As power generating companies seek to improve plant reliability, maintain efficiency, and increase outage intervals, many are turning to on-line performance and condition monitoring to augment traditional high-low control system alarms.
Such monitoring systems employ advanced modeling techniques for automated early detection of incipient problems and are an essential element in an effective asset management program. Read More
Early warning provides a reduction in repair costs,an increase in equipment run times, reduction in fuel costs, reduction in replacement power cost and economical use of planned or unplanned outages.
Traditional first principles models provide the ability to detect of abnormal behavior based on engineering relationships related to heat transfer, conservation of mass and energy, and fluid dynamics. Empirical models use historical data to accomplish the same goal on equipment where a first principles model is unavailable or overly complex, (turbine shaft, fan bearings, damper settings, etc.). This paper describes a new method which uses hybrid equipment models based on both empirical and first principles techniques. Such hybrid models incorporate elements unique to each method to provide comprehensive monitoring of all plant equipment.
The on-line monitoring system to be discussed uses hybrid physical-empirical models to detect abnormalities and alert plant personnel. Asset managers use the refined detection information from models to oversee the six major plant concerns of reliability, efficiency, environmental, chemistry, fouling, and cycle leakage. Numerous case studies demonstrating the results of the hybrid models at a coal-fired power plant are included.
GP Strategies Corporation
Article written by:
Dr. Elmer Hansen, Principal Engineer
GENERAL PHYSICS CORPORATION
Sam Rohmer III, Control Room Operator
CPS Energy