Finite Element Modelling of Transgranular Chloride Stress Corrosion Cracking in 304L Austenitic Stainless Steel

Austenitic stainless steels (ASS) have excellent resistance to general corrosion. However, these steels can be susceptible to localised corrosion such as pitting and crevice corrosion. In the presence of a tensile stress they can also exhibit stress corrosion cracking (SCC). In pressurised water reactor (PWR) nuclear plant incidents of SCC, especially chloride-induced SCC (CISCC), have been observed. Chloride ions which can lead to CiSCC of even low carbon austenitic grades can be introduced from many sources including the atmosphere and materials introduced into the reactor environment. Stress can result from primary loading or introduced as secondary stresses, such as residual stress, through machining or welding processes. Residual stresses are internal self-balancing stresses that can act alone or together with a primary stress to cause premature failure of a component. 15 mm lengths of 304L ASS tube were subjected to an in-plane compression of between 1–10 mm before unloading. This created regions of plasticity and on relaxation the specimen contains a complex state of residual stresses that can be modelled by finite element (FE) methods. The tube specimens were then boiled in MgCl₂ for 14 days before metallographic examination. A FE model of transgranular CISCC has been created by writing a VUMAT user subroutine implemented into the commercial FE code ABAQUS. The model is based on simple rules which include the initiation of surface corrosion pits from which, under mechanical control, SCC cracks may propagate. The model includes rules for SCC growth, based on hydrostatic stress state, and can incorporate the idea of grain orientation effects. Cracks created interact with and modify the residual stress field in the tube. Test results were then compared with model outputs. Crack morphologies and to a certain extent crack positions matched well with experiment. Attempts were made to calculate the crack tip driving forces from the model. The results also highlight the need to consider the importance of triaxial stress states, created by pits and cracks, and stress as a tensor rather than a scalar property. The effect of grain misorientation is also investigated, but so far, found to be of more limited importance for modelling transgranular CISCC.