(19a) Modelling of Chemical Reactions in Metallurgical Processes

Developing ways for the direct reduction of iron ores has attracted much research interest in the last three decades, since it can be considered as a core process in steel industry [1]. New advances in the iron making technologies offer up potential savings in energy and emissions. One of the most advantageous direct reduction processes are fluidized bed reactors. However, due to the harsh conditions inside these reactors, most investigations are carried out through computational tools. One such tool is the CFD-DEM method, in which the gas phase reactions and governing equations are calculated in the Eulerian (CFD) side, whereas the particle reactions and equations of motion are calculated in the Lagrangian (DEM) side.

In this work, two of the most common types of models to represent the reaction between solid particles submerged in fluids are implemented into the CFD-DEM library. These models are the Shrinking Particle Model (SPM), where the solid particle reacts with the fluid thereby changing its size, and the Unreacted Shrinking Core Model (USCM), where after the particle reacts with the fluid a product layer is formed around the particle that impedes the reaction rate [2].

The communication framework between the CFD and DEM sides are initially verified with the use of SPM. A case where a single particle reacts with the incoming fluid is carried out and the species mass balances are verified. In the metallurgical process of iron-ore reduction the oxygen is extracted from the iron-oxide through intermediary steps, therefore the SPM becomes insufficient. In this case, the different reaction steps of iron-ore reduction needs to be considered which leads to using the three-layered USCM representation. The layers that need to be considered are the Fe (Iron-Layer), FeO (Wüstite), Fe₃O₄ (Magnetite) and Fe₂O₃ (Hematite), which is the innermost core of the particle. The gaseous reactant diffuses through the porous layers of iron shell, wüstite and magnetite layers, while simultaneously a portion of the diffused gas reacts at the layer interfaces. Finally, the reactant gas reaches the hematite core and reacts with the core surface producing magnetite.

For the investigation of metallurgical processes, both of the mentioned models is implemented into the CFD-DEM library in such a manner that only the required data is communicated between the two phases with an adaptable communication interval. A similar case as with the SPM is also carried out with the implemented USCM for an iron-ore. The overall reduction rates are validated by comparing them to available literature. After the initial validation cases with only a single particle, the model capabilities are demonstrated with a lab-scale fluidized bed where various reactant gases such as hydrogen, carbon monoxide and a mixture gas of H₂ and CO are utilized. The simulation results are in good agreement with the fractional reduction obtained from the experiment. These results give us further insights about the general behavior of the reducing and product gases, the gas ratio, and temperatures as well as the effects that particle properties such as the particle porosity, diameter and pore size has on the rate of reduction.