Seminars

Global DRI production has increased by 25.6% over the past five years. Hydrogen-based flash steelmaking offers energy savings of about 30%. However, existing approaches oversimplify key grain-scale phenomena such as mass transport, reaction kinetics, microstructural evolution, and surface chemistry. Here we developed the Transient Reactive Grain Model (TRGM), which captures processes from micron-sized grains down to reactive nanopore surfaces, to solve the above-mentioned problem. Results highlight diffuse reaction zones, enhanced hydrogen transport from evolving porosity, and surface-dependent radical adsorption. However, deterministic models may cause overfitting problems. Therefore, we then extend TRGM with Uncertainty Quantification (UQ) to account for model and parameter variability. Monte Carlo propagation identifies influential parameters, generates prediction intervals, and shows that TRGM with UQ matches experiments while providing confidence bounds on reduction kinetics. Together, TRGM and UQ establish a robust, uncertainty-aware modeling framework to predict H plasma and molecular H2-DRI kinetics, optimize hydrogen utilization, and guide the design of flash steelmaking technologies.