Abstract-This paper presents a new Nonlinear Model Predictive Control (NMPC) design for an Ethanol Steam Reformer with Pd-Ag membrane separation stage. The reformer is used to produce pure hydrogen able to feed a Proton Exchange Membrane Fuel Cell. Mass and energy balances are used to obtain the nonlinear dynamic model of both the reforming and the separation stages. Constraints, system nonlinearities and flexible cost function are the main reasons to select an NMPC controller, which is tested against the ordinary differential equations as simulation model, and has an internal model based on the sample data technique.

The attempt to improve hydrogen quality has motivated research focused solely on the design of reactors for hydrogen production, paying little interest to their modelling and much less attention to its dynamic analysis. For these main reasons, this paper is focused on the dynamic analysis of a mathematical model of an ethanol steam reformer (ESR) for producing hydrogen. This analysis differs from other nonlinear analyses since it is applied to a specific seventh-order system, while literature presents analyses for at most fourth-order systems. The nonlinear model used is based on mass balances. It is represented by partial differential equations (PDEs), which are converted into ordinary differential equations (ODEs) by the finite-differences method over the space. A general nonlinear dynamic analysis based on equilibrium points and their local stability is carried out by using these ODEs. Dynamic analysis results show that the studied reactor has a stable equilibrium point for a physical-sense range of inputs (ethanol + water), which allow to have available key criteria for the design of control strategies.