LONG TERM LOAD FORECAST USING ARTIFICIAL NEURAL NETWORK METHOD: RAINBOW-ELEKAHIA COMMERCIAL 33KV FEEDER IN PORT HARCOURT, NIGERIA

Abstract

This paper presents the Long-term Commercial Electrical Load Forecast of the Rainbow-Elekahia feeder under the Port Harcourt Electricity Distribution Company (PHEDC) network using Artificial Neural Network (ANN) from 2020 to 2029. The data were trained with the instrumentality of a two-layer feed-forward neural networks curve fitting (FFNNCF) simulation tool within the MATLAB 2020 simulation environment. The historical load consumption of the mentioned feeder and Average Temperature was obtained from PHEDC Head Office Moscow Road, Transmission Company of Nigeria (TCN), Oginigba, and NIMET-Office Abuja; the summation of all these formed the data utilized for Training, Validation, and Testing of the proposed neural networks architecture. The forecasted results obtained prove that Curve-Fitting Neural Network (CFNN) is highly efficient for the commercial long-term load forecast as the justification of low error was investigated(evaluated) using Percentage Error combined with Root Mean Square Error(RMSE) and Mean Square Error(MSE) and the results obtained for the number of years shows that error is minimal. Generally, the forecasted load value for ten (10) years is 7354.3 MWHR which has shown a realistic and genuine forecast process that has been carried out through a reliable ANN application.