Estimation of the effect of radiation heat transfer in a gas turbine combustor wall temperature with computational fluid dynamics – conjugate heat transfer methodology

Ignition characteris In gas turbine combustion chambers, combustion gases (carbondioxide, water vapor etc.) which have very high temperature occurs as a result of combustion reactions. Therefore, huge amount of heat transfers from hot gases through the combustor wall. In order to determine the lifetime of the liner wall and reliability of the wall cooling technology, it is important to predict wall temperature accurately. The liner is heated by convection and radiation of combustion gases inside the flame tube and it is cooled by convection to the cold air in annulus and radiation to the outer casing. Radiative heat transfer occurs in two ways. These are non-luminous and luminous radiation which are due to complete combustion products (CO2 and H2O), and formation of soot particles in the fuel rich regions of the primary zone of a non-premixed diffusion flame. It is neccessary to take into account of radiation heat transfer for more accurate prediction of liner temperature. It is known that taking into account of radiation heat transfer reduces the peak liner temperature. Computational Fluid Dynamics (CFD) analyses became a popular tool with developing Conjugate Heat Transfer (CHT) solver because of the cost and difficulty of experiment under high temperature and high pressure environment. CHT methodology provides to model fluid volume (flame tube and annulus) and solid regions (liner) simulatenously. In this study, radiation heat transfer from hot combustion products (non-luminous radiation) is modelled and its effect on the combustor liner temperature is observed. Combustor of small scale turbojet engine is simulated with CFD-CHT computations and liner wall temperature is calculated. According to results, radiation heat transfer tends to reduce the wall temperatures, peak liner temperatures of 1310 K and 1140 K were predicted on the wall both for cases without and with radiation heat transfer respectively. Results are also compared with experimental data and it is clearly seen that neglecting radiation gives high unrealistic liner temperatures. Simulating the radiation heat transfer in combustor provides better agreement with experimental data. Simulations are performed with three different computational grids which have 900.000, 3.400.000 and 9.700.000 cells named coarse, medium and fine grids respectively. It is observed that medium and fine mesh results are very close to each other and it is decided that medium mesh is adequate for further simulations instead of fine mesh which has 9.7 million cells. Different combustion and turbulence models are implemented and all results are compared with experimental data with using medium mesh. Relizable k-ε turbulence model and Hybrid Eddy Break-up combustion model results are in better agreement with measured data. Parametric studies are performed with different casing and liner emmisivities, different Turbulent Prandtl/Schmidt numbers and all results are compared with measurements. Reynolds-Averaged Navier-Stokes (RANS) and Large Eddy Simulation (LES) computations are performed in order to see variation of the temperature field inside the flame tube and thereby temperature distribution in the combustor liner. The results of the two methodologies are also compared with experimental data.