Open AccessMedium term electrical load forecast of Abuja Municipal Area council using artificial neural network method
Author(s) -
E.C. Ashigwuike,
A.R.A. Aluya,
J.E.C. Emechebe,
Stephen Adole Benson
Publication year - 2020
Publication title -
nigerian journal of technology
Language(s) - English
Resource type - Journals
eISSN - 2467-8821
pISSN - 0331-8443
DOI - 10.4314/njt.v39i3.28
Subject(s) - artificial neural network , mean absolute percentage error , linear regression , term (time) , statistics , standard deviation , population , computer science , engineering , mathematics , artificial intelligence , physics , demography , quantum mechanics , sociology
This paper presents a medium-term electric load forecast for Abuja Municipal Area Council (AMAC) distribution network based on Artificial Neural Network (ANN). The technique results are compared with that of a conventional method (Multiple Linear Regression method), for the same data. The ANN proposed method takes into account the effect of temperature, time, population growth rate and the activities of different regions of city areas regarding lifestyle and types of consumers. The data of monthly to annual peak values are collected for the period from 2012 to first quarter of 2018. Hence, the Artificial Neural Network method presented a result with average MAPE of 0.00197 while the multiple linear regression having an average MAPE of 0.004545. The R-Value deviation was 8.06% and 34.42% for ANN and MLR methods respectively.
Keywords: Artificial Neural Network, Forecast, Load, Energy Demand, Capacity Allocation, Percentage error, Forecasting Accuracy.