New monitoring system for thermal power plants using digital image processing and sound analysis

Recently in thermal power plants, many components are operated on severe condition because of various demands for enormous electric power and frequent load regulation. In a control room, it is difficult to know accidents such as fire, fuel oil leak, and unusual sound, at a moment's notice. Maintenance operators, therefore, have to patrol the plant yard to find out them. To release from the patrol, the needs for automatic monitoring system to detect those accidents are increasing. We have developed a new intelligent monitoring system using advanced ima,ae and sound analysis, which is reference to human senses, to realize the request. This system can detect oil and steam leak from burners, unusual sound from machines, and, fire and smoke in the circumference. This paper describes the system outline, features, and methods of image processing and sound analysis.