Observer Based Fuel Delivery Control For PEM Fuel Cells

Undoubtedly,the increasing demand of energy and environmental concerns are the mostly cited significant issues in the 21st century.Therefore,hydrogen has drawn attention widely from both academia and industry due to its advantages such as exclusive abundance,pollution-free,and the highest energy density among all traditional fuels.The Proton Exchange Membrane(PEM)fuel cells is regarded as a potential hydrogen energy conversion plant for their high efficiency and low operation temperature.However,the practical PEM fuel cell system is a complexed interdisciplinary system which exists complex characteristics,such as strong nonlinearity and multi-physical field coupling with subsystems.These characteristics of PEM fuel cells are one of the difficulties in the controller design.The effective control strategies can not only improve its work efficiency,but also prolong the service life of the fuel cells.The other problem of the controller design is that it is expensive to measure the gas pressure in stack and the partial pressures in the fuel delivery system which are then supposed to be unknown in the controller design.In the anode of PEM fuel cells,the measurable outlet pressure is usually different with the actually utilized stack pressure.In this paper,the anode of PEM fuel cells is divided into three segments to reduce the modeling error between these two pressures and the impact of the spatial distribution of pressure,electrochemical reaction,and gas permeation.Specifically,we first model the fuel delivery system with anode recirculation and bleeding in a MIMO nonlinear compact form based on the segmented anode model.Then,a Lyapunov-based full state feedback controller is proposed based on the developed model to guarantee adequate hydrogen supply and maintain the anode hydrogen at a suitable concentration level.Moreover,a high order sliding mode observer is proposed to estimate the unmeasurable pressure and the partial pressure in the fuel delivery system with known disturbance.Furthermore,an output feedback controller based on the proposed observer is developed with guaranteed stability.Simulation results illustrate the effectiveness of the proposed control and the estimation approaches.