An expert system for startup scheduling and operation support in fossil power plants

An expert system is described which can support operations of fossil power plants in creating the optimum startup schedule and executing it accurately. The optimum speed-up and load-up pattern is obtained through an iterative modification of schedule parameters. The modification is based on fuzzy inference using quantitative calculations and qualitative knowledge. The quantitative calculations are based on plant dynamics models. The qualitative knowledge consists of schedule modification rules with fuzziness, which represent the relationships between stress margins and modification rates of the schedule parameters. Simulation analysis demonstrates that the system provides quick and accurate plant startups.<>