The steel industry provides large quantities of raw material for the society at the cost of large emissions of CO₂, however, and thus contributing to the global warming. Therefore, it is imperative for the steel industry to undergo a transition to low-CO₂ ironmaking technologies within the next 5-10 years. In order to reach sustainable ironmaking, recycled hydrogen as a reducing agent is the most promising path to follow [1], although the traditional blast furnace (BF) and its H₂-intensive variant next to Midrex and Energiron will still play an important role during the transition process. However, independent of the applied technology, it is common understanding that a thorough knowledge of the governing process parameters is critical to the design, control and optimization. Experimental investigations are very limited due to harsh operating conditions and involved costs so that innovative simulation technologies are more than welcomed to gain a deeper insight into the underlying physics of these processes, for which the extended discrete element method (XDEM) has crystallised as the most promising road to follow [2]. It treats the iron bearing material and coke as a discrete phase with its descend in the reactor and thermodynamic state i.e. drying, reduction through carbon monoxide and hydrogen and melting of ore particles and coke oxidation while a coupling to traditional computational fluid dynamics (CFD) describes the fluid and thermodynamics in the void space between the particles including an exchange of momentum, heat and mass between the gas and particles. This versatile and unique technology allows applying XDEM to a large variety of ironmaking reactors to even represent them as digital twins and to support the transition to “green” steel. A thorough analysis of predicted uncovers the hidden complexity and contributes significantly to a deeper understanding and thus, promotes strongly a shift from current empirical-based practice to a truly advanced multi-physics simulation technology. Hence, information from digital twins is already available at an early stage of design avoids costly re-designs. Problems are identified before they even occur and down-times are reduced. These findings support decision-makers and help them to make informed decisions. Consequently, a faster time-to-discovery and time-to-solution is obtained.