Numerical Investigation of a Double-Swirled Gas Turbine Model Combustor Using a RANS Approach with Different Turbulence–Chemistry Interaction Models

Abstract

In this work, numerical investigation of a gas turbine model combustor (GTMC) was carried out using two different turbulence–chemistry interaction models: the EDC (eddy dissipation concept) and TPDF (transported probability density function). GTMC with good optical access for laser measurements provides a useful database for swirling CH4/Air diffusion flames at atmospheric pressure. Modeling was performed by solving Reynolds-averaged Navier–Stokes (RANS) and Reynolds stress model (RSM) equations for a two-dimensional (2D) axisymmetric computational domain accompanied by swirl and the combustion chamber was investigated for both reacting and nonreacting conditions. A detailed reduced mechanism of DRM22 (with 22 species and 104 reactions) was used to represent the chemical reactions. Comprehensive comparisons were done for the predictions and measurements of velocity, mixture fraction, temperature, and chemical species concentrations of H2, O2, OH, H2O, CH4, CO, and CO2. Results showed an acceptable accuracy of predictions by considering computational cost. That means that the simplified 2D-axismmetric-swirl simulation has the ability to capture some important features and structure of combustion field in a double highly swirled chamber, like GTMC, with much lower CPU time in comparison with costly 3D modelings, although misses some details of flow field characteristics in comparison with more accurate large eddy simulation (LES) approachs. In terms of comparison between the turbulence–chemistry interaction models, TPDF led to a good prediction for major species and flame structure near the inlets while the EDC predicted more accurately downstream of the flow field.