Fatigue strength assessment of arc-welded aluminum joints by local approaches

This study investigates the applicability and reliability of the two stress-based local fatigue assessment approaches, namely the effective notch stress (ENS) and 4R approach, for welded aluminum joints. Statistical analyses were performed using a substantial number of experimental fatigue data from literature, limited to various arc-welded joints, manufactured from 5xxx and 6xxx alloys, and tested under constant amplitude loading. The 4R approach is an effective notch stress approach with mean stress correction that considers local elastic–plastic behavior. Several assumptions were necessary to apply the 4R method on welded aluminum joints and a novel Monte Carlo-simulation-based simulation decomposition (SimDec) global sensitivity analysis was performed to support these assumptions. The obtained results indicated that the ENS reference curve recommended by IIW is slightly nonconservative. Furthermore, it was identified that the mean stress, i.e. the R-ratio, has an influence on the fatigue strength, likely due to low residual stresses within the specimens. The 4R method decreased the scatter of the extracted experimental data significantly. Based on the results of this study, both investigated local approaches have high potential for fatigue applications and design of optimized and safe welded aluminum structures.