Continuous Caster Mould Digital Twin Development for Fluid Flow Control and Sliver Defect Minimization

钢铁制造商不断面临着在提高连铸机生产率的同时保持钢坯质量的挑战。这对于拉拔和熨烫(D&I)和超低碳(ULC)板坯来说尤其具有挑战性，因为客户对质量提出了严格的要求。这些板坯是扁平轧制成薄板(小于0.5 mm)，极易出现内部和外部缺陷。因此，提高这些牌号的拉速往往会对劣质成本造成不利影响。这些缺陷是由于连铸结晶器内的凝固钢壳卷住了非金属夹杂物(NMI)造成的，这些非金属夹杂物(NMI)通过浸入式水口(SEN)不断接收钢液和Ar气体。在较高的拉速下，结晶器内气泡状流的湍流和不稳定性显著增加。这加剧了捕获在液池内循环的NMI颗粒的风险。因此，连铸结晶器内的流动必须进行优化，然后在高连铸机吞吐量的条件下进行控制。连铸机数字孪晶及其应用将致力于提高AMD板坯和卷板的产品质量。这项研究计划还将利用物理模拟方法进行流动可视化，并针对不同的铸造条件量化和开发流动指数。还将开发一种基于降阶模型的CFD来预测钢中的缺陷(即结晶器壁附近的NMI颗粒捕获)。如果通过CFD预测NMI的包裹率，则将对流动指数进行实时控制和调整。在连铸机上，项目成果有望提高D&I和ULC钢种的最大拉速，减少缺陷，并开发实时缺陷预测技术。这将通过确定改进的SEN设计、优化的Ar流量和更好的工艺控制技术来实现。提高钢材质量带来的成本节约预计将超过250万美元/年。拟议的研究项目与AMD的战略愿景和业务目标保持一致，将为位于加拿大安大略省的AMD带来立竿见影的好处。

Steelmakers are constantly challenged to increase the productivity of their continuous casters (CC) while simultaneously maintaining steel slab quality. This is especially challenging for Drawn & Ironed (D&I) and Ultra Low Carbon (ULC) slabs because of stringent quality demands imposed by the customers. These slabs are flat rolled into thin sheets (less than 0.5 mm) and are extremely susceptible to internal and external defects. Therefore, increasing the casting speed for these grades often results in a detrimental impact on cost of poor quality. The defects are created by entrapment of non-metallic inclusions (NMI), such as alumina and mould slag, by the solidifying steel shell inside the CC mould, which continuously receives liquid steel and argon gas through a submerged entry nozzle (SEN). Bubbly steel flow turbulence and instabilities inside the mould are significantly increased at higher casting speeds. This exacerbates the risk of entrapping NMI particles circulating inside the liquid pool. Thus, the flow inside the CC mould must be optimized and then controlled at high caster throughput conditions. The continuous caster digital twin and its application will aim to improve product quality for AMD's slabs and coils. This research program will also utilize a physical modeling approach for flow visualization and quantification and development of a flow index for various casting conditions. A reduced order model-based CFD will also be developed for predicting defects (i.e. NMI particle capture near mould walls) in steel. The flow index will be controlled and adjusted in real time if NMI entrapment is predicted by CFD. At the caster, the project outcomes are expected to enable the increase of maximum casting speed of D&I and ULC steel grades, reduction in defects, and development of real time defect prediction techniques. This will be achieved by identification of improved SEN designs, optimized argon flow rates and better process control techniques. The cost saving arising from improving steel quality is expected to exceed $2.5 million/year. The proposed research project is aligned with AMD's strategic vision and business goals, and will create immediate benefits to AMD, located in Ontario, Canada.