About 70% of the steel used worldwide is produced via the Basic Oxygen Furnace (BOF). This process has limited automation, and normal operation relies on invaluable operators’ knowledge and past experience. In this paper, we present a framework for dynamic optimization of BOF operation. The dynamic optimization problems utilize a first-principles based dynamic model, and are solved using a hybrid method in which the states are integrated using a differential–algebraic equation (DAE) solver over an initial time interval and full discretization over the remainder of the time horizon. Optimization case studies are presented in which the impact of constraints and different objective formulations are explored. The results suggest that the proposed framework can potentially aid steelmakers to significantly reduce process costs while meeting production and quality targets. The framework is implemented in Python using the open-source software tool, CasADi.

全球约70％的钢材是通过碱性氧气炉（BOF）生产的。此过程的自动化程度有限，正常操作依赖于宝贵的操作员知识和过去的经验。在本文中，我们提出了BOF操作动态优化的框架。动态优化问题利用基于第一原理的动力学模型，并使用混合方法解决，其中在初始时间间隔内使用微分-代数方程（DAE）求解器对状态进行积分，并在其余时间进行完全离散化地平线。提出了优化案例研究，其中探讨了约束和不同目标公式的影响。结果表明，提出的框架可以潜在地帮助钢铁制造商大幅降低过程成本，同时达到生产和质量目标。该框架使用开源软件工具CasADi在Python中实现。