A theoretical and experimental study of glycerol steam reforming over Rh/MgAl2O4 catalysts

Abstract

In this paper, the back propagation artificial neural network (ANN) modeling approach was used for investigating of the steam reforming reaction of glycerol over Rh/MgAl2O4 catalyst in a packed bed tubular reactor. The experimental tests were carried out in order to provide appropriate experimental data set for training the model as well as for evaluating the ANN prediction ability of the catalytic glycerol steam reforming process. In this regard, important parameters in steam reforming process such as gas hourly space velocity (GHSV) between 35,000 and 70,000 ml·h−1 g−1, reaction temperature from 300 °C to 600 °C, water to glycerol ratio (feed ratio) from 3 to 9 and time on stream from 20 to 300 min were studied from experimental as well as modeling outlook. The obtained ANN results were in very good agreement with empirical values. The ANN model predicted the experimental results of glycerol steam reforming with high correlation (R = 0.9995) and very low error (MSE = 1.2317×10−4) for training data.