Predicting intergranular stress corrosion cracking of stainless steels in high-temperature water by incorporating crystallographic factor

This study systematically explores the influence of grain boundary (GB) characteristics on the intergranular stress corrosion cracking (SCC) against high-temperature water. Statistical analyses of GB cracking susceptibility were conducted using stainless steel specimens produced via conventional and additive manufacturing (AM) methods, subjected to various post-treatment conditions, encompassing over 12,000 GBs. Based on the crystallographic statistics, a preliminary investigation was preformed within the latent space to identify the metrics responsible for SCC nucleation. Contrary to conventional perspectives, corrosion-related parameters such as GB plane orientation and GB atomic packing density (GBAPD) are found to be less significant factors. Instead, mechanical parameters, notably GB normal strain/stress and slip discontinuity, emerged as critical factors. Specifically, the Luster-Morris m^′ factor demonstrated strong correlation with GB cracking susceptibility, whereas traditional Schmid factor variations showed minimal association. Lower m^′ values significantly intensified local strain/stress gradients by promoting slip discontinuities, thereby increasing susceptibility to SCC. A machine learning model was developed incorporating key GB parameters, achieving an accuracy of 85 % and demonstrating robust predictive capability across distinct microstructures.