On-Line Optimization and Control of an Hybrid Proton Exchange Membrane Fuel Cell System

ALEJANDRO MARCHETTI; DIEGO FEROLDI; MARTA BASUALDO

Mar del Plata, Buenos Aires

Congreso; 4º Congreso Nacional - 3° Congreso Iberoamericano de Hidrógeno y Fuentes sustentables de Energía - HYFUSEN2011; 2011

Instituto de Energía y Desarrollo Sustentable (IEDS) de la Comisión Nacional de Energía Atómica en colaboración con la Universidad Tecnológica Nacional

Proton exchange membrane (PEM) fuel cells are currently being considered as alternative powersources for vehicles. Because abrupt power changes may cause oxygen starvation and potentialdamage in fuel cells, these are usually combined with auxiliary power sources in vehicleapplications. Batteries and ultracapacitors for example, respond faster than fuel cells to suddenincreases in the current demand. Several energy management strategies have been proposed in theliterature to determine the power distribution between the fuel cell and the auxiliary source. Animportant challenge is given by the fact that the operation of the hybrid system must respect twoimportant operating constraints: (i) the state-of-charge (SOC) of the ultracapacitors must be keptwithin upper and lower limits; and (ii) the oxygen excess ratio of the PEM fuel cell system must bekept above a lower bound value in order to prevent oxygen starvation. Furthermore, wheneverpossible, it is desirable to take advantage of the available degrees of freedom in order to minimizethe overall fuel consumption.In this work, we investigate a power management strategy based on model predictive control(MPC). MPC has gained wide industrial acceptance because it can handle multivariable controlproblems with operating constriants. The manipulated variables are the fuel cell/ultracapacitorcurrent distribution, and the voltage applied to the compressor that is used to feed air into the fuelcell system. The SOC of the ultracapacitors is controlled within its limit values, and the fuelconsumption is minimized either by including a fuel cost term in the objective function of the MPCproblem, or by the inclusion of an upper level optimization problem which is executed on line. Theapproach is applied to a benchmark PEM fuel cell system with ultracapacitors as auxiliary powersource, using time-varying power load profiles resulting from standard driving cycles in vehicles.