Statistical analysis of forced outage durations of Eskom subtransmission transformers

Abstract

The main aim of this paper is to analyse the duration of forced outages of Eskom subtransmission transformers. Statistical techniques are used to analyse events recorded from 2007–2016. For various event variables, boxplots were plotted for various levels to understand and compare the distribution and spread of durations. Thereafter, the Kruskal–Wallis test was conducted to assess whether there was a variation in the durations for various levels of variables. Finally, multiple linear regression models were constructed with each variable as an independent variable against the duration as the dependent variable. The coefficients of determination for the models were calculated and used in ranking variables as predictors of duration. The study has demonstrated how insight into the duration of forced outages can be obtained by using statistical techniques. This insight can provide useful inputs to asset and performance managers when developing strategies for reducing the duration of forced subtransmission transformer outages.

Keywords

• Boxplots
• Coefficient of determination
• Kruskal–Wallis test
• Multiple linear regression
• Transformer forced outage