A0407
Title: Statistical inference for generalized power-law process in repairable systems
Authors: Pedro Ramos - Pontificia Universidad Catolica de Chile (Brazil) [presenting]
Abstract: Repairable systems are often used to model the reliability of restored components after a failure is observed. Among various reliability growth models, the power law process (PLP) or Weibull process has been widely used in industrial problems and applications. We propose a new class of model called generalized PLP (GPLP), based on change points, which can be treated as known or unknown parameters or interpreted as failure times, in which we consider the impact of all or some fixes about fault intensity function. In this context, GPLP, unlike the usual PLP, is not restricted to the assumption of minimal repair (RM), it is possible to consider other situations, such as perfect, efficient, and harmful repair. Some special cases of GPLP are presented, such as the main models for the analysis of repairable systems under the assumption of imperfect repair. The estimators of the proposed model were obtained using the maximum likelihood method. We evaluated the performance of the parameter estimators through Monte Carlo (MC) simulations. The proposed approach is fully illustrated two real failure time datasets.
