Numerical simulation of primary breakup of nonturbulent liquid jets in high-viscous gaseous crossflows
Mohammad Hashemi  1@  , Mehdi Jadidi  1@  , Ali Dolatabadi  1@  
1 : Department of Mechanical, Industrial and Aerospace Engineering, Concordia University
Department of Mechanical, Industrial and Aerospace Engineering, Concordia University, Montreal, QC, Canada H3G 1M8 -  Canada

In suspension plasma spray process, which is used for deposition of pseudo-eutectic composition as a thermal barrier coating, producing a controllable and non-pulsating spray is of a great importance. In order to generate nano-structured coatings, suspensions and solution precursors are injected into high-viscous plasma crossflows where the gas-flow Reynolds number approximately varies from 10 to 100. The coatings' quality strongly depends on the breakup of liquid jets in plasma crossflows. Therefore, an investigation into the impact of high-viscous gaseous crossflows on liquid jet atomization can be considered as a fundamental work to control and improve the recent thermal spray processes. In the absence of detailed atomization measurements, high-fidelity numerical simulations have provided a profound and comprehensive picture of atomization mechanisms in recent years. Although the numerical simulations of atomization process offer several advantages over experimental and analytical studies, they are computationally too expensive due to the huge number of cells that should be generated around the droplets and ligaments. In this work, an Eulerian-Lagrangian solver is used to tackle this problem. The motion of fluids and capturing the interface are solved by an open-source Eulerian computational fluid dynamics (CFD) code, called Basilisk, based on finite volume method. In the Eulerian solver, the incompressible variable-density Navier–Stokes equations are solved by using CFL-limited time step, Bell-Colella-Glaz advection scheme and the implicit viscosity formulation. The interface is tracked by a conservative, non-diffusive geometric volume-of-fluid (VOF) scheme. It is worth mentioning that Basilisk code has an algorithm which recognizes and counts each individual ligament or droplet. This feature provides the exact statistics of droplets without any assumption about its probability distribution. In this study, the fine droplets are filtered based on size and shape thresholds using the mentioned feature. Then, they are used as an input data for the Lagrangian solver and modeled by a Lagrangian point particle (LPP) approach. At first, the detailed simulation results at high density ratio (~845) are validated with previous measurements of surface wavelength, breakup location and column trajectory. Then, a parametric study is done on the atomization process in a wide range of gas-flow Reynolds numbers (5.5 − 5500) at high density ratio. The size, velocity and mass rate of droplets formed along jet column are studied and compared with experimental measurements. Ultimately, the connections between the effect of gas viscosity with the underlying jet deformation and breakup physics are found.

Online user: 1 RSS Feed