Aerodynamic Breakup of Vaporizing Micro-Drops : Effects of Reynolds and Stefan Numbers
Aerodynamic deformation and breakup are commonly observed phenomena in droplet and spray applications. In the past, research efforts have mainly focused on millimeter-sized drops, identifying a wide variety of breakup topologies and characterizing the breakup modes primarily through the Weber and Ohnesorge numbers. Much less attention has been paid to micro-drops, where the drop sizes are tens of microns, and the conventional assumption is that these tiny drops are too small to break. Nevertheless, for some applications, such as fuel injection and drop impact in high-speed flights, the relative velocity between the drop and air can be sufficiently high that even micro-drops can break. Furthermore, these applications often involve high-temperature flows, where aerobreakup is accompanied by significant heat and mass transfer. The Stefan flow, induced by phase change at the drop surface, significantly influences interfacial instability, drop deformation dynamics, and the resulting breakup morphology and outcomes.sThis talk will address our recent simulation results on the aerobreakup of vaporizing micro-drops, with the goal of systematically characterizing the effects of Reynolds and Stefan numbers on drop deformation and breakup dynamics. The sharp liquid-gas interface with phase change is tracked using the geometric Volume-of-Fluid (VOF) method. The Stefan flow at the interface is accounted for by solving the pressure Poisson equation with an additional source term due to phase change, using a conservative and compact source distribution method. The two-phase flow model is implemented in the adaptive multiphase flow solver, Basilisk, and comprehensive high-fidelity simulations were performed across various Weber, Reynolds, and Stefan numbers. The simulation results indicate that both a decrease in Re and an increase in St contribute to stabilizing drop breakup and lead to an increase in the critical Weber number. We will also present our efforts in utilizing these results to develop data-driven Lagrangian droplet models. To reduce the amount of data fed to the network, spherical harmonics are used to characterize the drop shape. The temporal data of the spherical harmonic modal coefficients, along with the drag coefficient, are extracted from the simulations and used to train a Nonlinear Autoregressive Network with Exogenous Inputs (NARX) recurrent neural network. The machine-learning (ML) model predictions show excellent agreement with the high-fidelity simulation results.
Contact Nathanaël Machicoane for more information or to schedule a discussion with the seminar speaker.