This is the authors' abstract. We don't add key points for this paper.
The corrosion behavior of weathering steels in marine atmospheric environments emerges from the synergistic interactions among alloy chemistry, exposure duration, and environmental loading, exhibiting strong nonlinearity and multiscale coupled responses. Based on 708 sets of outdoor exposure data, a machine learning model is developed to predict corrosion behavior, and the effects of Cr-Cu combinations are systematically investigated under high chloride conditions through model interpretation and field validation. Key features and environmental variables are identified via data preprocessing and feature selection. Among six compared models, GBDT demonstrates the best predictive performance on both test and external validation datasets. Subgroup analysis associates higher Cr and Cu contents with lower corrosion loss under high-chloride conditions. Outdoor exposure of WS over 24 months provides a further assessment of the model predictions and documents the evolution of Cr- and Cu-containing corrosion products in its rust layer. This study integrates machine learning prediction, interpretability analysis, and field validation to examine the association between Cr-Cu composition and corrosion loss in marine atmospheres, providing guidance for compositional optimization of weathering steels.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Yin et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: