Neural Computing Strategy for Predicting Deactivation of Fischer-Tropsch Synthesis With Different Nickel Loadings

Direct determination of the process deactivation model relates to catalyst stability and plays a crucial role in the process control. The present study aims at investigating the influence of nickel loading (10-20 wt%) on the deactivation model parameters of Ni/Al2O3 catalyst prepared by incipient wetness impregnation. Artificial neural network (ANN) predicts a steady-state activity of the catalyst for the ultimate purpose of a deactivation model selection. The results obtained from an ANN demonstrated that the first-order general power law expressions (GPLE1 model) could adequately predict the catalytic activity during long reaction time. Considering various loadings of nickel on an alumina support, better stability of 20Ni/Al2O3 catalyst was confirmed. Model parameters affirmed that a decrease in the loading of the nickel-made active phase increases the deactivation rate of the catalyst. [GRAPHICS] .