Artificial neural networks with response surface methodology for optimization of selective CO₂ hydrogenation using K-promoted iron catalyst in a microchannel reactor

CO₂ hydrogenation was optimized by a combination of AANs (Artificial Neuron Networks) with RSM (Response Surface Methodology) in a microchannel reactor using a K-promoted iron-based catalyst. This robust and cost-effective methodology was reliable to extensively analyze the effect of operating conditions i.e. gas ratio, temperature, pressure, and space velocity on product distribution of selective CO₂ hydrogenation. With experimental data as training data using ANNs and Box-Behnken design as design of experiment, the obtained model was able to present good results in a nonlinear noisy process with significant changes of critical operation parameters in an experimental design plan during CO₂ hydrogenation using K-promoted iron-based catalyst in a microchannel reactor. The achieved quadratic model was flexible and effective in optimizing either single or multiple objections of product distribution for CO₂ hydrogenation.