Learning Forum

[khaledhmz12]

Hello,

I simulate the combustion of propane-air flame in a swirl burner using DES turbulence model, I start a steady state simulation with Standard K-e model I got the following flame shape and this shape match the experimental.

but when i switch to DES i have a flame flashback I don't know why I try to refine the mesh but no thing changed... see next picture 

 

IF some ome experienced this problem and can help me in this please give me answer,

thanks in advance

[Essence]

Hello,
The results obtained depends on many factors like the mesh refinement, cell to cell ratio (<1.2), combustion models, initial and boundary conditions, and setting up the case properly.

You might get more information here: 8.2. Premixed Combustion (ansys.com)  17.2. Modeling Premixed Combustion (ansys.com)  8.2.1. Overview (ansys.com)

If you are unable to access the link, follow this Forum discussion https://forum.ansys.com/forums/topic/using-help-with-links/#latest

[khaledhmz12]

Hello,

thankyou for the replay, for the mesh I refined it very well and the total noead number is 803496, the simulation performed by species transport and for the setup I follow the LES  guide to setup it. the flame flashback is occurs only at high swirl number Sn=1.05 or when consider the rotation term in the realizable k-epsilon in the steady state, cartesian coordinate.

 

 

 

[Essence]

Did it solve your problem? What is the cell to cell ratio you used? You mentioned you are using species transport. Which combustion model are you using?

[khaledhmz12]

no not yet, cell to cell ratio you mean max aspect ratio? the combustion model is EDC with stiff chimistry

[Essence]

Cell to cell ratio meaning, growth rate in the mesh. Are you using CHEMKIN mechanism (encrypted or non-encrypted)? And as you are using DES which is transient, try using lower Under-relaxation factors initially until the residuals are stable and afterwards you may increase the URFs.

Try to converge the steady state first in the reaction and then using the converged solution, proceed with the DES. Use suitable time step and the number of iterations in each time step. Maintain the Courant number.

[khaledhmz12]

growth factor is 1.2, yes I use chemkin mechanism but I don't know if it encrypted or no, is there a  method to check it?, the Under-relaxation factors that I use is: 0,3 pressure, 0,6 density, 1 body forces, 0,55 momentum, 0,6 swirl,  0,8 Tur kinetic enrg, 0,8 tur dissipation rate, 1 turb viscosity, 0.95 species, and 0,95 energy is there any suggestion to play in this factor?

courant number i set up C <0.8 and the time step is 4e-6    

[khaledhmz12]

the flashback occur only when I use large mechanism reaction 20 species and 66 reactions, with 3 step there is no flashback and there is no problem in the simulation 

[Essence]

To check if it is encrypted or not, go to the Materials tab and check if the reactions rates, transport and thermodynamic dat are visible. If you are using encrypted mechanism, then you will only see the species but not the others. If you are using non-encrypted, then you will see the species as well as the others. This is to protect the IP. If you are using the default URFs, reduce them a little until the residuals are stable, then increase them.

[Essence]

If you are using encrypted mechanism then you should be using CHEMKIN CFD solver instead of Stiff Chemistry.

[khaledhmz12]

 

I check it and I think it is non-encrypted as you see in the example picture.

for URFs I reduce it as you see above

 

[Essence]

I see that you are using single step mechanism. CHEMKIN mechanism has many species and the reactions. Also, the fuel I see on the display is not propane.

[khaledhmz12]

nnn this is just an example but in my fluent panel it appears like this one