Forced Outage Analysis of Brazilian Thermal Power Plants using the Kruskal-Wallis Test

Abstract

In this paper, the forced outage data from the Brazilian National Interconnected System (SIN) is provided and the statistics from the thermal power plants are computed. The SIN is characterized by a marked seasonality in electricity supply. In addition, the expansion pattern of the Brazilian electric sector shows signs of exhaustion, and the demand for flexible thermal power plants, based on availability, requires outage management as a reliability-centered maintenance (RCM) strategy. In this work, the non-parametric Kruskal-Wallis test and Dunn’s pairwise-comparisons were chosen for evaluating the mean forced outage duration (MFOD) and the forced outage factor (FOF) using the data from the national electricity system operator (ONS) with R Software. The distribution fitting was provided using Weibull++ software from Reliasoft. Based on the MFOD and unit failure rate data, the FOF for Brazilian thermal power plants is 3.33% with a 90% probability and 95% confidence level. Finally, Brazilian thermal power plants were benchmarked against North American power plants.