The Community for Technology Leaders
Green Image
<p>Conventional heuristic-based expert systems for diagnosis and control cannot adequately handle unforeseen abnormal events. Our system combines heuristics with model-based reasoning to control a thermal power plant, and to diagnose and resolve abnormal events at the plant.</p> <p>In the field of diagnosis and control of thermal power plants, the more intelligent and flexible systems become, the more knowledge they need. Conventional diagnostic and control expert systems are based on heuristics stored a priori in knowledge bases, so they cannot deal with unforeseen abnormal situations in the plant. Such situations could occur if knowledge engineers do not implement the necessary knowledge.</p> <p>Skilled human operators can operate a plant and deal with unforeseen abnormal situations because they have fundamental knowledge about the structure and functions of the plant's component devices, about the principles of plant operations, and about the laws of physics. An operator's basic thought process comprises diagnosis of the situation, generation of plant control knowledge, and verification of generated knowledge. Operators can deal with such situations by repeatedly executing these steps, using the fundamental knowledge.</p> <p>We have developed a diagnostic and control expert system that simulates these thought processes. Our system combines model-based diagnosis for unforeseen abnormal situations, model-based knowledge generation for plant control, and knowledge-based plant control both with generated and a priori stored knowledge. Experiments on a simulated thermal power plant indicate that our system can effectively control the plant and deal with abnormal situations.</p>

M. Iwamasa and N. Sueda, "A Pilot System for Plant Control Using Model-Based Reasoning," in IEEE Intelligent Systems, vol. 10, no. , pp. 24-31, 1995.
91 ms
(Ver 3.3 (11022016))