Desenvolvimento de um modelo de predição de condições operacionais de altos-fornos siderugicos com base em redes neurais artificiais / Development of a prediction model for ironmaking blast furnace operational conditions based on artificial neural networks

AUTOR(ES)
DATA DE PUBLICAÇÃO

2006

RESUMO

This work describes the development of a model capable of evaluating and predicting iron ore reduction process in blast-furnaces based on raw materiais characteristics as well as burden and blast composition. It provides a planning and analysing toll to operators and process engineers. The motivation for this development resides on the lack of this kind of model in the ironmaking industry. The many models found are either toa simple and not capable of predicting raw materiais parameters effects on the furnace performance or toa complexo The latter are useful in technology potential identifications but not in the daily work or ordinary operation planning. Blast-furnaces and oxygen converters are the core units in a integrated steel works. The former one produces primary iron from oxides bared by iron ores and the latter one refines molten iron into steel, adjusting its chemical compositiono The developed and applied model is hybrid in nature, combining simulating algorithms based on mass and energy balances with variable lack of fitnesso The value of each lack-of-fitness index is estimated by a neural model in which the input quantify burden materiais characteristics as well as blow and charging conditions. The model has shown its predicting capacity during its industrial application, which lead to higher average productivity and lower specific fuel consumption are expected following better operation action and process planning efficiency. In addition, the model is now being associated to a plant mass balance module to asses alternative operation pattern simulation for the hole company which is a fundamental activity in strategic business planning

ASSUNTO(S)

redes neurais (computador) altos-fornos modelos matematicos artificial intelligence mathematical models siderurgy blast-furnaces inteligencia artificial artificial neural networks siderurgia

ACESSO AO ARTIGO

Documentos Relacionados