Transient Flow Structures in Continuous Casting of Steel

Sivaraj Sivaramakrishnan,Hua Bai,Brian G. Thomas,S. P. Vanka,Pierre H. Dauby,Mohammad B. Assar
2000-01-01
Abstract:Transient flow events can be very important to the generation of quality problems during the continuous casting of steel. In this work, several different tools are applied to investigate these phenomena. Transient flow is computed using Large Eddy Simulation (LES) models, while conventional K-e models yield time-averaged results. Particle Image Velocimetry (PIV) is applied to measure quantitatively the transient velocity fields in a water model of the nozzle and mold region. Electromagnetic (MFC) sensors on the wideface of an actual slab caster are used to measure the liquid steel velocity at four locations. Results using all four methods compare favorably for single-phase flow and give new insight into the flow phenomena. The slide gate creates a strong swirl at the outlet ports of the nozzle, which is also predicted using the K-e model. This swirl is seen to persist more than halfway across the mold, causing a characteristic staircase velocity vector pattern in the PIV measurements when viewed in a plane parallel to the wide faces. Flow across the top surface was found in PIV to contain periods of 5-10s when the velocities were three to four times their mean values. This is likely related to inlet conditions and would likely exacerbate shear entrainment of the liquid flux at the top surface and level fluctuations. Simulations of the MFC output indicate that accurate flow prediction is not possible unless the sensors are located in a region of relatively uniform flow, such as near the top surface. In both LES and PIV, the upper roll structure evolves chaotically between a single large recirculation structure and a set of distinct vortices. The lower rolls in PIV are significantly asymmetric for very long periods of time (~ 1-hour) and go through a repeating sequence of features. One of these features involves a short circuit between the upward and downward flow in the lower roll, which is also seen in the simulation. This appears to be inherent to the turbulent nature of the flow and is likely important to inclusion particle and bubble entrapment.
What problem does this paper attempt to address?