In this paper, the crack stress intensity factor in weld toe of T- joint under uniaxial and biaxial load is studied. A series of geometries of semi-elliptical surface cracks were selected as elementary crack geometries, the appropriate element size of crack tips are selected according to the convergence results. The calculation results show that the SIF of crack tip is affected by several factors such as aspect ratio, crack depth and load form in different level. By the comparison of SIF results of uniaxial and biaxial load, a new magnification factor is proposed in this paper, which captures the influence of the biaxial load on stress intensity factor for T-joint.
T-joint is a common type of connection in metal welded structures, it has an extensive application in marine structures, and so how to evaluate the fatigue life of this kind of welded joint effectively has a significant meaning in the safety study of marine structures. Previous studies for welded joint have tended to focus on the relatively simple butt joint, while for the complex geometry form of T-welded joint, the fatigue crack propagation is more complicated. It is proved that the fatigue cracks mainly initiate at weld toes (V Balasubramanian, 1999; Wang W X, 2002; Zhang L, 2007), and to estimate the fatigue life and fracture performance, SIF of surface cracks should be calculated first.
The mainstream method to calculate SIF are analytical method, finite element method (FEM) and hybrid method etc.(Almar-Naess,1985). The stress field at crack tip for welded joint is highly nonlinear so that analytical formula can be used to calculate the SIF only in a few simple cases. The finite element method is another way to get the complicated SIF, while the low computational efficiency and complex model mesh hinder the progress. So, in engineering application, a practical and non-accuracy method is pursued. The hybrid method base on the superposition principle and Green function in linear elastic fracture mechanics. Firstly, taking an existing formula as a foundation, and then multiplying different correction factors to modify the formula.