<p>Secondary cooling during continuous casting is a delicate process because the cooling rate of water spray directly affects the slab surface and internal quality. Undercooling may lead to slab surface bulging or even breakout, whereas overcooling can cause deformation and crack of slabs due to excessive thermal residual stresses and strains. Any slab which does not meet the required quality will be downgraded or scrapped and remelted. In order to remain competitive and continuously produce high-quality and high-strength steel at the maximum production rate, the secondary cooling process must be carefully designed and controlled. Efficient and uniform heat removal without deforming or crack the slab is a significant challenge during secondary cooling. In the meantime, the on-site thermal measurement techniques are limited due to the harsh environment. In contrast, experimental measurements are only valid for the tested conditions, and the measurement process is not only labor-intensive, but the result might be inapplicable when changes in the process occur. On the other hand, the high-performance computing (HPC)-powered computational fluid dynamics (CFD) approach has become a powerful tool to gain insights into complex fluid flow and heat transfer problems. Yet, few successful numerical models for heat transfer phenomena during secondary cooling have been reported, primarily due to complex phenomena. </p> <p> </p> <p>Therefore, the current study has proposed two three-dimensional continuum numerical models and a three-step coupling procedure for the transport of mass, momentum, and energy during the secondary cooling process. The first numerical model features the simulation of water spray impingement heat and mass transfer on the surface of a moving slab considering atomization, droplet dispersion, droplet-air interaction, droplet-droplet interaction, droplet-wall impingement, the effect of vapor film, and droplet boiling. The model has been validated against five benchmark experiments in terms of droplet size prior to impingement, droplet impingement pressure, and heat transfer coefficient (HTC) on the slab surface. The validated model has been applied to a series of numerical simulations to investigate the effects of spray nozzle type, spray flow rate, standoff distance, spray direction, casting speed, nozzle-to-nozzle distance, row-to-row distance, arrangement of nozzles, roll and roll pitch, spray angle, spray water temperature, slab surface temperature, and spray cooling on the narrow face. Furthermore, the simulation results have been used to generate a mathematically simple HTC correlation, expressed as a function of nine essential operating parameters. A graphic user interface (GUI) has been developed to facilitate the application of correlations. The calculated two-dimensional HTC distribution is stored in the universal comma-separated values (csv) format, and it can be directly applied as a boundary condition to on-site off-line/on-line solidification calculation at steel mills. The proposed numerical model and the generic methodology for HTC correlations should benefit the steel industry by expediting the development process of HTC correlations, achieving real-time dynamic spray cooling control, supporting nozzle selection, troubleshooting malfunctioning nozzles, and can further improve the accuracy of the existing casting control systems.</p> <p> </p> <p>In the second numerical model, the volume-averaged Enthalpy-Porosity method has been extended to include the slurry effect at low solid fractions through a switching function. With the HTC distribution on the slab surface as the thermal boundary condition, the model has been used to investigate the fluid flow, heat transfer, and solidification inside a slab during the secondary cooling process. The model has been validated against the analytical solution for a stationary thin solidifying body and the simulation for a moving thin solidifying body. The effects of secondary dendrite arm spacing, critical solid fraction, crystal constant, switching function constant, cooling rate, rolls, nozzle-to-nozzle distance, and arrangement of nozzles have been evaluated using the validated model. In addition, <a>the solidification model has been coupled with the predictions from the HTC correlations, and the results have demonstrated the availability of the correlations other than on-site continuous casting control. </a>Moreover, the model, along with the three-step coupling procedure, has been applied to simulate the initial solidification process in continuous casting, where a sufficient cooling rate is required to maintain a proper solidification rate. Otherwise, bulging or breakout might occur. The prediction is in good agreement with the measured shell thickness, which was obtained from a breakout incident. With the help of HPC, such comprehensive simulations will continue to serve as a powerful tool for troubleshooting and optimization.</p>