With the escalating number of vehicles on the road, great concerns are drawn to the large amount of fossil fuels they use and the detrimental environmental impacts from their emissions. A lot of research and development have been conducted to explore the alternative energy sources. The fuel cell has been widely considered as one of the most promising solutions in automobile applications due to its high energy density, zero emissions and sustainable fuels it employs. However, the cost and low power density of the fuel cell are the major obstacles for its commercialization. This thesis designs a novel converter topology and proposes the control method applied in the Fuel Cell Hybrid Vehicles (FCHVs) to minimize the fuel cell's cost and optimize the system's efficiency. Unlike the previous work, the converters presented in the thesis greatly reduce the costs of hardware and energy losses during switching. They need only three Metal-Oxide-Semiconductor Field-Effect Transistors (MOSFETs) to smoothly accomplish the energy management in the cold start, acceleration, steady state and braking modes. In the converter design, a boost converter connects the fuel cell to the DC bus because the fuel cell's voltage is usually lower than the rating voltage of the motor. In this way, the fuel cell's size can be reduced. So is the cost. With the same reason, the bidirectional converter connected to the ultracapacitor works at the buck pattern when the power is delivered from the DC bus to the ultracapacitor, and the boost converter is selected when the ultracapacitor provides the peaking power to the load. Therefore, the two switches of the bi-directional converter don't work complementarily but in different modes according to the power flow's direction. Due to the converters' simple structure, the switches' duty cycles are mathematically analyzed and the forward control method is described. The fuel cell is designed to work in its most efficient range producing the average power, while the ultracapacitor provides the peaking power and recaptures the braking power. The simulation results are presented to verify the feasibility of the converter design and control algorithm.

Indiana University-Purdue University Indianapolis (IUPUI) / This research proposes state-of-the-art multilevel converter topologies and their modulation strategies, the implementation of a conventional flying-capacitor converter topology up to four-level, and a new four-level flying-capacitor H-Bridge converter confi guration. The three phase version of this proposed four-level flying-capacitor H-Bridge converter is given as well in this study. The highlighted advantages of the proposed converter are as following: (1) the same blocking voltage for all switches employed in the con figuration, (2) no capacitor midpoint connection is needed, (3) reduced number of passive elements as compared to the conventional solution, (4) reduced total dc source value by comparison with the conventional topology. The proposed four-level capacitor-clamped H-Bridge converter can be utilized as a multilevel inverter application in an electri fied railway system, or in hybrid electric vehicles. In addition to the implementation of the proposed topology in this research, its experimental setup has been designed to validate the simulation results of the given converter topologies.

No description available.

This thesis presents an active clamp forward converter for use in the Energy Harvesting From Exercise Machines project. Ideally, this converter will find use as the centerpiece in a process that links elliptical trainers to the California grid. This active clamp forward converter boasts a 14V-60V input voltage range and 150W power rating, which closely match the output voltage and power levels from the elliptical trainer. The isolated topology outputs 51V, higher than previous, non-isolated attempts, which allows the elliptical trainers to interact with a central grid-tied inverter instead of many small ones. The final converter operated at greater than 86% efficiency over most of the elliptical trainer’s input range, and produced very little noise, making it a solid choice for this implementation.