Learning Forum

[Weiqiang Liu]

Hi all,

I am doing a porous soot combustion project. I reproduced the base case results in literature. The temperature magnitude is just the same with in the paper, though temperature distribution is slightly different. At first, I did not really pay attention to the difference of temperature distribution. I thought it was because we used different fluent solvers.

Then I continued to increase soot consuming rate like the author did. I found temperature would go extremely high at about 70 seconds and divergence happened. I checked my source code especially enthalpy source for several times and found no error. Then I tried to use default value of viscous resistance coefficients for porous media in fluent which is much smaller than the value recommended by the author. Then divergence never happens.

Therefore, I guess divergence in my case is because negative Y velocity namely filtration velocity is too small in my case and further convection is the only heat dissipation mechanism, so heat would accumulates in the corner and yields extremely high temperature. 

I put filtration velocity contour in the paper below:

I put filtration velocity contour of my simulation below:

You can see the Y velocity contour is totally different. There is a very obvious negative Y velocity near the end wall of inlet channel which dissipates heat very quickly. However, for my case, the Y velocity is very small and uniform. I think this is the reason divergence happens when I increase soot consuming rate.

I checked almost every details of my model. I still can not get similar Y velocity contour.

Can anybody give me some suggestions?

Best regards.

Weiqiang

[DrAmine]

Why new topic ?

Is the paper 2d or 3d?
What're the differences?

[DrAmine]

Moreover it is working with default resistances and with the ones it shows no filtration and it diverged later. Have you already investigated if there values are correct?

[Weiqiang Liu]

Hi, Amine,

you're right. I think something wrong with my viscous resistance UDF. In fluent, F_PROFILE is used to store calculated viscous resistance value. I am wondering how fluent deal with the interface between two porous zones like the picture below:

The upper layer is porous soot cake while the lower layer is porous wall. I need to calculate viscous resistance coefficients for these two zones. I am wondering how fluent deal with the interface between them. I mean viscous resistance coefficient is a face type data, right?

Best.

Weiqiang 

[Rob]

In a porous media the resistance is on the cell. In a porous jump the resistance is on the face, but won't effect flow parallel to the face. 

Between the two zones is an interior assuming you have a conformal mesh: that shouldn't do anything to the flow. 

[Weiqiang Liu]

So should I use C_PROFILE to store viscous value or F_PROFILE. Like you said, resistance is on the cell in porous zones.

Best 

Weiqiang 

[DrAmine]

Yes as a C_profile.

[Weiqiang Liu]

Hi Amine,

Thanks very much! I think I have a lot of problems of my case. Now I know that the Y-velocity contour in the paper is definitely right! Because soot was consumed at end of inlet channel at 70 seconds and most gas was forced to flow through this small opening, which yielded very large negative Y-velocity.

Best.

Weiqiang