A multi-agent system-based intelligent control for large-scale power plants.

摘要

While the demand in power is increasing, power plants are getting larger and more complex to run. The major problems of a large-scale power plant are the coupling problems among the systems (e.g., control systems, plant subsystems) and optimization problems in power plant operation. In order to solve these problems, this dissertation contributes methodologies to design a control system framework for reducing the complexities and for intelligent control systems to provide an optimal operation in a large-scale power plant.; The object of this research is to develop an Intelligent Control (IC) for a large-scale industrial system, namely power plant, based on a new framework. In order to deal with the difficulty in handling a large-scale system with the added complexity of the network of sensors/actuators, the concept of the Multi-Agent System (MAS) is employed as a major solution technique.; To operate a large-scale power plant, monitoring and control systems are distributed and automated for each subsystem in the power plant. The interaction should be performed on the state-of-the-art architectures of single agent and multi-agent system to obtain cooperative and negotiable operation in an unpredictable and dynamic environment. Based on the architectures of single agent and multi-agent systems, the intelligent monitoring and control systems will be developed to provide not only stable and maneuverable operation, but also more autonomous, efficient, and optimal operation in the power plant.; The control systems will consist of an intelligent reference governor for generation of optimal set-points, intelligent identification system for preservation of stability, intelligent heuristic optimal control systems for optimal control actions and adaptive gain tuning. The monitoring systems will consist of intelligent fault-diagnosis and on-line performance monitoring. Finally, the configuration of the Multi-Agent System based Intelligent Control (MAS-IC) will be adapted to develop the dynamic organization system for better performance.