Dynamic model of an industrial steam reformer and its use for multiobjective optimization

Abstract

An industrial side-fired steam reformer in a hydrogen plant is simulated under dynamic conditions. A rigorous model with verified reaction kinetics is used. This model incorporates aspects of heat transfer in the furnace and diffusion inside the catalyst pellet. The dynamic model is used to study the effects of a few (planned) disturbances that reduce the production of both hydrogen and steam and adversely influence the safety of the plant if corrective control action is not taken in time. The operation of the steam reformer is simulated in the presence of three idealized disturbances in (1) the inlet feed temperature, (2) the inlet feed rate of natural gas, and (3) the furnace gas temperature. The model is then used to obtain optimal operating conditions required to negate the effects of two disturbances using several control or decision variables. Two objective functions are minimized simultaneously: the cumulative (integrated over time) deviation of the flow rate of hydrogen and the cumulative deviation of the steam flow rate. The elitist nondominated sorting genetic algorithm NSGA-II is used to obtain solutions of this multiobjective optimization problem. Nondominating Pareto-optimal solutions are obtained.