Automatic Synthesizer of Operating Procedures for Chemical Plant by Use of Fragmentary Knowledge

Abstract

A methodology for developing a computer-based system that can generate a sequence of operations for starting up a chemical plant is given by making use of AI technology.
A computer-based system has been developed on the basis of such methodology and applied to the start-up of a practical chemical plant: a subprocess of an existing ethylene plant.
To facilitate the use of heuristic knowledge at each stage of decision-making during the generation of operating procedures, where relatively higher-level knowledge is needed, another inference mechanism, called HIGH-LEVEL INFERENCE ENGINE, is introduced by using a PROLOG-like logical programming paradigm.
A knowledge base is also organized in a generic fashion so that the system can be applied to chemical plants in general.