SHORT-TERM LOAD FORECASTING OF THE POWER HOLDING COMPANY OF NIGERIA ELECTRIC NETWORK USING ARTIFICIAL NEURAL NETWORK

Abstract

This paper is concerned with the application Of Artificial Neural Network (ANN) technique to forecast the short- term electrical load of the power Holding Company of Nigeria (PHCN) network. Electric power load forecasting is essential for the purpose Of optimal planning and operation Of a large system. It is also g useful tool for electric utilities in several applications such as economic allocation of generation, energy transaction, unit commitment. maintenance scheduling and optimal energy interchange between utilities. Important factors considered in this investigation include past load historical data. season and weather information and effects. Multi-layer Perception (MLP) was used in the implementation of the Feed Forward Backward propagation in order to create the ANN algorithm using Matlab-based software. Historical load data from National Control Center (NCC) Osogbo was used in the training of this network for Morning and Evening load patterns. ANN algorithm was grained with different hidden neurons ranging from 2-10 neurons. The best forecast was obtained using 2 hidden neurons for Morning and Evening short term load forecast and the average percentage errors were 1.68% and 3.90% as compared to the ones obtained from PHCN load forecasting techniques which were 50.56% and 70.42% respectively