Author : Celestin P. Zemtsop
Publisher :
Page : 115 pages
File Size : 44,56 MB
Release : 2011
Category : Eddies
ISBN : 9781303050381
Book Description
Swirl flows are useful in many engineering and environmental applications. They can be utilized to reduce emissions of pollutants, to improve ignition stability, and to stabilize a flame for clean combustion. The majority of swirl flows include both wall-bounded regions and free shear flow regions. Experimental research cannot provide all the information needed for the study of these flows. Numerical simulations of swirl flows provide more insights into the structure of the swirl flows. The direct numerical simulation (DNS) method is an accurate and reliable tool for flow simulations but computationally expensive. Reynolds-averaged Navier-Stokes (RANS) methods are computationally less expensive for the simulation of wall-bounded and free shear flows, but they turn out to be not well appropriate for turbulent swirling jet simulations. Large-eddy simulation (LES) is an appropriate modeling approach for turbulent swirl flow simulations, but the calculation of the free swirling jet using LES needs fluctuating inlet velocity conditions. The generation of such fluctuating velocity fields could be done by performing a LES of the nozzle flow. Unfortunately, the computational costs of such LES are comparable to those of DNS. A need thus arises to apply a hybrid RANS/LES method that combines two main advantages, the capability of LES methods in capturing instantaneous flow structures and the low computational costs of RANS methods. In this study, the performance of two hybrid approaches, a segregated RANS/LES method, and a unified RANS/LES method, is investigated regarding the simulation of turbulent swirling jet flows. The segregated model combines a RANS model in the nozzle region and a LES model in the jet region, whereas the unified model combines a RANS model in near-wall regions with a LES model away from the wall. This study provides evidence that the unified turbulence model is a better tool to predict swirl flows and, more importantly, a better tool to calculate the vortex breakdown. The accuracy of the numerical predictions is confirmed by comparing available experimental data for non-swirling and swirling jet flows with computational results. The validated model is used to study the mechanism of swirl effects, vortex breakdown, and scalar mixing.