• DocumentCode
    799956
  • Title

    Maximum likelihood parameter estimation in the extended Weibull distribution and its applications to breakdown voltage estimation

  • Author

    Hirose, Hideo

  • Author_Institution
    Dept. of Comput. Sci. & Electron., Kyushu Inst. of Technol., Fukuoka, Japan
  • Volume
    9
  • Issue
    4
  • fYear
    2002
  • fDate
    8/1/2002 12:00:00 AM
  • Firstpage
    524
  • Lastpage
    536
  • Abstract
    Although the Weibull distribution is widely used in a variety of reliability applications, difficulties in its treatment, particularly in three parameter cases in the maximum likelihood estimation, hinder us from using the distribution. The extended Weibull distribution proposed by Marshall and Olkin (1997) can avoid the difficulties which appear in the conventional Weibull distribution models. This paper shows the maximum likelihood estimation method in the extended Weibull distribution model. The paper also illustrates some typical applications for breakdown voltage estimation in which the extended models are superior to the conventional Weibull models. The central discussion is whether the shape parameters in the extended model accomplish the mass shifting effect of the distribution.
  • Keywords
    Weibull distribution; electric breakdown; maximum likelihood estimation; reliability theory; Kolmogorov statistic; Newton-Raphson; breakdown voltage estimation; continuation method; cumulative distribution; density functions; epoxy resin; extended Weibull distribution; extended two-parameter exponential; iterative methods; mass shifting effect; maximum likelihood parameter estimation; reliability; shape parameters; three parameter cases; transformer oil; Application software; Computer science; Data engineering; Maximum likelihood estimation; Parameter estimation; Reliability engineering; Shape; Systems engineering and theory; Voltage control; Weibull distribution;
  • fLanguage
    English
  • Journal_Title
    Dielectrics and Electrical Insulation, IEEE Transactions on
  • Publisher
    ieee
  • ISSN
    1070-9878
  • Type

    jour

  • DOI
    10.1109/TDEI.2002.1024429
  • Filename
    1024429