Knowledge acquisition and modification methods are described for building a quality diagnosis expert system. Knowledgeof a quality diagnosis expert is represented in a tabular form. The final quality properties are predicted from the operating conditions of the intermediate process and are compared with the target values. A pattern of deviations of thefinal quality properties from the target values is summarized in a table. A data file in the form of knowledge table is created or revised by an editor in an actual quality diagnosis system, and a knowledge base in an if-then format is created by a knowledge generation program. The practical application of this procedure and the architecture and human interface of a silicon steel glass film quality diagnosis expert system are described.
ASJC Scopus subject areas