Please use this identifier to cite or link to this item: http://repositorio.unicamp.br/jspui/handle/REPOSIP/101230
Type: Artigo de evento
Title: Revision About Accelerated Testing And Application In A Software For The Data Graphics Analysis
Author: Colmenero A.N.
Dedini F.G.
Abstract: This work presents a revision on different types of tests that are compatible with the nature of the data. They can be used to evaluate or to determine the reliability of equipment's and components in the different design stages of the design. The classification of the tests takes into account the type of censoring in the tests or the criteria to interrupt them. A classification of accelerated tests used in the industry is presented, along with the objectives and considerations of the most used models. Last of all, a software is presented for the processing of the accelerated tests data with constant stress and simple failure mode, for right censored data and with multiple censored data. The Weibull, Log-Normal and Normal distributions, combined with the most used models (Arrhenius and Inverse Power Law), are used. The software allows the processing of the accelerated tests' data by two methods: Minimum Square and Maximum Likelihood. The first offers the initial values to apply the method of Maximum Likelihood, and to assure the convergence of the Newton-Raphson method. The software offers graphs of residual, probabilities vs. time, stress vs. percentiles where it is possible to make an evaluation of the model of the distribution and of the variation or not of the failure mechanism. Statistical tests are used for the validation of the tests, allowing the possibility of using different confidence levels to accomplish the calculations.
Editor: ASME, United States
Rights: fechado
Identifier DOI: 
Address: http://www.scopus.com/inward/record.url?eid=2-s2.0-0343602885&partnerID=40&md5=fc0cd919eef98ac269d39bc676a3a2d7
Date Issue: 1999
Appears in Collections:Unicamp - Artigos e Outros Documentos

Files in This Item:
There are no files associated with this item.


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.