About this Abstract |
| Meeting |
2027 TMS Annual Meeting & Exhibition
|
| Symposium
|
Fundamentals of Sustainable Metallurgy and Materials Science
|
| Presentation Title |
Physics-constrained Constitutive Learning of rate-limiting timescales for efficient Hydrogen-based Direct Reduction for Green Steel Making |
| Author(s) |
Anurag Bajpai, Barak Ratzker, Dierk Raabe |
| On-Site Speaker (Planned) |
Anurag Bajpai |
| Abstract Scope |
Hydrogen-based direct reduction is central to low-carbon ironmaking, but its industrial efficiency is limited by conversion-dependent kinetic retardation during late-stage metallization. We present a physics-constrained, conversion-resolved constitutive framework that deconvolutes measured reduction trajectories into effective reaction and internal-transport timescales, then maps their dependence on operating conditions, pellet architecture, and composition using scientifically constrained additive modeling. The framework converts thermogravimetric trajectory information into interpretable constitutive maps, symbolic relations, and regime boundaries for hydrogen reduction of iron oxide pellets. The analysis shows that temperature and hydrogen partial pressure primarily shorten early reaction-controlled conversion, whereas late-stage reduction is governed by pellet-to-pore length scale, porosity, and tortuosity. Internal diffusion dominates the incremental time budget at intermediate-to-high conversion, and the reaction-to-diffusion transition shifts systematically with evolving porous architecture. These experimentally anchored relations identify stage-specific rate limitations and provide mechanistic design rules for pellet architecture and process selection in efficient green ironmaking at industrially relevant conditions. |
| Proceedings Inclusion? |
Undecided |
| Keywords |
Computational Materials Science & Engineering, Sustainability, Machine Learning |