Fatigue strength assessment of a butt-welded joint in ship structures based on a time-domain strain approach is performed in this study. The service life load histories applied to the butt-welded joint located on the deck of a bulk carrier are generated, accounting for the still-water and wave-induced loads. The rainflow counting method is applied to analyze the load histories, and the long-term distributions of the load range are compared with those based on the conventional spectral fatigue analysis. An approach of converting the load history to a series of closed notch stress-strain hysteresis loops and several open notch stress-strain hysteresis curves is proposed and demonstrated under variable amplitude loading. The approach is based on analytical notch stress-strain estimations and consists of several steps to consider the material memory effect, overcoming some limitations of the existing methods. To determine the fatigue damage for the variable amplitude loading, a design fatigue curve is derived considering the uncertainty in the fatigue lives and load sequence effects. The intrinsic fatigue limit concept is used to filter the small amplitude cycles that do not have a damaging effect. The fatigue strength of the butt-welded joint is analyzed, taking the weld-induced residual stress and misalignment effects into account explicitly. The notch mean stresses or strain amplitudes of the cycles are significantly enhanced because of the presence of a high level of weld-induced tensile residual stress or misalignment, resulting in highly severe fatigue damage.
1. Introduction
Complex ship structures containing geometrical and material discontinuities are prone to fatigue because of cyclic loads. Therefore, fatigue strength assessment has been an important criterion in the ship structural design (Guedes Soares & Moan 1991). Various fatigue design concepts for the assessment of welded joints, where fatigue failures mostly originate, are applied (Xu 1997; Radaj et al. 2006), and they can be classified into two types. The first one is based on S-N curves in combination with the Palmgren-Miner rule, and the second one is based on the crack propagation models and failure criteria.
