• DocumentCode
    2682281
  • Title

    Reliability analysis method of mechanical system with correlated failures

  • Author

    Tu, Hongmao ; Qian, Yunpeng ; Li, Juan

  • Author_Institution
    North China Syst. Eng. Inst., CNGC, Beijing, China
  • fYear
    2011
  • fDate
    12-15 June 2011
  • Firstpage
    137
  • Lastpage
    143
  • Abstract
    This paper presents a general mechanical system reliability modeling and analysis method, which consider the correlations of failure modes. Firstly, the performance functions which mathematically represent the failure modes are linearized based on FORM (First Order Reliability Methods), and the correlation coefficient matrix of those functions is calculated. Then, the PFTA (Probabilistic Fault Tree Analysis) method is adopted to establish the system reliability model which can take the correlations into account. Finally, effective algorithms such as G-FOMN (General First Order Multinormal Method) are used to quantify the system reliability, and the sensitivities of the system reliability index with respect to each failure mode, as well as the mean value and standard deviation of each random variable are also calculated. An engineering example demonstrates the applicability of the proposed method.
  • Keywords
    failure (mechanical); fault trees; matrix algebra; probability; reliability theory; FORM; G-FOMN; PFTA; correlated failures; correlation coefficient matrix; first order reliability methods; general first order multinormal method; mechanical system reliability; probabilistic fault tree analysis; reliability analysis method; standard deviation; Correlation; Indexes; Linear approximation; Mathematical model; Reliability; Sensitivity; Sensitivity; correlative coefficient; mechanical system reliability; performance function; system reliability model;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Reliability, Maintainability and Safety (ICRMS), 2011 9th International Conference on
  • Conference_Location
    Guiyang
  • Print_ISBN
    978-1-61284-667-5
  • Type

    conf

  • DOI
    10.1109/ICRMS.2011.5979249
  • Filename
    5979249