Over the past eighteen months, a specialist Drafting Panel has been producing recommendations, for the UK Department of Energy, for revision of the fatigue sections of the Department's document 'Offshore Installations: Guidance on Design and Construction l These sections refer to steel welded joints. This paper reviews briefly the basis for this revision to the guidance. Fuller discussion of the issues involved are given in a document to be published shortly by the Department of Energy.2
The past decade has seen a considerable development in the method used for the assessment of fatigue in offshore structures. The method is based on an S-N curve and developments have been in both the appropriate S-N curve to use and the definitions of stress (S) and life (N). Steel structures contain both simplenon-tubular butt or fillet welded joints as well as more complex fillet welded tubular joints. In the UK, the former have come to be assessed using a classification of welds into eight classes B, C, D, E, F, F2, G, W. Each class has a design S-N curve defined as - 2 standard deviations on the mean through the experimental data relevant to that class. Stress is taken as the nominal stress range, the local stress concentration of the weld detail being accommodated in the baseline data. Macrostress concentrations, such as holes, in which the weld detail is placed are however taken up in defining nominal stress. The fatigue life N is taken as complete failure in the test samples used to generate the data. The resulting S-N design curves have been incorporated in British Standard 5400 Pt 10, a Code of Practice for Steel, Concrete and Composite Bridges,3 and are shown in Fig. 1. The same curves are recommended for non-nodal joints in the suggested revision of the UK Department of Energy's Guidance Notes on fatigue,1 for protected joints or joints in an air environment.
Loading of a nodal tubular joint results in stress concentrations at and around the joint interface. For such joints, S-N curves have been developed based on two concepts. The first is to attempt to define a crude nominal stress to correlate failure. Such a stress, termed the punching shear stress, is the mean shear stress through the chord section normal to the chord surface at the intersection. Although the punching shear method is incorporated in the current D of En Guidance Notes,1 l it has been dropped in the proposed revision4 because of the development of the second concept, namely hot spot stress or strain range.
The hot spot stress is the peak stress around the intersection of two or more tubes. Its actual definition in the peak stress region is less precise. In Ref 1 (para 4.2.1.11) it is defined as 'that which is as near the weld as possible without being influenced by the weld profile'. The range of influence of the weld profile on stresses is not defined.
