Residual Stress Effect on Fatigue Strength of Non-Load-Carrying Cruciform Welded Joints of SM570Q Steel for Welded Structures

Abstract

It is common to evaluate the tatigue strength of welded joints by testing simple specimens incorporating the weld detail of interest, usually at R = 0. However, such specimens do not usually contain significant residual stress, whereas yield-magnitude tensile residual stresses generally exist in real welded structures. As a result, the effective maximum stress in the fatigue loading is equal to yield, giving more severe loading than R = 0. Thus, unsafe estimates of fatigue lives may be obtained from small-scale specimens at R = 0. In this study, the technique of applying fatigue loading that cycles down from yield to simulate the effect of high tensile residual stress (σmax = σy testing) was applied to non-load-carrying cruciform welded joints in 20 mm thick SM570Q steel. Comparative tests were performed on narrow specimens under σmax = σy loading and wider specimens, fabricated to ensure that they did contain high tensile residual stress, under R = 0. The results were in good agreement, confirming that the fatigue performance of large-scale welded structures can be reproduced using the σmax = σy test technique.