Design and Optimization of Hydrogen-Resistant Steels Based on First-Principles Calculations and Machine Learning
摘要
The hydrogen embrittlementHydrogen embrittlement (HE) issue in ultra-high-strength steels has always been a hot topic of concern for researchers, particularly regarding the unclear influence of alloying elements on HE resistance. In this study, a high-precision machine learningMachine learning force field (MLFF) for the Fe–C–H system was constructed by combining first-principles calculationsFirst-principles calculation and crystal structure prediction (CSP) with machine learningMachine learning. Using this MLFF, molecular dynamics simulationsMolecular dynamics simulations were performed to investigate the diffusionDiffusion behavior of hydrogenHydrogen atoms in steels with different carbonCarbon contents, and the hydrogenHydrogen diffusionDiffusion coefficients were calculated. It was found that the hydrogenHydrogen diffusionDiffusion coefficient generally decreased with increasing carbonCarbon content, in good agreement with experimental results. The algorithm model established in this study can analyze the influence of carbonCarbon content on the hydrogenHydrogen resistance of iron and steel materials, which is of significant importance for studying hydrogenHydrogen-induced damage in steel materials and composition design.