Comparative studies of fracture models for marine structural steels

Smooth and notched round specimens and smooth flat specimens are prepared from two steels, API-2W50 and DnV-DH32TM. From incremental tensile tests for smooth specimens, average true stress–logarithmic true strain curves are obtained. By using Bridgman's formula for round specimens and Choung's formula for flat specimens, average true stress is rectified. Then, tensile tests for notched specimens with various notch radiuses are carried out and damage tests are performed to identify material parameters of the damage mechanics model. Three yield models are taken into account: a shear fracture model based on a von Mises's yield function, a linear damage mechanics model based on Lemaitre's law of damage evolution, and a porous plasticity model based on Gurson's yield function. While the shear fracture model can forecast the tested plastic deformation path including final fracture for smooth specimens, it unrealistically estimates that the hydrostatic stress at the necking section largely contributes the plastic behaviors for notched specimens. The linear damage mechanics model assumes linear relationship between the damage value and plastic strain, and fails to accurately estimate the test results. The porous plasticity model presents very good agreement with experimental results regardless of the notch radiuses as long as material parameters are reasonably chosen.