Simulation and multiobjective optimization of an industrial hydrogen plant based on refinery off-gas

Abstract

A rigorous model is developed for simulating an existing industrial hydrogen plant based on refinery off-gas, which is made up of liquefied petroleum gas and off-gas from the membrane separation unit in a petroleum refinery. The presence of higher hydrocarbons in the reaction system is accounted for in the model equations for bulk gas and catalyst pellet. The reformer model is validated against three sets of industrial plant data, with good agreement. Thereafter, multiobjective optimization is performed using the nondominated sorting genetic algorithm to predict sets of Pareto-optimal operating conditions for improved performance. For a fixed feed rate of off-gas to the unit, two or three objectives, namely, maximization of product hydrogen and export steam rates and minimization of the heat duty supplied to the steam reformer, are targeted simultaneously. The optimal heat flux profile in the steam reformer is distinct from that predicted for methane feed (Oh, P. P.; Ray, A. K.; Rangaiah, G. P. J. Chem. Eng. Jpn. 2001, 34 (11), 1341. The optimal results obtained are better than industrial operating data.