Maintenance considerations in CIM design: An expert system perspective

Abstract

An unmanned manufacturing system (UMS) will typically be characterized by a large number of highly reliable components which are interconnected in a highly complex fashion. These characteristics suggest: (1) there will be a large number of ways in which a UMS can fail; (2) the mean time between repeat occurrences of a given type of problem will be quite large; and (3) the time and place where a failure creates an observable symptom may be quite different from the time and place where the failure occurred. This suggests that an UMS could be difficult to maintain. Maintenance personnel obviously need assistance for dealing with the complex cognitive problems which can arise in troubleshooting the UMS. This paper describes a procedure for developing an expert system to provide this assistance. Simulation is utilized to generate proxy data on how the system could be expected to behave when a given type of failure occurs. Transfer-function/intervention analysis is utilized to develop models which describes the system transient response to the various types of failures. Failure problems are diagnosed by comparing the shape of the actual transient response with the shapes of transient responses which correspond to the various types of failures.