Performance of Reliability Models for Repairable Systems: Dealing with Information Uncertainty

The paper presents an approach regarding the reliability estimation of complex repairable equipment of electrical utilities. Large quantities of maintenance data have been acquired over the past decades through maintenance data collecting systems. Before using the data, attention must be paid to the uncertainties in data and biases regarding the reliability estimates they generate. The aim of this paper is to propose a novel methodology regarding reliability assessment and evaluation of critical repairable equipment to support maintenance strategy considering uncertain maintenance data. The method integrates two different statistical models for analyzing historical failure data considering their uncertainties. A new validation framework is proposed to evaluate model performance considering those uncertainties. A case study is presented for Hydro-Québec's major transmission equipment. The limits of the models and their dependability on information uncertainty are discussed.