基于马尔可夫链蒙特卡罗的实时可靠性预测方法研究_徐正国.pdfVIP

Jour al of Mecha ical Stre gth 2007, 29( 5) : 765~ 768 X REAL- TIME RELIABILITY PREDICTION METHOD BASED ON MARKOV CHAIN MONTE CARLO XX XXX ( 清华大学 自动化系, 北京100084) XU ZhengGuo ZHOU DongHua ( Dep artment of Automation, Tsinghua University , Bejiing 100084, China) , , , , , , , ( MCMC) O213. 2 T B114. 3 Abstract In order to predict the product. s rea-l time reliabilit when the variance of error terms in its nonlinear degradation path model was unknown, a Ba esian anal sis method was presented. The noninformative prior of the standard deviation of error terms was in- troduced for Ba esian anal sis. Then, given the prior densit of the random coefficients in the degradation path model, the Markov Chain Monte Carlo method was adopted to generate samples of joint posterior distribution of random coefficients and unknown standard devia- tion. B these posterior samples, the product. s reliabilit during a certain period in the future was predicted. The numerical example b use of fatigue crack growth data shows the effectiveness of the proposed method. Key words Performa ce degradatio ; Reliability; Predictive mai te a ce; Bayesia a alysis; Markov Chai Mo te Carlo( MCMC) Corresponding author : X U ZhengGuo, E-mail: xzg02 @mails . tsinghua. edu. n, Tel: + 86- 10 The project supported b the National Natural Science Foundation of China ( No. . Manuscript received in revised form