A residual inspired approach for fault localization in DES

Abstract

In this paper an approach for fault localization in Discrete Event Systems (DES) is proposed. The presented diagnosis method allows fault localization using a fault-free nominal system model. Via a systematic comparison of the observed and the expected system behavior, it is possible to determine a set of fault candidates. Inspired by residuals known from diagnosis in continuous systems, different set operations are presented that carry out this comparison. After a fault has been detected and a first estimate concerning its localization has been performed, a special algorithm analyzes the further system behavior in order to determine a more precise fault localization. The algorithm also works on the basis of the nominal system model. The method is explained using a manufacturing system example.