Fault detection and diagnosis for nonlinear systems: A support vector machine approach

Rana Ortaç-Kabaoglu*, Ibrahim Eksin, Engin Yesil, Müjde Güzelkaya

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

3 Citations (Scopus)

Abstract

In this paper, a fault detection and diagnosis (FDD) technique for nonlinear systems based on support vector machines (SVM) is presented. Support vector regression (SVR) has been used in fault detection process and support vector classification (SVC) has been used in diagnosis process. In fault detection process, the confidence band idea represents the normal operating conditions of the system. The upper and the lower boundaries of the confidence band are modelled by two different SVR machines. A fault is detected when an output signal exceeds the upper or lower bounds of the generated confidence band. A support vector multi-classification method, one-against-all, has been used to classify the occurring fault within the group of expected and predefined faults in technical system. The performance of the proposed FDD method is illustrated on simulation example involving a two-tank water level control system under faulty conditions.

Original languageEnglish
Title of host publication2nd IFAC International Conference on Intelligent Control Systems and Signal Processing
PublisherIFAC Secretariat
EditionPART 1
ISBN (Print)9783902661661
DOIs
Publication statusPublished - 2009

Publication series

NameIFAC Proceedings Volumes (IFAC-PapersOnline)
NumberPART 1
Volume2
ISSN (Print)1474-6670

Keywords

  • Fault detection and diagnosis
  • Support vector classification and regression
  • Two tank water level system

Fingerprint

Dive into the research topics of 'Fault detection and diagnosis for nonlinear systems: A support vector machine approach'. Together they form a unique fingerprint.

Cite this