A process-cycle reactive CFD–DEM model for simulating long-duration operations of hydrogen-based direct-reduction shaft furnace

Direct-reduction shaft furnaces (SFs) are sensitive to burden heterogeneity, particularly under hydrogen-based operation because H 2 reduction is strongly endothermic. Resolving the resulting non-uniformity requires particle-scale transport and reaction histories over hours-long operation, which remain difficult to obtain experimentally or computationally. This study develops a process-cycle reactive CFD–DEM model for long-duration SF operation from charging to discharge, accelerating burden passage while preserving particle reduction histories. Verification over scaling factors of 10–400 shows close agreement with unscaled reduction trajectories, while validation against two experiments reproduces axial trends in reduction degree, gas composition, and temperature. Long-duration simulations capture startup, burden renewal, and established operation. Monodisperse simulations show that contact conditions modify bed structure, radial gas supply, and particle descent, producing a persistent under-reduced core under higher friction. For polydisperse burdens, contact conditions also govern size segregation and particle residence histories, linking product heterogeneity to coupled gas–solid transport and reaction characteristics.