A Near Field Lagrangian Dispersion Model for Sprays

Typically, RANS-based Lagrangian-Eulerian (LE) models of spray atomization are performed with the near-field region unresolved. As a result, physics, including radial dispersion, is not well captured. A new Near-field Lagrangian Dispersion Model (NFLDM) is presented to improve stochastic radial liquid dispersion. This Langevin-based model employs self-similar representations of mean velocity and Reynolds stress fields obtained from experimentally validated Volume of Fluid (VoF) simulations. Additionally, the model employs an autocorrelation function for computational time-step independence. A generalized model of the self-similarity is developed using VoF data from a range of spray simulations employing 2-4.5 MPa ambient pressure, a range of injection velocities, injections of n-dodecane and methanol, and multiple nozzles. The NFLDM is first compared against LE model results representative of typical industry simulations by using temporally-averaged Projected Mass Density (PMD) measurement, and the NFLDM is found to yield significant improvement over the standard LE. Then, the NFLDM is assessed over a wider range of conditions, yielding satisfactory PMD agreement with the VoF for the majority of the simulations.