Control and Analysis of Air, Water, and Thermal Systems for a Polymer Electrolyte Membrane Fuel Cell
Type of Degreedissertation
MetadataShow full item record
The polymer electrolyte membrane (PEM) fuel cell is a power source that can potentially replace the internal combustion engine in vehicles of the future. When hydrogen stored in a tank and oxygen from the air chemically react in a PEM fuel cell, electricity is generated and water and heat are produced as by-products. The management of fuel, water, and heat are crucial issues in order to ensure reliable operation and to maintain high efficiency with continuously changing loads. During the operation of a fuel cell, oxygen from the air is supplied to the cathode side of the fuel cell. Insufficient oxygen supply during dynamic loads causes oxygen starvation, which forces protons transported through the membrane to reduce the amount of hydrogen on the cathode catalysts. Reduced hydrogen could chemically generate heat on the platinum particles and result in local hot spots in the membrane electrode assembly (MEA), which could lead to failure of the fuel cells. Conversely, an excessive oxygen supply increases the parasitic power dissipated by an air supplier such as a blower, which leads to low efficiency of the overall system. Therefore, an optimal supply of oxygen is one of challenging issues addressed by researchers. When a chemical reaction takes places, water is produced in the cathode catalyst. A fraction of the water moves from the catalyst though the membrane and to the anode. In contrast, protons crossing the membrane take up water from the anode to the cathode. The water content in membranes directly affects proton conductivity. Insufficient water causes dehydration that decreases proton conductivity and increases voltage loss, which leads to a reduction in output power. However, the water produced by chemical reactions causes flooding in gas flow channels, gas diffusion layers, and catalyst layers, which blocks oxygen transport to the catalysts and reduces the catalyst activation areas. As a result, the degradation of components is accelerated. Therefore, excessive water in the cell should be removed to prevent flooding, and at the same time the membrane should be kept fully humidified. When operating, heat produced in the stack continuously changes as the load current varies. Consequently, the temperature inside the cells also varies. Variation of the temperature directly affects rates of chemical reaction and the phase change of water, and finally the water transport. Improper rejection of the heat might produce local hotspots and degrade the thin layers of the cell. Low temperatures decrease the rate of the chemical reaction and reduce efficiency. Conversely, elevated temperatures increase the reaction rate, ease the removal of water, and increase the mobility of water vapor in the membrane, which alleviates over-potential losses. In addition, parasitic power necessary for operating a coolant pump should be reduced in conjunction with the heat rejection strategy, which contributes to an increase in system efficiency. Therefore, thermal management is another challenging issue for reliable and efficient operation. The research conducted for my doctoral program focused on the development of fuel cell models for air, water, and thermal systems, and on the related components. The control-oriented models were used in the design of control strategies and in the analysis of integrated systems. The air and water supply systems consist of a blower, a humidifier, and inlet and outlet manifolds. The thermal management system is composed of a bypass valve, a liquid-to-liquid heat exchanger, a radiator with a fan, reservoirs, and pumps. The objectives of the proposed control strategies are to prevent oxygen starvation, maintain a proper water balance in the cells, reject excessive heat in the stack, and at the same time the parasitic powers are minimized. Considering the aforementioned system configurations and objectives, state feedbacks with integral controls were designed and optimized. The entire system was simulated and analyzed, and the resulting static and dynamic behavior obtained by experiments is described.