Title :
Knowledge-based supervision and diagnosis of complex process
Author :
Martin, Joseph Aguilar
Author_Institution :
Lab. d´´Autom. et d´´Anal. des Syst., CNRS, Toulouse, France
Abstract :
The knowledge about industrial processes concerns the structure and the behaviour as well as the relations between actions and effects. We describe here first the knowledge-based process description approach, leading to building a set of executable instructions. From that a causal graph can be implemented. Simultaneously, a knowledge-based variable analysis methodology is developed so that supervision and diagnosis actions can be taken from the performance characterisation values assigned to the nodes of the causal graph
Keywords :
diagnostic expert systems; graph theory; knowledge based systems; process control; causal graph; complex process; executable instructions; fault diagnosis; knowledge-based process description; knowledge-based systems; knowledge-based variable analysis; process control; supervisory control; Adaptive control; Buildings; Control theory; Design engineering; Difference equations; Industrial relations; Knowledge engineering; Ontologies; Open loop systems; State-space methods;
Conference_Titel :
Intelligent Control/Intelligent Systems and Semiotics, 1999. Proceedings of the 1999 IEEE International Symposium on
Conference_Location :
Cambridge, MA
Print_ISBN :
0-7803-5665-9
DOI :
10.1109/ISIC.1999.796659