Model predictive control for ethanol steam reformers with membrane separation

Abstract

This paper focuses on the dynamic modelling and the predictive control of an ethanol steam reformer (ESR) with Pd[sbnd]Ag membrane separation stage for the generation of pure hydrogen. Hydrogen purity necessary to feed a proton exchange membrane fuel cell (PEMFC) is required. A non-linear dynamic model of the ESR is developed together with a procedure for adjusting the model parameters in order to fit a bank of experimental data of a real ESR system. Static and dynamic analysis of the non-linear ESR model is presented. From this non-linear model, a linear, reduced order and discretised model is derived and a model predictive controller (LMPC) is designed for the ESR system. Control objectives are pure hydrogen flowrate tracking and ethanol inlet minimization. Comparisons between the non-linear and linear models are carried out to determine the control constraints. Finally, simulation results for the implemented LMPC controller are presented and discussed.