Crack Growth of Asymmetric Geometries and Non-Uniform Wall Thickness in Pressure-Containing Bodies Under Cyclic Loadings

Abstract The known art and computational simulation of crack initiation and propagation (FCGR) is understood and predictable for symmetric geometries with constant wall thicknesses, such as spheres cylinders and plates. Axes of principle stress and perpendicular locations of stress intensity at crack “tips” stay true on course over the cyclic life until the crack’s size overtakes the wall thickness, or the biggest value of ΔK, that does not cause any failure or show evidence of crack growth in 108, is considered as fatigue threshold. However, understanding cyclic loading acting on less symmetric bodies, such as odd-shaped components or varying wall thickness locations become more problematic. Changing geometries may lead to shifts in orientation and direction in primary stress axes as load paths create uneven stress fields and stress intensity through the body. The growing crack may introduce additional changes in geometry and subsequent load paths; all leading to stress intensity energy being directed toward or away from a crack tip, feeding or starving the crack’s continued growth, under cyclic loading. Predicted results may become more skewed (overly conservative) as geometries digress from simple failure models (and their formulas) used to calculate estimated fatigue life. This study investigates crack propagation within asymmetric geometries and non-uniform wall thickness in pressure-containing bodies subjected to internal and external pressures plus external mechanical loads, both mean and cyclic (such as typically found in subsea hardware). The dynamic relationship between changes in the amplitude and direction stress field axes, the crack field intensity with updated predictions of crack size growth and migration direction is analyzed to understand the difference in fatigue life estimates between current practice models and newer computer algorithms. Finite element simulation with a 3D extended finite element model (XFEM) focused on the simulating crack growth behavior is employed by reevaluating mesh stresses as the crack grows and migrates through the “irregular” wall. Results should provide a more rigorous (and often less conservative) method for prediction of fatigue life in asymmetric geometries and suggest insights on modifying geometry or material use to improve cyclic life.