Machine Learning Applied to Energy Efficiency of Large Consumers

The use of electric energy is rapidly increasing, so it is essential that users improve their understanding of electric consumption, thus reducing waste and improper use. The present work uses Machine Learning techniques (specifically Linear Regression and Random Forest) to model the relationship between electricity consumption and climatic conditions within the Federal Institute of Education, Science and Technology of Santa Catarina (IFSC) and the Ministries Esplanade (headquarters of the Brazilian Executive, in Brasilia). Historical data of active power, ambient temperature and atmospheric pressure obtained from the database of the project PGEN. The results show that the model allows predicting the instantaneous energy consumption of the localities with an average error of 22.75% KW. The building of the Ministries Esplanade obtained the lower errors, and the campus Florianopolis obtained the bigger erros.