Time to Event Analysis for Failure Causes in Electrical Power Distribution Systems

Abstract

Electricity is a fundamental resource for modern society. However, some threats are faced by Electrical Power Distribution Systems, which are responsible for delivering electricity to end consumers. Analyzing how much time these hazards will threaten these systems, causing failure events, is an essential area of study. Through statistical methods, it is possible to study this behaviour from time until failure, as well as to observe the influence of variables at this time,providing models to predict when a failure event will occur. In this study, Reliability Analysis Regression techniques are used on real data, constructing a model for all failures and for different groups of failures, using non-parametric and parametric methods to estimate the reliability and cumulative hazard curves. An analysis of the failure causes directly linked to weather events, using six weather variables, is also made.