The Development of Time-Series Based Outage Distribution Models for Feeder Lines In South Western Nigeria

Abstract

This paper reports the development of a Time-Series based outage forecast model on electrical power distribution feeder lines in the Molete District of the network of the Power Holding Company of Nigeria (PHCN) within the city of Ibadan in Nigeria. A five-year(1999-2003) monthly outage data (in hours per month) gathered on the feeder lines was analyzed with the Time-series analysis program in the statistical software package Systat to generate the autocorrelation and the model parameters. These parameters are: smoothing coefficient, Trend coefficient, seasonality index and standard error of the forecast. The models were developed using the parameters. The models were then used to obtain a 24-month outage forecast figures and plots for each of the ten feeder lines. The results indicate that seasonal factors influenced the monthly outage on the feeder lines with the forecast figures rising to their peaks within the rainny season months of June to August while the minimum forecast figures occur during the dry season period which is from November to February on each feeder line. The outage hours on the feeder lines increased with the lengths of the lines with the outage forecast figures rising to as much as 392 hours per month and 328 hours per month respectively on CRIN and Ring-Road, the two longest feeder lines, in the rain season while they were as low as 51.6 hours per month and 64.77 hours per month respectively on Black-House and Best-Oil, the two shortest feeder lines within the same period.