Dehydrogenation of isobutane to isobutene has recently received considerable attention because of the increasing demand for Methyl tert-butyle ether (MTBE) and Ethyl tertiary butyl ether (ETBE), which is used as additives for gasoline to increase the octane number and to substitute Lead. In this research, by using bench scale fixed bed reactor existing in BIPC, conversion of isobutane and the effective parameters on it, such as pressure, temperature, concentrations of feed,…has experimentally investigated. Due to simple geometry and small reactor diameter with respect to its length, one-dimensional model was developed to simulate fixed bed reactor for dehydrogenation of isobutane by using mass and energy conservation laws. Inter and intraparticle resistance was being neglected due to small diameter of catalyst particle. The 4 th Rung Kutta approach was used to solve governing equation including the feed and production concentration, gas temperature and pressure drop. These set of equations have been solved by programming in MATLAB. In the next step, artificial neural network is used to predict conversion of isobutane in fixed bed reactor. The experimental data have been used to create a GA-ANN model. 2 neurons for the hidden layers have been achieved by trial and error method and acceptable results are obtained. In order to increase the efficiency of neural networks, genetic algorithm is used to optimize the parameters of neural network. By increasing 150 degrees of centigrade in temperature in both experimental and mathematical model results, observed that conversion of isobutane increased about 80%. Also by decreasing of pressure and increasing in hydrocarbon to hydrogen fraction, conversion of isobutane increased. The results from the model are in good agreement with the experimental data. Finally, comparison of the results of the GA optimized ANN model with analytical model and experimental data have been done. Keywords : Isobutane, Dehydrogenation, Fixed Bed Reactor, Neural Network, Genetic Algorithm
