Discovering knowledge from automatically gathered production machine data

A system designed for the automated acquisition of knowledge for an expert system for process control has been applied to discovery of knowledge of automatically gathered production machine data. Data relating to the number of parts produced by a number of production machines has been provided. Meaningful information which is useful in terms of understanding the production process is derived from the data. This paper presents details of the format of the automatically gathered production data and identifies variables which are derived from the production data in order to extract knowledge in the form of production rules. The presented analysis concentrates on the number of parts produced in the production process and the relationship of the number of parts produced to a number of other factors derived from the automatically gathered production machine data. This relationship is represented by the production rules. The nature of knowledge extraction from the automatically gathered production machine data is illustrated by providing examples showing rules derived from the production machine data. Confirmation of a correlation of a priori expert knowledge of machine production and knowledge represented in the production rules has been provided. In conclusion, the usefulness of the extracted knowledge is identified to be that it allows the production operation to be understood with no a priori expert knowledge of machine production.