress concentration factor can be determined for the most weld types if the geometric parameters such as plate thickness weld toe radius and flank angle are known. However, no standardized method for the determination of these parameters exists. Beside the well-established 2D-measurement methods on cross sections with weld impression analysis, new 3D-methods based on contactless, optical measurement were applied in the last years for the geometrical analysis of welded joints. With these methods, long length of welds can be analyzed in a very short time and with low effort. However, the influence of the measurement system (geometrical accuracy, lateral resolution) was not quantified yet. Additionally, in all known cases of application different evaluation algorithms were used. This does not allow for a straightforward comparison of the investigated parameters and results. In this round robin study, the determination of weld toe radii and flank angles by different evaluation algorithms and 3D-measurement systems and by different institutes are compared. Furthermore, an approach for the direct determination of the stress concentration factors of fillet welds by translating the complex weld shape in a 2D-finite element simulation was implemented. The results of this direct approach are compared to the stress concentration factors determined indirectly using the geometric parameters and those calculated by established approximation formulas.