Knowledge-based factory supervision — The CIM shell

Abstract

Abstract The paper presents a unified methodology for analysis, design and implementation of AI based software modules, like expert systems for production planning, quality control and preventive maintenance, for the factory of the future, both large and small batch-oriented. An extension of the hierarchical NBS factory reference model towards a distributed intelligent controller model is proposed. Planning horizons and planning update periods determine the hierarchical levels whereas decision frames given from upper to lower levels are regarded as constraints determining a certain degree of self responsibility for each separate controller. The decision activities inside the factory are analysed using the GRAI method. The functions of a CIM shell based on AI techniques are explained as a toolbox supporting design and implementation of the intelligent work-cell controller. Further details will be given in a subsequent paper.