Resumo

PINA-MONARREZ, Manuel R.; RAMOS-LOPEZ, Miriam L.; ALVARADO-INIESTA, Alejandro  e  MOLINA-ARREDONDO, Rey D.. Robust sample size for weibull demonstration test plan. Dyna rev.fac.nac.minas [online]. 2016, vol.83, n.197, pp.52-57. ISSN 0012-7353.  https://doi.org/10.15446/dyna.v83n197.44917.

The efficiency of a Weibull demonstration test plan is completely determined by the total experimental time (T_a), which depends on the unknown sample size (n) and on the Weibull shape parameter (b). Thus, once b was selected, T_a depends only on n. Unfortunately, because n was estimated by the parametrical binomial approach, then if the confidence level C was higher than 0.63, n, and as consequence T_a, was overestimated, (For C<0.63, they were underestimated). On the other hand, in this paper, because the intersection between n and b, for which T_a was unique, was found with n depending only on R(t), then the estimation of T_a was optimal. On the other hand, since once b was selected, h was completely determined, then b and h were used to incorporate the expected failure times of the operational level in an accelerated life test analysis (ALT). Numerical applications are given.

Palavras-chave : Weibull demonstration test plan; Success run testing; Lipson equality; Accelerated life testing.

        · resumo em Espanhol     · texto em Inglês     · Inglês ( pdf )