Object-oriented model-based reasoning for diagnosing automated manufacturing systems

Abstract

In automated manufacturing, availability is the percentage of time that a production system is functioning properly. The author addresses the problems of availability by exploring the feasibility of a class of diagnosis algorithm. To develop this approach, a range of theory used in other areas was surveyed and synthesized. Petri net analysis was applied to fault detection. Rule-based inference and model-based reasoning were applied to failure isolation and hypothesis validation. A large part of the effort was the use of object-oriented software engineering to integrate the pieces into a unified algorithm. A software system was developed to realize the advanced algorithm and applied to a case study manufacturing subsystem for evaluation and analysis.<>