Dynamic Modeling, Control And Optimization Of PEM Fuel Cell System For Automotive And Power System Applications

This dissertation is focusing on fuel cell dynamic modeling and control, the optimization design of fuel cell systems, power electronics interface designs, and control for the fuel cell based hybrid system. First, a dynamic PEM fuel cell model is proposed as a nonlinear, multiple-input, multiple-output (MIMO) system so that feedback linearization can be directly utilized and the PEM fuel cells can be protected through the controller. For the design of the dynamic model PEM fuel cells, all possible water effects are considered, and the anode and cathode gas pressures, as well as the fuel cell voltage are defined as the control objectives. In terms of optimization of the fuel cell systems, the efficiency and the cost models of fuel cell system have been optimized under various operating conditions using a multi-objective optimization technique, the SQP (sequential quadratic programming) method. Second, due to a slow dynamics of fuel cell system by nature, it is recommended to have an additional auxiliary power such as ultra-capacitors or battery during transients to improve system performance in stationary power and transportation applications. Using ultracapacitors and bidirectional converter with the fuel cell system can be a possible solution to the problem of slow dynamics of the fuel cell system. In this research, we present a combined small signal ac equivalent circuit model consisting of bidirectional converter, PEM fuel cell, and ultracapacitor for the purpose to design an appropriate controller for the bidirectional converter during buck and boost mode. Transient performance of the fuel cell based hybrid system including fuel cell model, ultracapacitor and bidirectional converter is simulated. At last, a supervisory control strategy for a hybrid power system consisting of fuel cell, solar cell and energy storage is proposed. Using ultra-capacitors and photovoltaic (PV) panel together with a fuel cell system is another feasible consideration to provide a sufficient power supply for residential areas. Four modes of operations of this hybrid power system are defined for the proposed supervisory control. The proposed hybrid power system is simulated and analyzed based upon the supervisory control algorithm in Matlab/Simulink environment. Therefore, this dissertation will be a good platform for modeling,control and optimization of the fuel cell system and its applications such as hybrid power systems.