Systematic construction of qualitative physics-based rules for process diagnostics

Abstract

A novel first-principles-based expert system is proposed for on-line detection and identification of faulty component candidates during incipient off-normal process operations. The system performs function-oriented diagnostics and can be reused for diagnosing single-component failures in different processes and different plants through the provision of the appropriate process schematics information. The function-oriented and process-independent diagnostic features of the proposed expert system are achieved by constructing a knowledge base containing three distinct types of information, qualitative balance equation rules, functional classification of process components, and the process piping and instrumentation diagram. The various types of qualitative balance equation rules for processes utilizing single-phase liquids are derived and their usage is illustrated through simulation results of a realistic process in a nuclear power plant.KeywordsNuclear Power PlantControl VolumeProcess ComponentLeak LocationFaulty ComponentThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.