Author(s)

S. Cauvin & B. Braunschweig

Abstract

Graphical knowledge representation in the Alexip system for petrochemical process supervision S. Cauvin & B. Braunschweig Institut Frangais du Petrole, 1-4 Avenue de Bois Preau BP 311, 92506 RUEIL MALMAISON Cedex, France ABSTRACT : Monitoring a petrochemical process run with complex instrumentation and control systems entails using sophisticated! reasoning methods to solve problems in which the dynamic process, behaviour has to be taken into account and to exploit available! knowledge as well as possible. The Alexip knowledge-based systemt contains several reasoning methods. In this paper, we will deal withi the implementation of one of them. The aim is to precisely! determine which events are governing the process changes at ai given time in order to be able to suggest the proper corrective actions to be undertaken to maintain or bring back the unit into a state ofi optimal operation. The method we describe involves a graphical! representation of the knowledge used by a

Keywords