Burning Velocity Model (BVM) to simulate pure premixed turbulent flame
 

Hi,

I have been using BVM model to simulate pure premixed turbulent METHANE-AIR flames under lean conditions inside a gas turbine combustor. I have some queries with regards to accuracy of BVM in predicting the species profile for pure premixed flames.

1) The flamelet library generated by the RIF tool in cfx (Using GRI-MECH 3.0) maps the species profile for a range of mixture fraction values. However in a 100% premixed flame, mixture fraction has a constant value (zero mixture fraction variance as well) and the reaction zone is modeled by reaction progress variable. How will the cfd solver update the species profile from the flamelet library for pure premixed flames then considering the fact that RIF maps the species concentration for a range of mixture fraction values?

2) Using constant mixture fraction and 0 mixture fraction variance as boundary conditions for pure premixed flames, I simulated the flame using BVM. As far as temperature profile is concerned the agreement between simulation and experiment result is excellent. However the species profile for the intermediate radicals is extremely poor. For example the OH specie contour is very similar to CO2 i.e. minimum in fresh reactant and maximum in the burnt products. However the OH concentration should only exist in the reaction zone (0 in fresh and burnt products). Same goes for other intermediate radicals such as CH2O, O, H2 etc.

3) Now FLUENT 15.0 has a similar model implementation under the name of FGM (Flamelet generated manifold). An additional feature in this model is a separate option for premixed flamelet generation (In CFX RIF there is no such option). I simulated the same problem using FGM in fluent for premixed flamelet and the was able to get right results for the intermediate radicals (such as OH, CH2O, O etc.). From this observation can I deduce that CFX-RIF does not have the capability to simulate correctly the species profile for intermediate radicals, since the flamelet library generated by cfx-RIF is for non-premixed flamelet?

I will be really grateful to the experienced users in the field of turbulent reacting flow for their precious feedback.