Journal of South China University of Technology (Natural Science Edition) ›› 2011, Vol. 39 ›› Issue (6): 47-52.doi: 10.3969/j.issn.1000-565X.2011.06.009

• Mechanical Engineering • Previous Articles     Next Articles

Comparison of Parameter Estimation Methods for Reliability Model

Wang Xiao-fengShen Gui-xiangZhang Ying-zhiChen Bing-kunZheng ShanLiu Wei2   

  1. 1. College of Mechanical Science and Engineering,Jilin University,Changchun 130022,Jilin,China; 2. School of Life Sciences,Changchun University of Science and Technology,Changchun 130022,Jilin,China
  • Received:2010-12-29 Revised:2011-02-13 Online:2011-06-25 Published:2011-05-06
  • Contact: 王晓峰(1984-) ,男,博士生,主要从事数控机床可信性技术研究. E-mail:289799164@163.com
  • About author:王晓峰(1984-) ,男,博士生,主要从事数控机床可信性技术研究.
  • Supported by:

    国家自然科学基金资助项目( 50875110) ; 国家科技重大专项项目( 2010ZX04014-011)

Abstract:

In order to solve the problems existing in the parameter estimation of the Weibull distribution for the reliability determination of complex systems,both the x-direction and y-direction parameter estimation methods are introduced,and the statistical characteristics of the two methods are analyzed. Then,a comparison between the two methods is carried out via the Monte Carlo simulation. Finally,by reasonably selecting a parameter estimation
method,a Weibull distribution model of complex products is established. The results show that,for the reliability model of complex systems,when the sample size n is more than 14,the x-direction parameter estimation method is applicable. However,when n is not more than 14,a suitable parameter estimation method should be selected according to the mean square deviation of the sample.

Key words: reliability, parameter estimation, small sample, Weibull distribution, Monte Carlo simulation