Model Based on Artificial Neural Networks for Forecasting Electricity Consumption: A Holistic Approach

Daniel Orlando  Garzón Medina; Jose  Calixto Lopes; Thales  Sousa

doi:10.48011/sbse.v1i1.2495

Daniel Orlando Garzón Medina Federal University of ABC
Jose Calixto Lopes Federal University of ABC
Thales Sousa Federal University of ABC

DOI: https://doi.org/10.48011/sbse.v1i1.2495

Keywords: Artificial intelligence, Artificial neural networks, Load forecasting, Multimodel forecasting, Residential forecasting

Abstract

Electrical demand forecasting is a key tool in making operational and strategic decisions in power companies, whose lack of accuracy can lead to high economic costs. In this sense, forecasting allows network operators to make power dispatch, maintenance program, reliability analysis and operational safety decisions. Therefore, the present work proposed the use of Artificial Neural Networks (ANN) to project the demand of the Colombian residential sector. The model presented for the forecast was based on socioeconomic variables obtained from official Colombian government data sources such as population growth, gross domestic product and residential electrical consumption. The work was developed with the aid of the MATLAB® software where a model with appreciable assertiveness margin were proposed.