A detailed comparison of two sub-grid scale combustion models via large eddy simulation of the PRECCINSTA gas turbine model combustor

作者：Wang, Ping^*; Froehlich, Jochen; Maas, Ulrich; He, Zhi-xia; Wang, Cai-jun

来源：Combustion and Flame, 2016, 164: 329-345.

DOI：10.1016/j.combustflame.2015.11.031

• Large eddy simulation
• Turbulent premixed swirling flame
• Thickened-flame model
• Flame surface density model
• PRECCINSTA combustor
• Flame surface area

A detailed comparison of the performances of a dynamically thickened-flame (DTF) model and a flame surface density (FSD) model is conducted through large eddy simulation (LES) of two fuel-lean turbulent premixed swirling flames in the well-known PRECCINSTA gas turbine model combustor. The two models are implemented in the same in-house Finite-Volume code, with the same numerical method, and all the computations are performed on the same block-structured body-fitted curvilinear grid. Very good agreement between the LES and experimental data is obtained. It shows that for predicting the statistics of velocity, temperature and major species, both the FSD model and the DTF model predict very similar results for the flames under consideration. However, for predicting the distributions of minor species CO the FSD model performs less good as the DTF model, which is analysed to be result from the strategy used to generate the chemistry look up table for the FSD simulation. The coherent flow structures, the flame surface area as well as the power spectra density of velocities are also analysed and compared. The present analysis of the resolved flame surface area might be the first one in the literature of LES of turbulent premixed flame. With the employed model constants, the FSD model yields an approximately 1.4 times as large flame front surface area as the DTF model does. It is also interesting to find that the resolved flame surface area by FSD model is approximately 44% of the real flame surface area for the case phi = 0.75, while for the DTF model the value is around 31%. Additionally, a cross-comparison with results in the literature is performed as well and the present LES results are as good as those obtained in the literature.

• 出版日期2016-2
• 单位江苏大学