Neural network approximation of iron oxide reduction process

The kinetics of Fe2O3 to FeO reduction process was investigated using the thermogravimetric data. The authors' previous experimental results indicated that initially the reduction of hematite is a surface controlled process, however once a thin layer of lower oxidation state iron oxides (magnetite, wüstite) is formed on the surface, it changes to diffusion control. In order to analyze the time-behavior of Fe2O3 reduction under various process conditions, artificial neural network (ANN) was tested for modeling of this complex reaction pathways. The data used included the reduction of hematite in various temperatures by CO, H2 and a mixture of CO and H2. The ANN model proved its applicability and capability to mimic some extreme (minimum) of reaction rate within specific temperature range, when the classical Arrhenius equation is of limited use.