Simulation of Slag Droplet Entrainment by Volume of Fluid and Lagrangian Particle Tracking Coupled Modeling.

The slag droplet entrainment is a common phenomenon in steel refining processes, which may lead to inclusions and defects. In the multiphase flow system, the distinct interface and tiny blobs possess a wide range of spatial and temporal scales and make it hard to be simulated. In numerical methods, the volume of fluid (VOF) approach is appropriate for capturing the interface, but for the unresolvable tiny blobs, the Lagrangian particle tracking (LPT) is preferable. This work newly implements a bidirectional VOF-LPT transformation algorithm for developing a multiscale solver in OpenFOAM to simulate the slag droplet entrainment. The interIsoFoam solver is selected as the main solver to resolve the interface, and the resolution is improved with using the geometric reconstruction and the adaptive mesh refinement (AMR). For capturing tiny droplets, a connected component labeling (CCL) method is adopted for detecting discrete droplets in the VOF field, and then the VOF-to-LPT transition takes place for saving computational costs. Conversely, the LPT-to-VOF transformation for droplets touching the interface is also incorporated to achieve the bidirectional transition. The solver is first validated by a simple case, indicating that the two-way transition algorithm and the Eulerian-Lagrangian momentum coupling are accurate. Then the solver is applied to simulate the slag layer behavior for revealing the mechanisms of slag droplet formation and entrainment. Two main mechanisms of slag droplet formation are identified, and it is found that fewer discrete droplets are generated when the surface tension increases.

Investigation of cavitation noise using Eulerian-Lagrangian multiscale modeling.

We have employed the large eddy simulation (LES) approach to investigate the cavitation noise characteristics of an unsteady cavitating flow around a NACA66 (National Advisory Committee for Aeronautics) hydrofoil by employing an Eulerian-Lagrangian based multiscale cavitation model. A volume of fluid (VOF) method simulates the large cavity, whereas a Lagrangian discrete bubble model (DBM) tracks the small bubbles. Meanwhile, noise is determined using the Ffowcs Williams-Hawkings equation (FW-H). Eulerian-Lagrangian analysis has shown that, in comparison to VOF, it is more effective in revealing microscopic characteristics of unsteady cavitating flows, including microscale bubbles, that are unresolvable around the cloud cavity, and their impact on the flow field. It is also evident that its evolution of cavitation features on the hydrofoil is more consistent with the experimental observations. The frequency of the maximum sound pressure level corresponds to the frequency of the main cavity shedding for the noise characteristics. Using the Eulerian-Lagrangian method to predict the noise signal, results show that the cavitation noise, generated by discrete bubbles due to their collapse, is mainly composed of high-frequency signals. In addition, the frequency of cavitation noise induced by discrete microbubbles is around 10 kHz. A typical characteristic of cavitation noise, including two intense pulses during the collapsing of the cloud cavity, is described, as well as the mechanisms that underlie these phenomena. The findings of this work provide for a fundamental understanding of cavitation and serve as a valuable reference for the design and intensification of hydrodynamic cavitation reactors.