CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Main CFD Forum (https://www.cfd-online.com/Forums/main/)
-   -   How to get an accurate CFD solution to aeroheating problem for hypersonic flows? (https://www.cfd-online.com/Forums/main/63545-how-get-accurate-cfd-solution-aeroheating-problem-hypersonic-flows.html)

 Leining D April 12, 2009 10:49

How to get an accurate CFD solution to aeroheating problem for hypersonic flows?

Hi, everyone:
I am glad to disscuss CFD with you.
I am puzzled with a proplem, which is how to get an accurate CFD solution to aeroheating on body with hypersonic velocity. In my work,I found that heating flux on wall obtained by CFD had large errors comparing with experimental data. Do you have any ideas about that?
Thanks!

 harishg April 12, 2009 12:17

If your mach number is > 5 your equations change due to dissociation of molecules. Assuming its less than 5 what numerical scheme do you employ ?

 gory April 12, 2009 20:54

If the grid is unstructured, a current standard FV code has a big trouble in heating prediction even for a very simple geometry like a 2D cylinder, which I learned from this paper:

Computational Aerothermodynamics Simulation Isues on Unstructured Grids, Peter A. Gnoffo and Jeffery A. White, AIAA-2004-2371, June 2004.

 Leining D April 13, 2009 07:37

Quote:
 Originally Posted by harishg (Post 212635) If your mach number is > 5 your equations change due to dissociation of molecules. Assuming its less than 5 what numerical scheme do you employ ?
In my problem,Mach number is equal to 10 and the body is a cone with blunt nose. It is supposed to be a simple shape with simple mesh. Because experimental data is from shock wind tunnel in which experimental time is very short, high temperature effect such as dissociation of molecules should not be present. Therefor, I think that emoloying Navier-Stokes equations in perfect gas as governing equations is reasonable. The scheme I used is Roe's splitting +MUSCL with Minmod Limiter. Do the scheme result in large error?
Thank you!

 Leining D April 13, 2009 07:42

Quote:
 Originally Posted by gory (Post 212654) If the grid is unstructured, a current standard FV code has a big trouble in heating prediction even for a very simple geometry like a 2D cylinder, which I learned from this paper: Computational Aerothermodynamics Simulation Isues on Unstructured Grids, Peter A. Gnoffo and Jeffery A. White, AIAA-2004-2371, June 2004.
My grid is structured. Do you have any suggestion?
Thank you!

 harishg April 14, 2009 11:26

I worked in a group where people did similar simulations for a blunt nose with structured grid and obtained pretty good results. I do not remember the numerical scheme they exactly employed but its either ENO or Godunov scheme with a TVD R-K method. I have seen some simulations in AIAA journal which shows the successful application of ENO schemes for flows with shock. MUSCL might be highly dissipative but if you have experimental results, you can always check the contours to see the ability of the scheme.

 All times are GMT -4. The time now is 17:08.