Hydrogen diffusion in thick-walled S690 SAW joints: part 2 – predictive modeling of heat input and microstructure influence

Abstract

High-strength low-alloy (HSLA) steels such as S690 are widely employed in thick-walled welded structures, where hydrogen-assisted cold cracking (HACC) remains a persistent concern. While microstructure-specific hydrogen diffusion coefficients (D_H) for weld metal (WM), heat-affected zone (HAZ), and base material (BM) were experimentally established in Part 1 of this study, their quantitative influence on hydrogen accumulation and effusion has not yet been fully clarified. This work presents a transient, spatially resolved numerical model for simulating hydrogen transport in multi-pass submerged arc welds. The model integrates experimentally determined D_H values with realistic thermal cycles and temperature-dependent boundary conditions. Developed in Python, the simulation tool is purpose-built for hydrogen diffusion in welded joints. It offers a focused, transparent alternative. It offers a focused, transparent alternative to general-purpose finite element platforms. Parametric analyses demonstrate that, although the diffusion coefficients vary by up to 50%, their impact on overall hydrogen retention is minor. In contrast, plate thickness, bead geometry, cooling time (t₈/₅), and interpass temperature exert a dominant influence on hydrogen distribution. Despite clear microstructural differences between the thermomechanically rolled (S690MC) and quenched and tempered (S690Q) variants, including opposite HAZ hardness responses (softening in S690MC, hardening in S690Q) in the (pen)ultimate weld bead, the simulations confirm that their diffusion behavior and hydrogen solubility are closely aligned. Consequently, differences in D_H and solubility exert only a minor influence on hydrogen retention compared to thermal exposure and joint geometry. These findings support the interchangeable use of both steel grades in terms of HACC risk due to hydrogen diffusion kinetics under comparable welding conditions.