Ronald J. O'Malley

Computational Fluid Dynamics Study of Molten Steel Flow Patterns and Particle-Wall Interactions Inside a Slide-Gate Nozzle by a Hybrid Turbulent Model

Melt flow patterns and turbulence inside a slide-gate throttled submerged entry nozzle (SEN) were studied using Detached–Eddy Simulation (DES) model, which is a combination of Reynolds–Averaged Navier–Stokes (RANS) and Large–Eddy Simulation (LES) models. The DES switching criterion between RANS and LES was investigated to closely reproduce the flow structures of low and high turbulence regions similar to RANS and LES simulations, respectively. The melt flow patterns inside the nozzle were determined by k–ε (a RANS model), LES, and DES turbulent models, and convergence studies were performed to ensure reliability of the results. Results showed that the DES model has significant advantages over the standard k–ε model in transient simulations and in regions containing flow separation from the nozzle surface. Moreover, due to applying a hybrid approach, DES uses a RANS model at wall boundaries which resolves the extremely fine mesh requirement of LES simulations, and therefore it is computationally more efficient. Investigation of particle distribution inside the nozzle and particle adhesion to the nozzle wall also reveals that the DES model simulations predict more particle–wall interactions compared to LES model.

Comparison of CFD Simulations with Experimental Measurements of Nozzle Clogging in Continuous Casting of Steels

Abstract Measurements of clog deposit thickness on the interior surfaces of a commercial continuous casting nozzle are compared with computational fluid dynamics (CFD) predictions of melt flow patterns and particle–wall interactions to identify the mechanisms of nozzle clogging. A submerged entry nozzle received from industry was encased in epoxy and carefully sectioned to allow measurement of the deposit thickness on the internal surfaces of the nozzle. CFD simulations of melt flow patterns and particle behavior inside the nozzle were performed by combining the Eulerian-Lagrangian approach and detached eddy simulation turbulent model, matching the geometry and operating conditions of the industrial test. The CFD results indicated that convergent areas of the interior cross section of the nozzle increased the velocity and turbulence of the flowing steel inside the nozzle and decreased the clog deposit thickness locally in these areas. CFD simulations also predicted a higher rate of attachment of particles in the divergent area between two convergent sections of the nozzle, which matched the observations made in the industrial nozzle measurements.