Per: ALOISIO SIMOES RIBEIRO (RHI Magnesita), Eduardo Roberto Menezes (), Mateus dos Santos Souza (), Lucas Gabriel Seibert (), Victor Luiz Cruz Morais (), Anderson Carvalho Nogueira (), Alex Martins da Silva (), Rodrigo Dalla Vecchia (), Bruno Moreira Nabinger ()
Abstract:
There is significant interest in accurately modeling the operational variables of the Ironmaking process to predict the refractory wear in blast furnace main runners. Understanding refractory wear variables on blast furnace main runners can allow us to promote safe, low-cost, sustainable, high-performance and quality iron ore processing in the blast furnace. Despite this, the task is challenging due to the complex interactions between process variables, which are not entirely comprehended. Often, decisions in the industry are grounded in experience or undirect methods. This study aims to introduce a robust model that can effectively guide process engineers by forecasting the influence of ironmaking variables in the refractory wear of the blast furnace main runners. Based on an algorithm that combines unsupervised learning techniques for clustering and a meta-regressor, it was possible to obtain a high accurate model to predict the refractory wear in the blast furnace main runner, resulting in valuable information to decision-making process regarding the blast furnace main runner campaign