Advanced integrated condition monitoring

A key aspect of effective aging management for plant components is an accurate knowledge of the current condition of the components. Condition monitoring is often recommended from a Reliability Centered Maintenance analysis as offering a practical means for routinely determining changes in the condition of a component which can be used to facilitate Predictive or On-condition maintenance regimes. Traditional approaches to condition monitoring such as vibration, oil analysis etc. are based upon detecting changes in a single measured parameter. The interpretation of changes in these parameters requires an understanding of the current operating environment, for example plant output power or machine speed. This paper presents a new condition monitoring concept which uses changes occurring in more than one measurement to provide a more accurate and sensitive detection of component degradation. This is achieved by integrating conventional condition monitoring methods such as vibration and trend analysis with advanced state-of-the-art robust model based diagnosis and knowledge based diagnosis. A key feature of the advanced diagnostic algorithms developed as part of this approach is their ability to automatically assimilate data from-n different sources. This integration automatically takes into account normal plant response to operating changes thereby providing adaptive thresholding of alarm limits. Consequently maximum sensitivity to component degradation is ensured whilst minimizing false alarm rate. This paper discusses the application of this integration concept to pumping systems and steam generators