Application of the adaptive-predictive controllers to plant safety surveillance utilizing on-line Plant Analyzer

The plant safety surveillance concept described in this report continuously provides information that is not accessible to measurements and which, in conjunction with the existing plant data, offers promise for a thorough understanding of the plant status and of the causes of upsets when they occur. The concept is based on using the Plant Analyzer Controller, which works in tandem with the Plant Analyzer, to control inputs to the Plant Analyzer so that its outputs continuously track the selected plant measurements. The basic premise is that, if the key plant measurements are being well reproduced (tracked) by the Plant Analyzer, the wealth of information which is being continuously computed (most of which cannot be measured), can be trusted and used to ascertain the plant status. Research results presented in this report show that the act of tracking immediately uncovers upset conditions which are, otherwise, either not amenable to measurements or take a long time to diagnose using conventional methods. The Plant Analyzer (a systems safety code capable of running in real time on a plant computer), and the Plant Analyzer Controller, would be used ''on-line'', i.e., connected to the plant computer which provides information on plant measurements. Their outputs would be graphically displayed in a comprehensive and easily understood manner. The version of the Plant Analyzer Controller described in this report is based on the Adaptive-Predictive, digital control theory. 154 figs.