| Abstract Scope |
Direct reduction of chromite (DRC) is a pilot-ready ferrochrome production method that enables metallization at 1300 °C, below conventional smelting temperatures (>1600 °C), reducing energy consumption and avoiding the need to melt refractory minerals. To support scale-up and commercialization of DRC, we developed a rapid, non-destructive method for 3D assessment of metallization and microstructure. In prior work, we developed a workflow combining CaCl₂-assisted TGA reduction experiments, X-ray micro-computed tomography (µCT), deep learning segmentation, and image analysis to extract phase distribution, morphology, porosity, and tortuosity at varied reaction extents. Here, we apply pretrained models to µCT scans from larger-scale DRC trials using matched ore, reductant, and flux, and identical scan settings. Metallization extent is inferred by comparison with reference datasets. This approach enables same-day product quality assessment in experimental campaigns where ferrochrome is produced by DRC, and it can be extended to development of other innovative ferroalloy production methods. |