Model predictive control of a combined electrolyzer-fuel cell educational pilot plant

Abstract

In today's era of renewable energy, hydrogen fueled proton exchange membrane (PEM) fuel cells are considered as an important source of clean energy. As the technology is emerging fast, many universities and colleges have adopted fuel cells in their educational program. In this paper, we will present the modeling and control of the fuel cell pilot plant present in Clean Energy Trainer, which is used by students and researchers in many universities. The plant under consideration is a laboratory-scale pilot plant designed mainly for verifying the applicability of theoretically studied control strategies on the real-world application. The plant is a series connection of electrolyzer and a PEM fuel cell stack with one input and one output. The control of such a plant is the challenging research problem due to the nonlinearities, slow dynamics, dynamics and physical constraints. The control oriented data-driven model of the plant is developed and validated through a series of experiments. To tackle the electrolyzer-fuel cell control problem, we present a model predictive control (MPC) scheme that can take into account the physical constraints of the plant. In addition to the controller, a disturbance observer is designed to cope with the external disturbances and to avoid adverse effects on the system performance. Subsequently, the developed control scheme is successfully implemented in realtime. Highly satisfactory results are obtained, regarding reference tracking, constraint handling, and disturbance rejection.