Modeling and Forecasting of Short-Term Half-Hourly Electric Load at the University of Ibadan, Nigeria.

ABSTRACT In this study, the short-term load pattern for the University of Ibadan was investigated and a multilayered feed-forward artificial neural networks (ANN) model was developed to forecast the time series half-hourly load pattern of the system using the load data for a period of 5 years (2000 to 2004). The study showed that the mean half hourly load for the period of study ranged between 1.3 and 2.2 MW, and the coefficient of determination (R2 -values) of the ANN predicted and the measured half-hourly load for test dataset decreased from 0.6832 to 0.4835 with increase in the lead time from 0.5 to 10.0 hours.