Quality Prediction for Blast Furnaces

Changes in operational conditions of the blast furnace may interfere in the grade of silicon in an unpredictable way. The furnace manager may then have to spend several hours to bring the furnace's output quality back to desirable ranges. Such a delay may cause one or two runs to have pig iron with silicon grade out of the expected specification, a costly loss.

SAD Screenshot SAD is a decision support system that production managers can use to forecast the silicon grade in the upcoming runs, reducing the variability of this indicator in the final product, therefore improving the quality of the production process. SAD can be used to explore the effect of different possible corrections, evaluating the best way to fix deviations.

The system acquires data in raw format directly from the supervisory system database, and all data preprocessing is done automatically, simplifying integration.