Short Term Electrical Energy Demand Forecasting using Artificial Neural Network Technique
Authors: *Ita, M.A., Adebisi, O.I., Amusa, K.A. And Vincent, R.O.
DOI Info: http://doi.org/10.5281/zenodo.10442738
ABSTRACT
Under-estimation or over-estimation of electricity demand can affect the power infrastructures negatively, misleading the planners and wastes resources. Therefore, a precise and dependable model for forecasting electricity demand becomes highly imperative. This study developed an artificial neural network (ANN)-based short term load forecast model for projecting electrical energy demand. Hourly load data from 132/33 kV Ikeja West Transmission Station, Ayobo, Lagos State and temperature data from Nigeria Meteorological Agency, Oshodi, Lagos State, Nigeria were obtained for six months (April to September 2022). The back-propagation algorithm was used in training the ANN model. The three-layer network was trained using 80% of the load and temperature data while validation and testing of the model employed 10% each of the data. The model accuracy test was performed using five metrics which includes mean square error (MSE), mean absolute error (MAE), mean absolute percentage error (MAPE), chi square (X2) and F- test. The results obtained revealed that the developed model performed excellently with minimal error of 2.54, 1.04, 3.11%, 1.55 and 1.26 for MSE, MAE, MAPE, X2 and F-test respectively. Indication from the results shows that ANN model when properly trained with the appropriate data has the potential to effectively predict electrical energy demand on short term basis.
Affiliations: Department of Electrical and Electronics Engineering, Federal University of Agriculture, Abeokuta, Nigeria.
Keywords: Artificial Neural Network, Back Propagation Algorithm, Electrical Energy Demand, Load Forecasting, Short Term
Published date: 2023/12/30
