AbstractsEngineering

Marine and offshore structures are continuously exposed to different loading types. Cyclic loads results in cyclic stresses in structural components, which may result in degradation of the structural integrity. This process is called fatigue. Fatigue damage of structural components usually occurs without any —global— visible sign. Ultimately, damage in critical components of a structure may lead to catastrophic fracture, and must therefore be prevented. Currently, for uniaxial fatigue of welded joints in marine and offshore structures several design techniques are available which are approved by classification societies. Research indicates that uniaxial fatigue life prediction methods results in a lifetime overestimation when they are applied on structural details exposed to multiaxial stress states. Differences in calculated lifetime and real lifetime can be up to a factor 14. During this research a new method in order to calculate the fatigue lifetime of a multiaxial loaded weld, proposed by researchers of Battelle is implemented. The object of interest is a flex-lay tower structure located on a vessel which is able to install flexible pipes, risers and in-line structure at sea. For the new proposed method the mesh insensitive Verity structural stress, based on nodal forces and moments is determined along the weld. The coupling with a hydrodynamic analysis of the vessel results in a normal and shear stress signal, which is used for further fatigue calculations. A path dependent maximum range (PDMR) cycle counting method is programmed and applied. This method uses the path length as measure for the effective stress, taking directly the effect of the shear stress into account. Comparison of the results of the new proposed method is made with the prescribed methods by classification codes. The class codes instruct in case of a multiaxial loaded weld, summation of the damage due to normal and shear stress. An implicit assumption in the approach is that the two stress types act independently, and both contribute to the fatigue damage in the same way as they would do when acting alone. The damage as prescribed by the class codes is calculated using the hot spot stress and rainflow cycle counting. For a proportional loaded weld in the flex-lay tower leg, the new proposed method shows a factor 3 higher damage than the method prescribed by the current codes. The damage due to the shear stress is negligible in the method using the codes. No non-proportional stress state suitable for analysis is found; therefore a realistic one is obtained by manipulation of the proportional data. Comparison of the proportional and non-proportional stress histogram shows that the non-proportional stress histogram shows on average more cycles in the region of the mode and higher ranges at the cost of the cycles at the ranges positioned left of the mode. For the non-proportional loaded weld, the new proposed method shows a factor 4.2 higher damage as calculated using the current class codes. The focus of this research is on the loading…