Fault diagnosis with dynamic fuzzy discrete event system approach

Kiliç E., Karasu Ç., LEBLEBİCİOĞLU M. K.

14th Turkish Symposium on Artificial Intelligence and Neural Networks, TAINN 2005, İzmir, Turkey, 16 - 17 June 2005, vol.3949 LNAI, pp.117-124, (Full Text) identifier

  • Publication Type: Conference Paper / Full Text
  • Volume: 3949 LNAI
  • Doi Number: 10.1007/11803089_14
  • City: İzmir
  • Country: Turkey
  • Page Numbers: pp.117-124
  • Middle East Technical University Affiliated: Yes

Abstract

Determining faults is a challenging task in complex systems. A discrete event system (DES) or a fuzzy discrete event system (FDES) approach with a fuzzy rule-base may resolve the ambiguity in a fault diagnosis problem especially in the case of multiple faults. In this study, an FDES approach with a fuzzy rule-base is used as a means of indicating the degree and priority of faults, especially in the case of multiple faults. The fuzzy rule-base is constructed using event-fault relations. Fuzzy events occurring any time with different membership degrees are obtained using k-means clustering algorithm. The fuzzy sub-event sequences are used to construct super events. The study is concluded by giving some examples about the distinguishability of fault types (parameter, actuator) in an unmanned small helicopter. © Springer-Verlag Berlin Heidelberg 2006.