• DocumentCode
    3506973
  • Title

    Error analysis and enhancing of the ellipticity estimation of steel coils

  • Author

    Schleicher, Daniel C H ; Zagar, Bernhard G.

  • Author_Institution
    Inst. for Meas. Technol., Johannes Kepler Univ., Linz, Austria
  • fYear
    2009
  • fDate
    3-5 Nov. 2009
  • Firstpage
    2056
  • Lastpage
    2062
  • Abstract
    Measuring the ellipticity of coils is an important topic for the quality assurance in the steel industry. The concentric ellipse fitting algorithm discussed in a previous paper, is an important step to estimate the ellipticity of steel coils, when only a few elliptic curve segments can be used as input data to be processed. In this paper the concentric ellipse fitting algorithm as given by Schleicher et al. is enhanced and adapted for various numerical calculation methods. The variance of the ellipticity estimation is analyzed for different estimators. At first for the simple ellipse fitting problem with only three parameters to be estimated and further for the concentric ellipse fitting problem. Based on these analyses a weighted concentric ellipse fitting algorithm is presented, which is shown to improve the estimation results especially for short data record lengths. Finally the improved algorithm is applied to an industrial application, where the ellipticity of steel coils has to be estimated.
  • Keywords
    coils; curve fitting; error analysis; quality assurance; steel industry; ellipticity estimation; error analysis; numerical calculation method; quality assurance; steel coils; steel industry; weighted concentric ellipse fitting algorithm; Algorithm design and analysis; Analysis of variance; Coils; Curve fitting; Elliptic curves; Error analysis; Metals industry; Parameter estimation; Quality assurance; Steel;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Industrial Electronics, 2009. IECON '09. 35th Annual Conference of IEEE
  • Conference_Location
    Porto
  • ISSN
    1553-572X
  • Print_ISBN
    978-1-4244-4648-3
  • Electronic_ISBN
    1553-572X
  • Type

    conf

  • DOI
    10.1109/IECON.2009.5415083
  • Filename
    5415083