Diagnosis via qualitative parameter estimation

George MacLeod Coghill, Hind Al-Ballaa

Research output: Chapter in Book/Report/Conference proceedingPublished conference contribution

Abstract

Parameter estimation is a well established numeric
fault detection method of the process supervision task inside the
control engineering community. This method requires accurate
numeric knowledge. Such information is not available or expensive to obtain for ill-defined systems. This paper presents, under
a pure qualitative context, an approach for model based fault
diagnosis via parameter estimation. Estimating parameters
qualitatively in this work is based on finding out whether the
malfunction is caused by an increase/decrease of the parameter
included in the model of a faulty component. In fact achieving
that is a challenge because the lack of discriminating power
traditionally associated with qualitative reasoning techniques
makes diagnostic engines compete for producing high resolution
diagnoses. No attempts to identify in pure qualitative context
the way those candidates failed.
Original languageEnglish
Title of host publicationProceedings of the 11th UK Workshop on Computational Intelligence (UKCI)
Place of PublicationManchester
PublisherUniversity of Manchester
Pages80-86
Number of pages7
Publication statusPublished - Sept 2011
Event11th UK Workshop on Computational Intelligence - Manchester, United Kingdom
Duration: 7 Sept 20119 Sept 2011

Conference

Conference11th UK Workshop on Computational Intelligence
Country/TerritoryUnited Kingdom
CityManchester
Period7/09/119/09/11

Keywords

  • artificial intelligence
  • parameter estimation
  • qualitative model based diagnosis

Fingerprint

Dive into the research topics of 'Diagnosis via qualitative parameter estimation'. Together they form a unique fingerprint.

Cite this