hi all, I am trying to simulate methane flow and combustion in big furnace. The problem is that I cannot find a definition for best cell size in the closure of fuel inlets (where combustion occures). I know that there are many factors that may has impact on this but can anyone give just a rought advices hot to set up a best mesh resolution for conditions as follows: - Re ~10^6, - Size of domain ~10m, diameter of fuel ports ~0.5 m, - gas velocity ~20m/s - eddy dissipation model.

Are there any forumlas for calculation optimal mesh size for combustion (like for example CFL condition for flow in explicite solvers) or something?. Can enyone give any guideline? In ansys tutorials there is a combustor with elements of size ~4mm in the closure of fuel inlets and there is a comment that there is "not very fine mesh". In my situation, where combution will take place in the volume of many cubic meteres (neglecting rest of the furnace where there is only burned gas flow and mesh must not be that fine) it is not possible to give such small size of mesh elements.

I would be very gratefull for any help, Luk

Hi Luk, I never did such a case nor heard detailed recommendations for it. So if I would have this task, I would try to the classic best practise study. As it is too expensive for the real case, you have to construct a quasi two-dimensional case. This case should be as similar to the real one as possible. The required mesh resolution may depend also from flame type you model (premixed or non-premixed). It can also differ with turbulence/flame interaction. The CFL number is based on theoretic considerations of convective flows. This timescale may differ completly from the timescales of reaction. However pure EDM should be very robust. As CFX is implicit the thing should work with high CFL numbers.

May be somebody has experience with similar case and can give detailed directly applicable tipps.

Greets, Joey

Hi Joey, Thanks for Your reply. Of course the CFL can be only an example of the ciriteria where You can calculate the mesh resolution for simplest solver. I agree that having implicit solver the mesh for combustion (which has timescale many times faster than flow itself) can be completly different. I only wanted to know if there is a best practite for rough assesment of required mesh.

Greetings, Luk

Hi,

The required mesh and timestep is strongly problem dependant so with an implicit solver you have to do a sensitivity study.

Glenn Horrocks

Hi, I have found that (about the simple combustion example from manual) that knowing combustion timescale (for methane combustion it is rougt t=1.37*10^-4s) and gas velocity in the example (V=40m/s) the mesh seems to fit to rough criteria x=V*t and is within 4 up to 5 mm.

Luk