[twinturbotom]

All,
I would appreciate any help or direction you can offer.

I'm using reactingFoam to simulate the mixing of two gases. I hope to show the lighter CH4 collect at the top of the domain. Unfortunately the CH4 does not seem to disperse as expected.

I have 2 inlets and 1 outlet. one inlet is air (O2 & N2 located at the bottom) the other is CH4 located on the side.

Attached are my case files and a few of the relevant images showing U and CH4 fields.

My online research leads me to believe that I can use reactingFoam with/without combustion and reactions; however, I have had no success after numerous modifications and recommendations from other threads.


I modified the counterFlowFlame2D tutorial in the following way:
-Changed the geometry and used topoSet/refineMesh (same results across various meshes).

-Input files (U,T), and the turbulenceProperties to RASModel vs. the default laminar.

-I have toggled combustion and reactions on/off with no real difference.

Where am I going wrong here? I'm using OF 2.3.x.

Thank you for your time and consideration here!

[twinturbotom]

All, I found reactingFoam to be sensitive to mesh. With a few adjustments I was able to achieve great results. Thanks!

[flowAlways]

Hello Tommy,
Thank you for this info.
yes indeed the reactingFoam results are sensitive to mesh. I have not done mesh sensitivity analysis myself but have read the same at many sources(atleast this one which i can recall Development and validation of a local time stepping-based PaSR solver for combustion and radiation modeling , although it is LTSreactingFoam, but the fundamental ideas are same). I have been using reactingFoam for sometime now and have also done validation studies for Sandia FlameD(benchmark problem). The results are good but do not agree very well with literarture. I am not sure if that is because of the basic turbulence(k epsilon) and chemistry(PaSR, GRI 3.0) models that I am using or due to the sensitivity of the solver to mesh. I would post my results here when I have done sensitivity tests using the benchmark Sandia FlameD. However, since you have disabled chemistry and combustion, I fail to understand the reasons for the sensitivity to mesh. Can you please provide more details ...what are the few adjustments you did? ... what is the no of control volumes for the corase mesh(that gives you wrong results) and the fine mesh that you used. This would give us more information.