VALIDATION OF A STOCHASTIC BREAKUP MODEL FOR TURBULENT JETS IN HIGH-SPEED CROSSFLOW: ASSESSMENT OF TURBULENT INTERACTIONS AND SENSITIVITY TO BOUNDARY CONDITIONS

Abstract

Improving the mixing of fuel and air by injecting a turbulent liquid fuel jet into a high-speed cross-flowing gas can reduce the emissions of gas turbine applications. To facilitate and hasten the development of such low-emissions technologies, accurate predictions of the spray characteristics are needed. The objective of the present study is to validate the predictive capabilities of a stochastic breakup model for turbulent transverse jets over a wide range of representative pressures and atomization characteristics. The effect of turbulence modeling is also assessed to provide accurate and computationally less expensive Eulerian-Lagrangian transient approaches. To do so, the predictions made with the large eddy simulation (LES) approach for different subgrid-scale (SGS) models and with the synthetic eddy method (SEM) are compared to the ones made using the unsteady Reynolds-averaged Navier-Stokes (RANS) approach with and without a turbulent dispersion model. The sensitivity of the numerical methodology to the upstream velocity profile, pressure, and momentum flux ratio were also assessed. Properly accounting for the upstream gas velocity profile was found to be critical to ensure accurate predictions of the spray characteristics. The unsteady RANS (URANS) turbulent approach coupled with the turbulent dispersion model showed good agreement with experimental data, but the LES approach tends to overpredict the spray penetration and underpredict the Sauter mean diameter (SMD). This could be due to the lower turbulent interactions it predicts, which may lead to lower momentum transfer between the phases.