Modelagem de dados de sistemas reparáveis com fragilidade

• Autor(a) principal: Feitosa, Cirdêmia Costa
• Data de Publicação: 2015
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFSCAR
• Texto Completo: https://repositorio.ufscar.br/handle/ufscar/7233
• topic: Estatística; Sistemas reparáveis; Fragilidade; Abordagem bayesiana; Minimal repair; Perfect repair; Imperfect repair; Frailty; Bayesian; Repairable systems; CIENCIAS EXATAS E DA TERRA::PROBABILIDADE E ESTATISTICA

The usual models in repairable systems are minimal, perfect and imperfect repair, and, in the literature, the minimum repair model is the most explored. In repairable systems it is common that the same type of components are studied and in these cases is relevant to verify the heterogeneity between them. According to Vaupel et al. (1979), the standard methods for analysis of repairable systems data ignore the heterogeneity not observed and in some cases this should be considered. Such variability can be estimated from frailty models, characterized by using a random e ect. It is proposed that the minimum repair model with frailty in order to estimate the heterogeneity not observed between systems. For this model it was conducted a simulation study in order to analyze the frequentist properties of the estimation process. The application of a real data set showed the applicability of the proposed model, in which the estimation of the parameters were determined from maximum likelihood and Bayesian approaches.