Umudike Journal of Engineering and Technology

The effective performance of any system is determined usually by its present and future reliability status, using appropriate forecast techniques. Reliability is an inverse characteristic of the system failure rates. In recent years in Nigeria, satisfactory degree of reliability has not been obtained in our distribution system, most especially in urban commercial cities. One of the quantitative forecasting techniques is the use of least square regression analysis.  Using Federal University of Agriculture Abeokuta (FUNAAB) 33 kV distribution network in Abeokuta, Ogun state as a case study, the available five years (2010-2014) fault data were used to determine the forecast model for the network. The appropriateness of the initial developed model equations  from the observed fault data were diagnostically checked using minimum sum of the  error square () and coefficient of the determination (R²). Where the model equation fails to satisfy R² conditions, R² adjusted () was used.  Results from the analysis showed that the technique produced satisfactory forecast model equations. The results from the analysis also showed that the mean time between failures reduces in hours as failure rate increases while reliability will decline from 74.79% in 2010 to approximately 0% in 2019, if no urgent actions are taken.