MODELO DE PREVISIBILIDADE DE GERAÇÃO DE GÁS DE COQUERIA

Abstract

The steel sector faces critical moments in the midst of the economic crisis, being necessary to adapt to the market, seeking every day its competitiveness. With this, it is necessary to walk with the processes seeking the optimization of the raw material mixes and the lowest costs in the processes. Coke is a plant in the steel sector that compromises much of the final cost of the product. The need to guarantee the lowest cost and the best possible yield is analyzed by the processes and their respective raw materials. Aware of this need, this work aims to verify the existence of linear correlation of variables with the generation of coke oven gas, and from this, formulate an equation that translates the appropriate model, which allows a better prediction of coke oven gas generation, and can improve The prediction of the energy matrix and to compose the optimization of the mix of coals used in the plant. The analysis will be performed based on statistical analysis of historical data, collected from 01/10/2012 to 09/30/2016. With the statistical analysis of the data, using a multiple linear regression, a model was proposed that estimates the coke oven gas generation, from the coke production, the moisture of the coal mixture, the percentage of low volatile coal in the coal mixture and Of the volatile matter of the coal mixture. It was possible to obtain results that demonstrate coherence between the responses of the obtained models and the real system responses, thus being able to have a more assertive forecast of this generation. Keywords: Production