Display of diagnostic information from multiple viewpoints in an anomalous situation of complex plants

The paper describes the objectives, approaches, and current status of an intelligent information display agent called semantic representation interface (SI). The SI displays diagnostic information from a functional viewpoint as well as from behavioral, operational, and structural ones to plant operators in an anomalous situation. The functional information of the plant will help operators to understand the overall plant situation and the necessity of a recommended counter action written in operation manuals or suggested by the SI. Diagnostic information of anomaly identification and finding possible counter actions are derived by qualitative reasoning based on a functional model. The selection of suitable counter actions which will be developed in the near future is made using a fast future prediction code generated by an intelligent modular simulation system. The derived diagnostic information is displayed in a graphical way so that an operator can understand the information as a pattern. In addition to the graphical display, necessary graphs to understand the plant situation and the derived diagnostic information are automatically shown. The applicability of the diagnostic techniques applied the qualitative reasoning is discussed through applications to an oil refinery plant.