Welded joints are important for fatigue strength evaluation of ship and offshore structures. However, current techniques for fatigue evaluation of welded joint under variable load is not accurate enough. Also, it cannot consider the effect of load history which is one of the important features for the variable loads. Therefore, many experimental attempts are conducted for storm model to consider the variable loading. However, studies of storm loading usually ignore the effect of calm sea loading which constitute a large portion of the marine phenomena. Because it has been believed that the contribution of calm sea loading is not dominant for fatigue life in storm loading. In this paper, fatigue tests are conducted for the specimens with transverse attachment made of high tensile steel under variable amplitude axial loading based on storm model. Considered loadings include repeated single storm, 6 or 8 kind storms sequenced randomly, and storms including calm sea condition while the mean stress and the maximum stress of loadings are changed. Moreover, the effect of three variables are investigated; 1) root mean square (RMS) value of stress amplitude, 2) mean stress shift and 3) maximum stress which can characterize the storm loading on fatigue life. In addition, experiments with calm sea loading are conducted and the effect of calm sea loading is also investigated. The storm and calm sea loadings are generated from IACS-34 wave scatter diagram. 5% strain drop criteria is introduced to define crack initiation life. Experimental results including the test results from previous study are evaluated and compared with DNV-CN 30.7 (2005) and Matsuoka’s method for the estimation of crack initiation and propagation life. From the result, it is concluded that the fatigue strength under storm loading can be evaluated by RMS value of stress amplitude. And mean stress shift is more likely to relate to fatigue strength than maximum stress. The effectiveness of the calm sea loading is depend on the existence of mean stress shift. Regarding fatigue life evaluated by DNV and Matsuoka method, both of them have almost same accuracy.

This content is only available via PDF.
You do not currently have access to this content.