Automação e Otimização de Controle via MQ e RNA para Redução das Emissões de Gases Causadores de Efeito Estufa (GHG) Geradas por Plantas de Alumínio.

Abstract

Nowadays the regulatory restrictions and global concern with the environment are leading the aluminum industry to develop a sustainable model production, with propose to reduce the environmental impacts of its economic activity. Thus, becomes necessary improvements in the operational and control standards for the aluminium production. These needs have major objectives, decrease green house gases (GHG) energy consumption and increase in productive. As technological alternatives such as smart feeders for Point Feeders pots and the development of new control for automatic adjust of the number of manifolds to be broke in the next cycle for Side Break pots will help to improve the decrease of Green Houses Gases. The smart feeders had a significant decrease in the anode effect frequency and consequently a decrease in anode effect time too. For the VSS Side Break pots were possible to create a decision matrix using the Least Square estimation (LS) of the resistance slope and curvature to adjust the number of manifolds. Another approach that showed promising results in the simulation was the neuronal networks for pattern recognition, especial class knows by probabilistic neural network.