[Weiqiang Liu]

hello, all,

I am doing a micro-combustion with fluent. The geometry include gas phase and solid wall. The diameter of the channel is 0.6mm. The thickness of the wall is 0.2mm and the length of the channel is 10mm.

In my case, I imported both gas phase and surface mechanism in chemkin format. The gas phase is n-dodecane mechanism with 54 species and the surface mechanism is Detchem of methane on Pt. The surface reaction rate of gas phase species other than methane are defined by UDF.

The species transport panel is shown in the attachments. I also have 67 surface reactions , in which 24 are imported from surface mechanism file and the rest 43 are defined by user defined function. In addition, my UDF can build, load and hook normally.

however, fluent complains about 'divergence detected in AMG solver: temperature' at the beginning of iteration. So I tried some methods recommend on CFD forum.

First , I tried to decrease under relaxation factor of species and energy equations. it doesn't work.

Then, I tried to turn off energy equation first and got a converged cold flow field and then opened energy equation to continue iteration. Unfortunately, it does not work.

At last, I thought it was the quality of my mesh that caused this error. So I checked the quality of my mesh in ICEM and it is perfect.

I am wondering is it because the gas phase mechanism is too simplified, which makes the reactions very stiff. or is it a defection of Fluent to simulation complicated chemical reaction flow?

I can not upload the case file because it's size. I can send you guys the case file if you want.

Thanks very much!!

Weiqiang

[LuckyTran]

Originally Posted by Weiqiang Liu

or is it a defection of Fluent to simulation complicated chemical reaction flow?

Just no. Don't cook up a defect because your case isn't converging. Fluent solves exactly what you tell it to solve. And it is the same as when you use other software too.


You have plain good-old divergence issues. You just need to put in the time to figure out why it is diverging.

[Weiqiang Liu]

Originally Posted by LuckyTran

Just no. Don't cook up a defect because your case isn't converging. Fluent solves exactly what you tell it to solve. And it is the same as when you use other software too.


You have plain good-old divergence issues. You just need to put in the time to figure out why it is diverging.

OK thanks very much. you give me confidence. I am trying to solve this.

[MahmoudHisham]

Originally Posted by Weiqiang Liu

OK thanks very much. you give me confidence. I am trying to solve this.

Hi, Have you managed to sort it out as I am having the same problem, thanks

[Bin Hao]

Hi buddy, have you solved this problem yet?
I also encounter the problem.