Effect of Electromagnetic Braking (EMBr) on Turbulent Flow in Continuous Casting

Fluid flow in the mold region of the continuous casting process is responsible for surface defects, slag entrainment, and other steel quality problems. Thus, it is very important to choose nozzle geometries and operating conditions that produce flow patterns within an operating window that avoids these problems. Operating conditions which control mold flow problems include the mold cross section, casting speed, submergence depth, mold powder, argon gas injection and electromagnetic forces. The application of a magnetic field is an attractive method to control mold flow because it is nonintrusive and can be adjusted during operation. However, the application of a magnetic field can change the flow pattern in non-obvious ways. Understanding how a magnetic field affects highly turbulent mold flow is both an important and challenging task. It is difficult to take measurements in operating commercial steel casters, so experimental studies are limited. Physical water models are problematic because water is unaffected by a magnetic field. Conducting fluids, such as tin, mercury 5 and eutectic alloys such as GaInSn, have been used to study the effect of magnetic fields on flow in continuous casters. Numerical studies of mold flow have been extensively used to understand the continuous casting process, including the effect of magnetic fields. Most of the studies exploring mold flow use Reynolds-averaged Navier-Stokes (RANS) or unsteady RANS (URANS) which accurately predicts the mean flow behavior. However, transient behavior and flow stability is more important to mold flow quality, and has received relatively less attention. Only a few recent studies, using Large Eddy simulations (LES) without EMBr and with EMBr, have been performed to understand the transients involved in the process. Cukierski et al. observed that application of local EMBr weakens the upper recirculation region and decreases the top surface velocity. Harada et al. compared the effects of local and ruler EMBr systems and claimed that both configurations increase surface velocities and dampen high velocities below the mold, and that configuring the ruler configuration below the nozzle ports has better braking efficiency and also results in better surface stability. Similar behavior was observed by Chaudhary et al. in a computational model of a physical model with insulated walls. The predictions agreed well with measurements in the same system. However, raising the ruler magnetic field to center it across the nozzle ports resulted in severely unstable transient flow, with large scale fluctuations of the jets and great asymmetries. Adding conducting side walls with the same ruler configuration over the nozzle produced stable transient jet behavior. Li et al. also observed that the incorporation of accurate wall conductivity is necessary as it affects the braking efficiency of the magnetic field. In the current study, mold flow with ruler EMBr fields was simulated using a high-fidelity, fine-grid LES code, CUFLOW, incorporating the influence of the conducting shell. CUFLOW is an in-house Computational Fluid Dynamics code using graphics processing unit (GPU). CUFLOW was first validated by comparing with previous measurements taken in a scaled GaInSn model. It was then applied to simulate the full scale real caster with and without the ruler EMBr field. Time averaged and transient flow patterns, surface velocities, surface level profiles and surface level fluctuations were computed to investigate the effect of ruler EMBr on the details of the flow phenomena, and to investigate similarity criteria for scaleup.