The scale-adaptive simulation of non-reacting and reacting jet in crossflow: Non-uniform fuel injection to decrease wall temperature

Abstract

• SAS is employed to non-reacting and reacting jet-in-crossflow configurations. • SAS improves the prediction of Reynolds stress compared with RANS approaches. • SAS coupled with FRED model predicts velocity field well for reacting jet-in-crossflow. • Arrangement of non-uniform fuel injection is proposed to decrease wall temperature. Jet in crossflow (JIC) is widely used in many industrial applications, but the lack of accuracy on JIC flow prediction with Reynolds-averaged Navier Stokes (RANS) approaches has been recognized through some numerical studies. Thus new numerical method with both low computational cost and reasonable accuracy should be applied to JIC simulation, and the combustion systems also require reliable reacting JIC prediction specially. In the present study, the Scale-Adaptive Simulation (SAS) is employed to both non-reacting and reacting JIC configurations. Significant improvement has been observed on the prediction of Reynolds stress component with SAS as compared with RANS approaches, therefore the scalar field is also predicted accurately by SAS modeling. For the reacting JIC flow, the SAS model coupled with the finite-rate/eddy-dissipation (FRED) model is proved to predict the velocity field fairly well, and the flame-vertex interaction can be well captured by this method. The attachment of the heat release regions tends to increase the wall temperature near the nozzle. Subsequently an arrangement of non-uniform fuel injection is proposed and numerically studied, showing that this arrangement increases the flame lift-off and thus decreases the wall temperature effectively.