The working state of the blast furnace (BF) hearth is vital for achieving high efficiency, quality, low fuel consumption, and extended lifespan in BF production. In this study, optimizing the BF hearth by considering three key factors is focused on: hot metal temperature (HMT), silicon content ([Si]), and hearth activity index (HAI). A multi‐objective and multistep optimization method, combining prediction models of HMT, [Si], and HAI with a genetic algorithm, is proposed to enhance the hearth's performance. A Pareto optimal solution set screening method is also established based on target thresholds, weights, and operation priorities, improving its applicability in industrial production. Through this model, significant improvement in hearth working state is demonstrated, verified through case analysis and industrial applications. The prediction model achieves high accuracy in forecasting HMT, [Si], and HAI, with hit rates of 92.26%, 93.45%, and 94.64%, respectively. During the application, the pass rates of these indicators increase by 8.32%, 16.67%, and 13.51%, showcasing the effectiveness of the approach in industrial settings.
Read full abstract