[Iniesta8]

Hello, everyone
Recently, I found that the heat flux results of a 2D laminar cylinder are mismatched with the fact. As we all know, for a hypersonic flow over a cylinder, the peak heat flux is located at the stagnation point. The true distribution of heat flux is like this But when I use some software or CFD codes to do the simulation, the result is that the peak heat flux is always a bit deviation to the stagnation point. The figure is attached. The grid is 2D and the height of the first layer is 1e-6m.

I don't know why the CFD results are wrong, can anyone give me some advice?
Many thanks!!!

[wkernkamp]

The stagnation point will have the highest temperature if the wall has no heat flux. I suppose you gave the wall a fixed temperature. This means the wall is going to cool the flow near the stagnation point. At the stagnation point, there is very little fluid movement, so the cooling is more effective there relative to adjacent locations.

[Iniesta8]

Originally Posted by wkernkamp

The stagnation point will have the highest temperature if the wall has no heat flux. I suppose you gave the wall a fixed temperature. This means the wall is going to cool the flow near the stagnation point. At the stagnation point, there is very little fluid movement, so the cooling is more effective there relative to adjacent locations.

Thanks for your reply,
Yes, I used an isothermal wall boundary for the cylinder wall. But in theories, the temperature gradient of the stagnation point is the highest and many experimental results prove that the stagnation point has a maximum heat flux. Here you can see a comparison of experimental and CFD results. experiment.JPG

[Iniesta8]

Originally Posted by sellardsh

I think you will be able to find the answer to this question using this link. The article exclusively deals with the Simulation of Conjugate Heat Transfer in Thermal Processes with Open Source CFD

That is not a link to the article!!! And my case is a pure fluid flow without the simulation of heat transfer in solid.

[sbaffini]

What code did you use to get those results? Is it an immersed boundary code? If in house code, what method did you use?

Btw, you are right, higher heat flux should be at stagnation point for your fixed temperature bc. But the problem I see here is slightly different. The max heat flux seems shifted because of those oscillations that just seem to get worst at the stagnation point.

Give more details and we will try to help

[sbaffini]

Also, I see you are postprocessing results in paraview? How were they written to file and processed?

[Iniesta8]

Originally Posted by sbaffini

Also, I see you are postprocessing results in paraview? How were they written to file and processed?

Hi
Thanks for your reply. The code I used is Hy2Foam, which is based on the OpenFoam framework. Now I think the reason for the unphysical result is that my mesh is not dense enough in the flow direction. But after refining the mesh, the peak heat is still located at a tiny distance from the stagnation point.

[sbaffini]

This almost certainly excludes postprocessing issues and also the code is not IB. Have you maybe done some of the tutorials of the code, to see if the problem is there as well?

[Iniesta8]

Originally Posted by sbaffini

This almost certainly excludes postprocessing issues and also the code is not IB. Have you maybe done some of the tutorials of the code, to see if the problem is there as well?

Hi,
I have done some tutorials like a blunt cone. And the result is fine. This question just appears when I doing a 2D blunt body simulation. I also have done the simulation with some commercial software like fluent and CFD++. But it also needs a very dense grid to make the peak heat almost located at the stagnation point.
From my side, this mismatch with the fact can almost avoid via a very dense grid. And it will cost much more time to achieve convergence. I still want to know why it needs a so dense grid when doing 2D blunt body simulations.

[LuckyTran]

More than the peak heat not being at the center, there is the greater issue that the heat flux is not smooth anywhere. I suspect there is some c-grid shenanigans that doesn't go away no matter how you refine the mesh because the same skewed cells are messing you up every time. It doesn't happen in the tutorial because, well, you followed the tutorial instead of doing things on your own. There is a lot more amuck than the peak heat not being in the center, don't bury them under the rug.

[Iniesta8]

Originally Posted by LuckyTran

More than the peak heat not being at the center, there is the greater issue that the heat flux is not smooth anywhere. I suspect there is some c-grid shenanigans that doesn't go away no matter how you refine the mesh because the same skewed cells are messing you up every time. It doesn't happen in the tutorial because, well, you followed the tutorial instead of doing things on your own. There is a lot more amuck than the peak heat not being in the center, don't bury them under the rug.

Hi
Thank you for the advice.
Yes, I know what you mean. But when I doing a sphere simulation, this mismatch will not take place. And the mesh is generated by myself. And the fluctuations always exist less or more because of the skewed cells. Could you tell me how to avoid these "c-grid shenanigans"?