[oscar1404]

Hello,

I am trying to simulate a lean premixed flame H2+air (phi=0.45) at high speed (Bulb velocity 15m/s in bunsen diameter 8 mm) combustion but I am facing a lot of troubles. I changed mechanism and tried with the different models on FLUENT. Does someone know any trick or advise to this study case?

Thanks

[Starcatcher]

what exactly is the problem? does the case not ignite, or what?

[oscar1404]

It goes in blow-off, but experimental data confirms that the flame is stationary

[Starcatcher]

ok, tell me your fluent settings.
for example, are you using finite rate chemistry? is there a paper about this chem. mechanism?
are you applying TCI/EDC, or simmilat. The EDC model in fluent is reducing reactoin rate too strongly, sometimes. This could lead to flame extinguishment. Also what turbulence model are you using?
maybe you can add a screenshot of the geometry.

[oscar1404]

Thank you for the response.
After some research I could correct some issues in my simulation. I was using the premixed model (Zimont and Peters) but now I understand it is not applicable. I am using now the EDC and the mechanism available in Ansys Fluent, with thermal diffusion and full multicomponent difussion. For now it seems to work better. But, the problem that I am facing is with the residuals, they are fluctuating in values above 10e-3, I would like to reduce them. I am using a QUICK scheme, and the chemistry equilibrium is solved for each iteration. Here the geometry, I am using a automatic refinement of the mesh (cell volumes of 3e-11)

[Starcatcher]

try to use simple scheme. alternatively complex with a very high cfl number.
are you calculating stationary or transient. stationary should help. Also, you can use discretisation of first order for a first conversion.

[Starcatcher]

you need to use a chem mechanism which was developed for the given gases, mixture ratio and pressures and reactant temperatures. This you should see from the publication, where this mechanism was presented. So what mechanism are you using? or did you just define a one-step reaction by yourself?

[Starcatcher]

I think the best thing would be if you would just publish screen shots of your setup. I could go through all settings

[oscar1404]

I am using the available mechanism from Ansys. But I am considering that due the geometry (free jet, with three pressure outlets), the residuals of 10e-3 are acceptable. What do you think?

But now, I am not sure about the temperature. It is arriving until 1470 °C which is almost 100 degrees higher than the adiabatic temperature for this equivalence ratio. What do you think are the possible reasons for the higher temperature?

[oscar1404]

the simulation is also having the "dasac failure at Temperature= 303.15 K: Continuing..." and it does not change along the simulation. I tried to use "chemical agglomeration", but it does not work. I am trying to reduce the error for integration, but the problem but it does not seem to work.

[Starcatcher]

Hey sorry for the late reply,

I am not sure if ANSYS has a proper mechanism. can you provide details on it? In your case a mechanism should have around 8 species. You should not use one from ansys but look into the literature and contact the authors of a mechanism. Attached i send a mechanism for h2-o2. I don't have one for h2-air but it gives you an idea how a mechanism shall look like. You can always use the GRI 3.0 but it's too big. You can try to find sombody on google scholar who reduced this mechanism for h2-air applications.

in the integration parameters in ansys use "direct integration" without ISAT. There, set the absolute error tollerance to1e-10 and the relative error tollerance to 1e-05.

[oscar1404]

hello, thanks for the response,
The mechanism available in Ansys is extracted from M. J. Lighthill. "A Comprehensive Kinetic Mechanism for CO, CH2O, and CH3OH Combustion". Int. J. Chem. Kinet. 39. 109-136. 2007. Indeed, it is not a model intended for H2 originally. I change the integration setup and I disable the ISAT and set the errors as you indicate, but now it goes in flashback. I am trying with other mechanisms (Konnov and Connaire) but when I use them the flame goes in blow off.

Before disabling the ISAT option, I got an solution (no desac integration errors) but there is an strange zone before the flame front. This zone has a rapid change in T and H2 fraction, it seems no physical. Do you know how to solve this error?

If you know a better mechanism or some tip to avoid the blow off on the simulation I will appreciate,

Thanks

[Starcatcher]

is this the mechism you uses?
https://www.sciencedirect.com/scienc...10218019300501

can you add a colorbar to your piture? what are the maximum temp?

preferably you need a mechanism with n2 in it. Otherwise you can use this konnov mechanism and keep N2 inert. This should do the trick for a rough solution.

about ISAT, I am a little surprised that it helps so much for stability. usually it's purpose is just for acceleration in some cases.

In order to identify the sourse of the instability, I really suggest you to put screen schots of all your settings. This means at least
- a screen shot of turbulence settings (the menu is called "viscous" in fluent"
- a screenshot of rective settings (including edc parameters and chemical solver parameters)
- a screen shot of discretisation scheme settings (i think it is called "control" and "solution" in fluent)
- a screenshot of the calculation settings (called "calculation" in fluent)