Bayesian approach in the power electric systems study of reliability

Abstract

This work aims to highlight prerogatives and advantages of the Bayesian approach in the reliability studies of the modern power electrical systems. The new organization of the electric energy sector and the consistent degree of technological innovation make the data more uncertain related to the operation of the electric systems components, with the consequent lack of information for an efficient estimation of the reliability according to the classical approach. Sometimes, this uncertainty concerns the same reliability models, missing adequately the validation of the model adaptations. In this paper, referring to the uncertainty of data, we define a general probabilistic model, for reliability studies. Subsequently, Bayesian methodologies are framed in an ampler problem list, based on the definition of an opportune "vector of state" and of a vector describing the system performances, aiming to the definition and the calculation or the estimation of system reliability. The purpose of our work is to establish a useful model based on information "a priori", with respect to the Bayesian approach. This allows getting tools for the identification of probabilistic models, information on usury process and damage models (cumulative damage, multiplicative, etc.), suitable to compensate the lack of data.

Keywords: Reliability - Power System - Bayes Theorem - Weibull Model - Probability