Sammanfattning
In this study we compare two types of Large Eddy Simulation (LES) turbulent combustion
models with experimental data for a low swirl stabilized turbulent lean premixed flame.
Such flames are a great challenge to numerical simulations since they are unsteady and sensitive
to boundary conditions, and details of the experimental set-up. The two classes of LES
turbulent combustion models considered are the flamelet and finite rate chemistry models.
Individual models of each category may be very different, but in the former the flame is
considered infinitely thin, whereas in the latter the chemical kinetics and the diffusion governs
the flame behavior. As representative of the flamelet models we here use a G-equation
model, and as representative of the finite rate chemistry models we use the thickened flame
model and the partially stirred reactor model. Predictions are being compared with measurement
data for an atmospheric low-swirl methane/air flame. The experimental measurement
data include data from stereoscopic PIV, filtered Rayleigh scattering and acetone LIF,
providing information about the velocity, temperature and fuel distribution. All LES show
reasonable agreement with the experimental data, predicting a lifted weakly swirling, flame
oscillating back and forth just above the rim of the burner. A more detailed comparison of
the predictions with the experimental data show that best quantitative agreement is obtained
by one of the finite rate chemistry models, whereas the best qualitative comparison is
obtained by the flamelet model. Causes for the difference in qualitative and quantitative behavior
are elaborated on in the concluding remarks section.
models with experimental data for a low swirl stabilized turbulent lean premixed flame.
Such flames are a great challenge to numerical simulations since they are unsteady and sensitive
to boundary conditions, and details of the experimental set-up. The two classes of LES
turbulent combustion models considered are the flamelet and finite rate chemistry models.
Individual models of each category may be very different, but in the former the flame is
considered infinitely thin, whereas in the latter the chemical kinetics and the diffusion governs
the flame behavior. As representative of the flamelet models we here use a G-equation
model, and as representative of the finite rate chemistry models we use the thickened flame
model and the partially stirred reactor model. Predictions are being compared with measurement
data for an atmospheric low-swirl methane/air flame. The experimental measurement
data include data from stereoscopic PIV, filtered Rayleigh scattering and acetone LIF,
providing information about the velocity, temperature and fuel distribution. All LES show
reasonable agreement with the experimental data, predicting a lifted weakly swirling, flame
oscillating back and forth just above the rim of the burner. A more detailed comparison of
the predictions with the experimental data show that best quantitative agreement is obtained
by one of the finite rate chemistry models, whereas the best qualitative comparison is
obtained by the flamelet model. Causes for the difference in qualitative and quantitative behavior
are elaborated on in the concluding remarks section.
Originalspråk | engelska |
---|---|
Titel på värdpublikation | AIAA 2008-513 |
Förlag | American Institute of Aeronautics and Astronautics |
Antal sidor | 14 |
DOI | |
Status | Published - 2008 |
Evenemang | 46th AIAA Aerospace Sciences Meeting and Exhibit - Reno, Nevada, USA Varaktighet: 2008 jan. 7 → 2008 jan. 10 |
Konferens
Konferens | 46th AIAA Aerospace Sciences Meeting and Exhibit |
---|---|
Land/Territorium | USA |
Ort | Reno, Nevada |
Period | 2008/01/07 → 2008/01/10 |
Ämnesklassifikation (UKÄ)
- Atom- och molekylfysik och optik
- Strömningsmekanik och akustik