(source from www.twi.co.id, http://www.civil.usyd.edu.au/images/history/trahair_history_fig_15_small.jpg)


Fatigue crack growth calculations were performed on offshore tubular joints caused weld defect(porosity,slag inclusion, undercut, hydrogen diffusible). The stress intensity factors required for such calculations were obtained from T-butt solutions . The applicability of the solutions to tubular joints was first demonstrated by comparing the fatigue life of a base case with that obtained from a mean S–N curve, and the influence on fatigue life of various factors including load shedding, the size of initial defects, weld geometry, etc. was investigated. The solutions were then used to predict the lives of tubular T-joints from an experimental database. The results show that the solutions underestimate the fatigue life; this underestimation was shown to be primarily due to ignoring the combined effects of load shedding and the intersection stress distribution. In general, however, the trends in the predicted fatigue lives with joint geometry and other details were seen to be superior to predictions from the S–N approach, with the solutions significantly reducing the dependency on loading mode exhibited by the test data.
Stress concentration factors (SCF's) calculated from parametric formulas were also compared with those obtained from finite elements models. The results of this work have the intention of verifying the validity of the refinement of fatigue analysis on critical joints of the platform and the evaluation of its consequences in the inspection plan. Obtained results show that finite elements analysis is recommended for joints whose type and/or behaviour is not consistent with standard models, used by computational programs such as ANSYS 5.7, STRUCAD 3.4 for calculation out plane bending, axial tension, in plan bending and fatigue life(chord and brace).