Optimization of welded K-node in offshore jacket structure including the stochastic size effect

Abstract

Offshore structures are exposed to cyclic loading and thereby at risk of fatigue failure, especially at the welded joints between trusses in offshore jacket foundations. In this paper, an optimization framework for welded joints considering fatigue damages is presented. The framework can be used to optimize the orientation and location of welds connecting the joint in a welded K-node in offshore jacket structures considering fatigue damage. It is known that longer welds are at higher risk of fatigue failure compared to shorter welds. To account for this, the statistical size effect is modelled using the Karhunen-Loéve expansion which can be used to take into account the effect of longer welds in the damage assessment. The approach is included in the optimization framework and it is shown that using this method it is possible to simulate the effect of having a good welding quality and a poor welding quality using the correlation length and coefficient of variation. The proposed optimization framework is validated on a simple plate structure and the effect of the statistical input parameters in the statistical size effect is examined. The results show that the proposed optimization approach is robust in predicting the expected tendencies for a simple plate structure. Furthermore, a full-scale welded K-node in an offshore jacket structure manufactured using cast steel is optimized with respect to mass. The optimization is performed considering a good quality weld and a poor-quality weld by including the statistical size effect. The results show that using high quality welds will result in lower mass of the K-node.