Modelling of Failures of Electric Network Elements Based on Regression Models Accounting the Trend

Abstract

The article analysed various approaches to modelling emergency outages on electrical networks, some of which were previously used by the authors to predict and probabilistic assessment of the studied parameter. Based on the correlation analysis, it was revealed that the number of emergency shutdowns that took place in the city of Irkutsk in 2010–2017 is significantly affected by the heat parameter, which is taken as the sum of the average monthly temperatures. Since it was previously revealed that in the series of emergency outages for some years and months there are significant linear and polynomial trends, time is taken as the second factor. According to the results obtained, linear and nonlinear factor models calculated from data for a given year are equally significant. Monthly series of outages in most cases more accurately describes models with a polynomial trend. The obtained regression equations can be used to model emergency outages with a given probability and to assess the risks of their occurrence.