Quantitative Evaluation of Sinter Reducibility Under Simulated Blast Furnace Conditions Using Microstructure Estimated by Hyperspectral Imaging

Precise control of sinter reducibility is essential for stable blast furnace operation. Each mineral phase present in sinter, such as hematite, magnetite and calcium ferrite exhibits different reducibility. In XRD analysis, the requirement for sample pulverization leads to the loss of mineralogical texture information. This makes it difficult to quantitatively correlate the complex mineral phases present in the sinter with reducibility. This study introduces a novel quantitative approach using hyperspectral imaging to distinguish specific mineral morphologies. Reduction experiments simulating blast furnace thermal and gas conditions were conducted on several sinters. Multiple regression analysis was applied to correlate mineral fractions and macroporosity with reduction rates across three distinct reduction stages. In the low-temperature stage, hematite, macroporosity and acicular calcium ferrites were identified as the primary drivers of reduction. In the intermediate stage, acicular calcium ferrites continued to enhance reactivity, whereas coarse calcium ferrite showed a significant negative influence. In the high-temperature stage, macroporosity strongly promoted reduction, while coarse calcium ferrite and magnetite hindered it due to the formation of shell-like metallic iron structures which impede gas diffusion. These findings demonstrate that hyperspectral imaging combined with multi-stage regression analysis offers a useful tool for designing optimal sinter mineralogy for blast furnace performance.