Filter for detecting and isolating faults for a nonlinear system

L. Giovanini, A. Dutka

Research output: Contribution to conferencePaper

37 Downloads (Pure)

Abstract

In the paper the problem of detecting and isolating multiple faults for nonlinear systems is considered. A strategy of state filtering is derived in order to detect and isolate multiple faults which appear simultaneously or sequentially in a discrete time nonlinear systems with unknown inputs. For the considered system for which a fault isolation condition is fulfilled the proposed method can isolate p simultaneous faults with at least p+q output measurements, where q is the number of unknown inputs or disturbances. A reduced output residual vector of dimension p+q is generated and the elements of this vector are decoupled in a way that each element of the vector is associated with only one fault or unmeasured input.
Original languageEnglish
Number of pages6
Publication statusPublished - Jul 2003
EventSafe Processes 2003 - Washington, United States
Duration: 1 Jul 2003 → …

Conference

ConferenceSafe Processes 2003
CountryUnited States
CityWashington
Period1/07/03 → …

Keywords

  • filter
  • detecting
  • isolating faults
  • nonlinear system
  • analytical redundancy
  • unbiased
  • directional residuals
  • dynamic observers

Fingerprint Dive into the research topics of 'Filter for detecting and isolating faults for a nonlinear system'. Together they form a unique fingerprint.

  • Cite this

    Giovanini, L., & Dutka, A. (2003). Filter for detecting and isolating faults for a nonlinear system. Paper presented at Safe Processes 2003, Washington, United States.