Determination of a Function for a Degradation Process by Means of Two Diagnostic Bond Graphs

  • This paper proposes a novel approach to a bond graph model-based failure prognosis for systems represented by a mode switching linear time-invariant model. A function for a degradation process is numerically determined by using a first stage and a second stage diagnostic bond graph model (DBG). Evaluation of the Analytical Redundancy Relations (ARRs) from the first stage DBG provides residuals that enable to detect the onset of incipient faults. The second stage DBG model accounts for parametric degradation by means of an unknown function. ARRs derived from the second stage DBG make use of the residuals of the first stage ARRs and constitute an implicit relation for the unknown degradation function. The computation takes place online concurrently to the monitoring of a real system and the measurement of its output signals. Once the time evolution of a degradation process has been computed up to some time instant, the time evolution of an ARR residual using it can be projected into the future in order to estimate the Remaining Useful Life (RUL) of a system. With progress of time the computation of a degradation process can be repeated and estimations of a RUL can be improved.

Export metadata

Additional Services

Share in Twitter Search Google Scholar Availability
Document Type:Conference Object
Parent Title (English):IFAC-PapersOnLine
First Page:636
Last Page:642
Date of first publication:2018/10/11
10th IFAC Symposium on Fault Detection, Supervision and Safety for Technical Processes, August 29-31, 2018, Warsaw, Poland
Tag:Diagnostic bond graphs; Mode switching LTI model; Mode-dependent ARRs; Model-based failure prognosis; Parameter degradation model; Remaining Useful Life; System health monitoring
Departments, institutes and facilities:Fachbereich Informatik
Dewey Decimal Classification (DDC):000 Informatik, Informationswissenschaft, allgemeine Werke / 000 Informatik, Wissen, Systeme / 004 Datenverarbeitung; Informatik
Entry in this database:2018/08/10