TWI Industrial Member Report Summary 144/1981
By T R Gurney
Background
Fatigue tests have been carried out on longitudinal non-load-carrying fillet welded joints using simple loading sequences in which each constant amplitude load cycle had secondary load cycles of one or two magnitudes appended to it. The loading was entirely tensile so that it was anticipated that stress interaction would occur to give Sigma n/N > 1. In fact Sigma n/N was, in almost every case, less than 1.0. Based upon an empirical fit of an equation to one set of results it proved possible to predict the remainder of the results with reasonable accuracy. However the new rule is not consistent with the application of simple fracture mechanics methods.