Control of a Fuel Delivery System for Polymer Electrolyte Membrane Fuel Cells based on a Two-Phase Transient Model
Type of Degreedissertation
MetadataShow full item record
Polymer electrolyte membrane fuel cells (PEMFCs) are a type of fuel cell that converts the chemical energy released by the reaction of hydrogen (fuel) and oxygen into electrical energy and generates water and heat. The fuel delivery system (FDS) is designed to supply hydrogen from a storage tank to the fuel cell stack and, in some designs, reuses the exhausted fuel. In this research work, a hybrid FDS used in fuel cell vehicles is proposed, which uses an ejector and a blower dependent upon loads that circulate unconsumed hydrogen to increase efficiency of fuel usage. In addition, stoichiometric ratio (SR) of the hydrogen, defined as the ratio of the supplied hydrogen flow rate to the consumed by the reaction in cells, should be maintained a constant to prevent fuel starvation at abrupt load changes. Moreover, the hydrogen pressure imposed to the stack should follow any change of the cathode pressure to prevent large pressure difference across thin membranes. Furthermore, liquid water, impurities and contaminant species in the anode gas flow channels should be purged out in time to prevent flooding and catalysts poisoning in cells. Design of model-based controls for the FDS is a challenging issue. A transient two-phase model of a single cell was developed for the design of controls, which considered the phase changes of water and two phase flows in cells. The model was experimentally validated by a segmented single cell that allows for measurements of current distributions and visualization of liquid water in gas flow channels. The experiment results of I-V curves shown that the air humidity in gas flow channels had larger influence on the cell performance than the air flow rates did. The images of liquid distribution in the channels indicated that most liquid water was accumulated near the outlet of gas flow channels and the amount of liquid water in the channels was affected by the air humidity and flow rates. The I-V curves and liquid water amount variation in channels have the similar trend with the simulation results of the transient model of the single cell. An anode model of a stack including two-phase phenomena was developed based on the transient model of the single cell, which was integrated to a set of control oriented models of FDS components. The integrated model was analyzed and linearized to develop a state feedback controller with integral and observer (SFB), which was compared with other two classic controls such as the proportional and integral (PI) and static feed-forward (SFF) controllers. It was found that the FDS could not be stabilized because of the liquid water accumulation in the system and cells without purging process. A dynamic purging process based on the time integral of stack current was designed and implemented to control the liquid water amount in the system. The simulation results of SFB, PI and SFF controllers with FDS model shown that the SFB controller had the best tracking and rejection performance on the control of the supplied hydrogen pressure and stoichiometric ratio under the disturbance of step change of stack current and purging process. In the simulation results, the liquid water was found in the anode side of fuel cells and manifolds in FDS. The amount of liquid was also effectively limited in a small range to prevent flooding in FDS and cells.