Geir Stian Landsverk (geir.stian.landsverk@itk.ntnu.no)
Wed, 05 Apr 2000 14:58:19 +0100
Heisann,
Alle ønskes velkommen til PROST seminar
mandag 10 april kl. 14.15 - 15.00
i rom B-337, Elektrobygning B.
Foredraget er ved Joe Qin fra Department of Chemical Engineering,
University of Texas at Austin (http://control.che.utexas.edu/qinlab) og
omhandler "Dynamic Fault Detection and Identification with Maximized
Sensitivity" - se sammendrag under.
Vårens program finner du på http://www.itk.ntnu.no/groups/PROST/seminars.html.
Det vil ikke bli noen matservering, og foredraget varer fra 14.15-15.00.
-Geir Stian
>Title: Dynamic Fault Detection and Identification with Maximized
>Sensitivity
>
>Abstract:
>
>In this talk we present a new method for the
>detection, reconstruction, and identification of faulty sensors using a
>normal process
>model. The normal process model can be built from first principles or
>statistical methods such as partial least squares (PLS) or
>principal component analysis (PCA). The model {\em residual} is
>used to detect sensor faults that demonstrate a deviation from the
>normal process model. To identify
>which sensor is faulty, a {\em structured residual approach with maximized
>sensitivity} (SRAMS) is proposed to make one residual insensitive to one
>subset of faults but most sensitive to others. The confidence
>limits for the EWMA filtered structured residuals (FSR) are determined
>using statistical inferential techniques.
>In addition, other indices
>including generalized likelihood ratio
>(GLR) test, cumulative sum (Qsum) and cumulative variance (Vsum) of the
>structured residuals
>are proposed and compared for the identification of faulty sensors. The
>fault magnitude is then estimated based on the model and faulty data. Four
>types of sensor faults, including bias, precision degradation, drifting,
>and complete failure are simulated to test the proposed method.
>Data from an industrial boiler process are used to test the effectiveness
>of
>the proposed methods. Both single faults and simultaneous double faults
>are detected and uniquely identified with the proposed method.
-------------------------------------------------------------
Geir Stian Landsverk
Dr.ing. student - Department of Engineering Cybernetics
Norvegian University of Science and Technology
7491 TRONDHEIM, NORWAY
Phone:++(47) 7359 4386 (office), ++(47) 7388 9310 (private)
E-mail: Geir.S.Landsverk@itk.ntnu.no
http://www.itk.ntnu.no/ansatte/Landsverk_Geir/
--------------------------------------------------------------
This archive was generated by hypermail 2.0b3 on Wed Apr 05 2000 - 19:42:35 MET DST