Numerical investigation of primary break-up of conical swirled jets

Primary break-up is a hard element to be described into the atomization chain, since experimental works are rare especially for swirl atomizers. In the past, different models had been developed to define the mechanism which leads to break-up of jets and, in this way, the characteristics of the sub-sequent produced spray. These models had been validated against experimental data in simplified conditions, for example using round jets; however they cannot be generalized for all the other categories. Moreover, these models were based on simplified assumptions, for example they neglected turbulence. Thus their application to conical swirled jets is tricky and could produce misleading results. In absence of experimental data, Volume of Fluid Direct Numerical Simulations (VOF DNS) could help to provide more information about the produced primary spray and its characteristics, such as droplets velocity components, location, size and shape in the whole investigated domain. However, in order to simulate conical swirled jets from aeronautical pressure swirl atomizers, realistic velocity profiles of both liquid and gas phases together with the characteristics of the external environment are required as input parameter. Semi-empirical or analytical correlations, indeed, may provide an estimation of these data, but they can be properly applied only to a small group of test cases, if compared with the huge amount of possible configurations with different geometries, liquid properties and operating conditions. VOF RANS and LES are performed to provide the internal nozzle flow characteristics, and the subsequent initial jet characteristics. For this reasons, in this work VOF Reynolds Averaged Navier-Stokes (RANS) and Large Eddy Simulation (LES) have been performed to provide a proper simulation of the internal nozzle flow and the subsequent initial jet characteristics. In addition, a sensitivity analysis has been performed to evaluate the effect of various turbulence model (i.e. RNG k-e, Reynolds Stress Model and LES) on the final numerical results. Then, these informations have been applied to reproduce the following jet development and its subsequent break-up. To reach this goal a DNS code from the University of Stuttgart, Free Surface 3D (FS3D), has been adapted and used. As shown, the achieved results could be useful to define the principal phenomena involved in the atomization process. Moreover, comparison with a well known analytical method are presented in order to underline possible drawback and improvements. For both the internal and the external flow numerical simulations, a grid dependence study, as well as the effect of various operating conditions, has been investigated to exclude any important and unwanted dependences.