Learning Forum

[Weiqiang Liu]

Hello, all

I have some simulation results of methane catalytic combustion with fluent. However, they don't reproduce literature results very well. I am writing this discussion to ask for some suggestions.

The computation domain is 2D planar with both fluid and solid domain. There is catalyst coated on the solid wall. However the leading edge of the wall has no catalyst (inert) and the thermal boundary condition of the leading edge is set to 300k which is the same temperature with the inlet mixture (I am doing this for the aid of convergence recommended by some authors). The length of the the channel is 20mm. The height of the channel is 1mm.

The gas mechanism I use is GRI 3.0 and the surface mechanism I use is Detuchmann mechanism. Both the two mechanisms are widely used mechanisms. Before I imported the mechanism, I checked the mechanisms in chemkin and found no problem. 

The inlet velocity of the mixture is 0.5m/s and the equivalence ratio is set to 1. However, I don't know the mole fraction of methane because the author did not mention it in the paper. Therefore, I just set the oxidizer to air and make the equivalence ratio to 1. The outlet is pressure outlet. All the walls are set as adiabatic.

The case was initialized by standard initialization and a 2000K temperature was patched to the whole computation domain to ignite the mixture (recommended by literature).

After 500000 times iterations for one week!!!!! , I got the following results. However, the residuals are still oscillating.(I just can not wait anymore)

1. Temperature field.

the maximum temperature is 1090K which is too low compared to literature. I don't know if it is because the mole fraction of methane is lower than those in literature.

2. contours of main species

The problem is the reactants are consumed too slowly namely the products are produced too closed to the rear part of the channel. the contour of OH mass fraction even shows no OH has been produced.

3.methane consuming rate along the catalytic wall.

the consuming rate of methane also shows reaction happens in the rear part of the channel, which does not make sense. In literature, almost all methane is consumed in the front part of channel.

4. the surface coverage of O(s) and PT(s)

the surface coverage of PT(s) is too low along the catalytic wall compared to literature values.

I am wondering is it because the results are still not converged. But somebody told me that 100000 iterations are more than enough for my case while I have 500000!!!!

Or are there something wrong with my model settings?

Can anybody give me some suggestions?

Thanks so so so  so much!!!!!!!!!!!!!!!!!!!!!!!!!!

weiqiang

[Amine Ben Hadj Ali]

Which reaction model are you using? Have you rechecked every single reaction if it is set properly and translated correctly? Do your boundary conditions correspond to the paper you try to reprdouce? Convergence has nothing to do with number of iterations but rather with the time scales limiting the flow and reaction. Have you checked the imbalances under Flux report?

[Weiqiang Liu]

Hi Amine,

my reaction model is species transport with laminar finite rate. I have 340 gas phase reactions and 24 surface reactions. I downloaded them from the website of a research group. In what way can I check the reactions one by one? I just imported these mechanisms in both chemkin and fluent to guarantee they don't complain any error.

To check the imbalances of flux, do you mean to check the total surface heat flux of inlet and outlet . I've already checked the mass flow rate and it's balanced. For a reacting flow, what kind of flux balance do I need to check? I really have no idea.

Best regards,

Weiqiang