Determinação da janela ótima de manutenção para sistemas reparáveis na presença de covariáveis

Abstract

An important field of application of reliability is in maintenance data for industrial systems and equipment. Most of the time, these systems undergo several preventive and corrective maintenance throughout their useful lifespan, and, for this reason, they are classified as repairable systems. This thesis presents a literature review with practical emphasis on the application of the Power Law Process for modeling repairable systems in two different software: Minitab for its graphical interface and wide application in industry; and R Studio, software that uses R language for the most diverse applications in statistics. The thesis presents as a deepening the proposal of a model that considers the inclusion of covariates in the Power Law Process’s intensity function assuming a single shape parameter for different levels that can represent batches, models, manufacturers, types of components, etc. in practical applications. As a first case study, real maintenance data for electrical transformers presented by Gilardoni and Colosimo (2007) are used to replicate the same analysis presented in their article. In the second case study, failure data for locomotive brake components of two distinct types are analyzed using the two versions of the Power Law model: the standard and the one with covariates.