Investigation of water gas-shift activity of Pt-MOx-CeO2/Al2O3 (M = K, Ni, Co) using modular artificial neural networks

The water gas shift activity of promoted Pt-CeO2/Al2O3 catalysts were investigated in this work. The catalysts were prepared by incipient to wetness impregnation and tested using a microflow reaction system. It was found that K has beneficial effects under product-containing feed compositions while Co and Ni promoters worsen catalyst performance. The reaction temperature and feed H2O/CO ratio positively affect the catalytic activity, whereas CO2 and H-2 addition to the feed decreases CO conversion, as expected. The experimental results were also modeled using modular neural networks, at which the catalyst preparation and operational (reaction) variables were used together in the same network because they are interacting but processed differently because they are dissimilar in their form (i.e. categorical versus continuous) and their effects on catalytic activity. It was concluded that the effects of catalyst preparation and operational variables and their relative importance could be comprehended more accurately by using this approach, which may be also employed in other similar systems. Copyright (C) 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.