Failure rate estimation, a dangerous nonsense in a Bayesian view

✍ Scribed by C.A. Clarotti

• Book ID: 103970833
• Publisher: Elsevier Science
• Year: 1988
• Tongue: English
• Weight: 359 KB
• Volume: 20
• Category: Article
• ISSN: 0951-8320
• DOI: 10.1016/0951-8320(88)90092-0

✦ Synopsis

A BS TRA C T

Everything must be changed in order that everything may be left unchanged. This applies to the introduction of Bayesian techniques in reliability and PRA practice, and in particular to the estimation of probability-distribution parameters. Incoherence of estimating parameters in a Bayesian frame is enlightened in the paper by showing some incorrect results which estimation of parameters can lead to.

R(p, Oi, O j)

number of system components. reliability of rth system component. reliability of a k-out-of-n system under the hypothesis: pr=p, l <r<n. reliability of a k-out-of-n system under the hypothesis: pi = 1, pr=p, r¢i, i<r<n.

as above except that Pi = 0. reliability of a k-out-of-n system under the hypothesis: pi=l, pi=O, pr=-p,r¢i,j, l ~r<n.

as above but with i and j reversed. reliability of a k-out-of-n system under the hypothesis: p~=pj=l,p,=p,r¢i,j, l <_r<n. as above except that p~ = pj = O.