Hello again!

I am trying to model combustion within a cylindre. The problem is that ignition occurs near the walls due to the k/e ratio. Wall functions don't help. My next approach is to write a UDF, but I would have to implement the whole EBU-combustion-model again (from my point of view). Is there anybody who has experienced the same problems and has worked out possible solutions? Please contact me!

Thank you!

Regards,

Peter

Appendix:

At www.rz.unibw-muenchen.de/~j9lp0407/ one can see 2 tiffs of my problem. Ignition shall only take place in the area on the lower left. There I patched with a higher temperature and higher product mass fractions. The problem is that the eddy dissipation model does not work correct in the laminar sub-layer near the wall. Instead, there is a high reaction rate.

Regards,

Peter

Peter,

Are you using the finite-rate/eddy-dissipation model or standard eddy-dissipation model ? Also, could you give more information about the problem you are trying to model ?

Dave

OK. I am modelling premixed combustion (2d) of n-octane/air in a cylindre. Ignition shall take place in the lower left corner. I initialised with T=786K, p=922400 Pa. The material settings are piecewise polynomial for the cps, except n-octane (no data in the database), sutherland-theory for viscosity. The I marked the ignition area and patched with T=2600K and mass fractions representing fully burned mixture. Due to the k/e-Viscosity-model ignition takes place near the wall after about 10exp-4 seconds. I have also used a boundary-layer like grid and the k-w-model for viscosity, but achieved the same results. My idea now is to write a UDF that avarages the turbulent timescale. I think I would have to re-program the whole combustion model for that purpose, so I asked whether someone has already written a UDF modifying the combustion model. Other hints for solving this weired problem are also welcomed. I hope, this describes my problem. If you need more detailed information, please write again.

Regards,

Peter