Per: MARCOS ANTÔNIO RODRIGUES (TERNIUM BRASIL LTDA.), LEONCIO MACHADO DE REZENDE (TERNIUM BRASIL), Vilson José Anunciação (TERNIUM BRASIL), André Machado da Silva (TERNIUM BRASIL), Evandro da Silva do Carmo (TERNIUM BRASIL), Arthur Araujo Maia Farah (Vetta - SMS), Fernanda da Silva Machado (Vetta - SMS)
Abstract:
The objective of this work will be to present the development of the project, which aims to optimize the generation and distribution of steelmaking gases and vapors from production processes at Ternium Brasil. Using machine learning techniques to create prediction, simulation and optimization models, as well as techniques that make it possible to evaluate scenarios that maximize the economic viability of operations, enabling the reduction of the burning of steel gases in flares, and their subsequent replacement by gas natural, maximizing the greatest generation of electrical energy with the energy of available waste gases. The project aims to optimize steel gases and vapors based on the production planning of production areas, although forecast models use production orders to define consumption and generation forecasts, these are only input variables for the dispatch optimizing system. of gases/vapours, and under which no decisions or controls are made.
The seasonality of Steelmaking gas generation combined with its complex analysis of energy generation generated by race, as well as the logistics of distribution of steelmaking gases from the Ternium Brasil site, was exponentially one of the most important and relevant factors for the development of this project Machine Learning that uses algorithms to assimilate changes in the generation and consumption scenarios of steelmaking gases and vapors.