Institutional Repository of Vilnius University
Pagreitintų eksperimentų duomenų modeliavimas, kai netinka AFT modelis /
Abstract
dc:descriptionInvestigation of the reliability of highly reliable components, and systems is an important industrial problem. Relatively short time of reliability experiments usually does not allow toobserve failures of units. One of possible ways out is to do accelerated life testing (ALT), i.e. to test units under higher than usual stresses which shorten the life of them. In such a case more failures may be observed during the experiment. The most popular accelerated failure time (AFT) model is rather restrictive because under this model the failure time distributions under different constant stresses differ only in scale. Some authors use a generalization of the AFT model for constant stresses supposing that under different stresses not only scale but also shape parameters are different. Unfortunately, this changing shape and scale (CHSS) model has an unnatural property for accelerated experiments: the survival functions under different stresses intersect in some point. That is why we are trying a new model, which lets us avoid survival functions intersections. We used this new model for estimation of reliability characteristics under normal stress using accelerated experiments. Real data analysis was done. We showed that the AFT model is not adequate for these data. The likelihood ratio test rejected strongly the AFT model. We obtained much more adequate results with the new model.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2017
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Brazdeikaitė, Greta,
- Contributors dc:contributor
-
- Bagdonavičius, Vilijandas
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD35408096&prefLang=en_US