Utilização de aprendizado de máquina para classificação de perfis de consumo de energia elétrica nas diferentes regiões do Brasil

Abstract

The accurate forecasting of energy consumption can significantly contribute to improve distribution management, and potentially contribute to control and reduce energy consumption rates. Advances in data-based computational techniques are becoming increasingly robust and popular as they achieve good accuracy in results. Among these techniques, Machine Learning (ML) techniques have been widely used in several different domains. This study proposes the development of a classification model that is capable of classifying energy consumption profiles, using Machine Learning methods. The application of Machine Learning techniques in energy production can indicate great potential for controlling and managing the production and distribution of electric energy, which can bring greater efficiency, improve production and optimize distribution. In this study, we combine ML techniques the transfer learning, that is able to use pre-established knowledge in new contexts (different Brazilian regions), making the energy forecasting process more efficient and robust. The application of transfer learning resulted in average accuracies above 90% in the Bagging, Boosting, Random Forest methods for all data used as transfer targets