Identificação de sistemas dinâmicos não lineares para o controle preditivo de climatização, usando modelos Hammerstein

Abstract

This work aims at the development of a methodology to obtain, automatically, the non-linear static part of a model in Hammerstein for the predictive control of climatization in non-residential buildings, based on machine learning. The objective of this work is to help the achievement of an eventual alternative to the existing model, developed through Genetic Programming. The work is motivated by the need to reduce the energy consumption and emissions of carbon dioxide (CO2) into the atmosphere by buildings, given the considerable increase in the number of buildings built and their current impact on the environment.