Modeling of NH3–NO–SCR reaction over CuO/γ-Al2O3 catalyst in a bubbling fluidized bed reactor using artificial intelligence techniques

Comparative study of the artificial neural network and mechanistic model was carried out for NO removal in a bubbling fluidized bed reactor. The effects of temperature, superficial gas velocity and ammonia/nitric oxide ratio on the NO removal efficiency were determined and their optimum conditions were estimated by the experimental study, the artificial neural network and mechanistic models as well. The optimum values of ammonia/nitric oxide ratio, temperature and superficial gas velocity for the maximum NO removal efficiency were found to be 1.5, 300 °C and 0.098 m/s, respectively. A mechanistic model was implemented in our previous study [Muhammad F. Irfan, Sang Done Kim and Muhammad R. Usman, 2009] and it was found that this model fitted well only at specific condition i.e. maximum conversion temperature (300 °C). However, it failed to perfectly match with rest of the experimental data points at other temperatures and parametric conditions as well. To improve this, an artificial neural network modeling strategy was applied and its predictions were evaluated which were favorably matched with the experimental data rather than the mechanistic model.