Turbulence Modelling and Flow Analysis of a Trapped Vortex Combustor

Abstract

In this paper, a computational study is presented on the turbulence structure and aerodynamics of a Trapped Vortex Combustor (TVC) by means of Reynolds–Averaged Navier–Stokes (RANS) calculations. The aim is to assess the ability of three turbulence models — RSM, Realizable k—ε and SST—kω — in predicting flow characteristics and pressure drop within the combustor. Fluent of ANSYS was used to simulate the scratching process under different Reynolds numbers and cavity geometries, and simulation results were compared with data from literature. A grid insensitivity analysis demonstrated 30,000 cells to be the optimum mesh refinement. It is also found from the results that RSM model provides the best predictions of pressure loss and gives a maximum discrepancy slightly above 12.1% compared SST—kω (23.9%) and Realizable k—ε (26.7%). Parametric studies showed that for Reynolds numbers up to 300% the recirculation zone would strengthen factor of three and turbulence intensity by an order of magnitude, respectively, while flow imprints did not show a considerable shift. The authors suggest an optimum H/Df ratio of the depth—to—flow diameter is found to be around 0.6, where a pressure drop reduction (to less than one—third) due to vortex generation inside and downstream of the cavity is minimum. Both small (H/Df 1) cavities generate multiple vortices, resulting in higher aerodynamic losses. The novelty is to compare RANS models for TVC optimization and determine Reynolds number and cavity geometry effects. These results provide useful information for the design and optimization of TVC—based propulsion systems, highlighting the significance of turbulence model and combustor geometry in favoring combustion efficiency and stability.