An Intelligent Control System for a Hybrid Fuel Cell with Gas Turbine Power Plant

Open Access
Yang, Wenli
Graduate Program:
Electrical Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
July 15, 2009
Committee Members:
  • Kwang Yun Lee, Dissertation Advisor
  • Kwang Yun Lee, Committee Chair
  • William Kenneth Jenkins, Committee Chair
  • Jeffrey Scott Mayer, Committee Member
  • Robert M Edwards, Committee Member
  • optimization
  • system modeling
  • hybrid power plant
  • fuel cells
  • Control systems
  • fault diagnosis and accommodation
Fuel cell power plant is a novel, clean and efficient energy source in distributed generation, and received extensive attentions from researchers, developers, and governments in recent decades. As one of the most advanced fuel cell technologies, hybrid fuel cell power plant has shown its potential for applications and is already under commercialization. A hybrid fuel cell with gas turbine power plant was envisioned as a base-load power source for distributed generation. As an emerging technique, the need of advanced control systems, which are essential components that guarantee reliable and efficient operations for the power plant, has motivated this investigation. This dissertation seeks to develop an intelligent control system to improve the energy conversion efficiency and the reliability of the hybrid fuel cell power plant. Toward this goal, an intelligent overall control system is established in the dissertation by developing and integrating a hybrid plant model, an optimal reference governor, and a fault diagnosis and accommodation system in the comprehensive control system. The hybrid plant model provides a novel modeling method that combines a mathematical model and a neural network model, which can identify plant parameters and uncertainties from operational data and can considerably improve the model accuracy for the following and future analysis and research work. The optimal reference governor is achieved by particle swarm optimization algorithms and a neural network state estimator to generate optimal setpoints and feedforward controls to improve plant efficiency. A nonlinear multi-objective optimization framework is developed by integrating heuristic optimization and artificial neural network technologies. Meanwhile, a fault diagnosis and accommodation system is implemented with fuzzy logic to detect and regulate system faults, preventing instabilities and damages to the power plant during system failures. The capability of the fuzzy theory in detecting and regulating system faults is demonstrated. The individual control systems are finally integrated into a comprehensive system that provides overall management for the hybrid power plant. With the integrated control system, the power plant can have high energy conversion efficiency in normal operations and can be well regulated during system failures. As a result, an intelligent autonomous control system is achieved to perform high quality plant-wide control, by which both efficiency and reliability can be guaranteed. Moreover, the presented intelligent control system and its design approach are not only valid for the hybrid fuel cell power plant, but also capable of other types of power plants, where efficiency and reliability need to be improved and guaranteed.