Title: Prediction of research octane number in catalytic naphtha reforming unit of Shazand Oil Refinery

Authors: Abdolreza Moghadassi; Alireza Beheshti; Fahime Parvizian

Addresses: Department of Chemical Engineering, Faculty of Engineering, Arak University, Arak 38156-8-8349, Iran ' Department of Chemical Engineering, Faculty of Engineering, Arak University, Arak 38156-8-8349, Iran ' Department of Chemical Engineering, Faculty of Engineering, Arak University, Arak 38156-8-8349, Iran

Abstract: In this work, an artificial neural network model was developed to predict the research octane number of an industrial catalytic naphtha reforming unit. Needed date set was provided from Shazand Oil Refinery which was continuously measured. The H2/HC ratio, feed flow rate, pressure, specific gravity and ASTM D-86 distillation data of the feed stream were considered as input variables. Various neural networks were trained and tested in order to find the best ANN model. A three-layer network including two hidden layers was found with minimum mean square error of 0.28 for testing. Comparison between estimated and experimental values of octane number exhibits good agreement. The results show the capability of ANN model to predict the octane number of a typical catalytic naphtha reforming unit with an acceptable error. Therefore, developed ANN model can be applied in other similar units for octane number prediction.

Keywords: artificial neural networks; ANNs; catalytic naphtha reforming; CNR; research octane number; RON prediction; modelling; oil refineries; Iran.

DOI: 10.1504/IJISE.2016.077696

International Journal of Industrial and Systems Engineering, 2016 Vol.23 No.4, pp.435 - 447

Received: 24 May 2014
Accepted: 19 Oct 2014

Published online: 13 Jul 2016 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article