Question about rhoCentralFoam and rhoPimpleFoam?

iLyrics · June 30, 2016, 11:36

Hello foamers. I'm now making an effort to generating turbulence inflow in for a flat in the supersonic flow. At the very first, I was doing the test case by the rhoPimpleFoam. The results were acceptable. However, when I turned to the rhoCentralFoam, I cann't reproduce the same results. The puzzle troubled me a lot. I just can't find out where is going wrong. In fact, I just modified the boundary condition of the inlet plane, namely U, T and rho field. Any suggestions?

iLyrics · June 30, 2016, 11:38

How to attatch the picture?

iLyrics · September 4, 2016, 11:28

Let's answer myself. As to my experience, the rhoCentralFoam solver is adopting a dissipative scheme(TVD/NVD), while the the rhoPimpleFoam is adoping a non-dissipative scheme(linear scheme). So you can use the rhoCentralFoam solver to deal with the discountinuity problem, but you can't expect it to generate turbulence. The rhoPimpleFoam solver is just the opposite side.

g.freeman · September 7, 2016, 05:10

Quote:

Originally Posted by iLyrics

Let's answer myself. As to my experience, the rhoCentralFoam solver is adopting a dissipative scheme(TVD/NVD), while the the rhoPimpleFoam is adoping a non-dissipative scheme(linear scheme). So you can use the rhoCentralFoam solver to deal with the discountinuity problem, but you can't expect it to generate turbulence. The rhoPimpleFoam solver is just the opposite side.

That's the answer

usv001 · December 10, 2016, 03:24

Hello all,

I am faced with the same dilemma as iLyrics. I have implemented two different synthetic turbulence generation methods. I tried to test them using rhoCentralFoam and found that the turbulence dissipated very fast but quite well preserved in rhoPimpleFoam (see picture).

I find the answer given by iLyrics quite reasonable but I need to use rhoCentralFoam since I want to test cases with shocks. So, is there anyway out?

Regards,
USV

iLyrics · December 10, 2016, 04:28

The key point is that you should adopt a low dissipation scheme, which could deal with shock wave and resolve the turbulence deep into the wall boundary layer.
Most importantly, the grid should meet some requirements, especially combined with LES methods. So realization of turbulence inflow generation within the OpenFOAM frame is not an easy case.

usv001 · December 10, 2016, 07:57

Hello Lee,

Thanks for the quick reply. So, do you think that if I were to use higher order schemes with proper shock capturing (e.g. ENO, WENO), this problem shouldn't arise?

The mesh I am using now is close to DNS requirements, actually. So, I believe mesh shouldn't be a problem.

Regards,
USV

June 30, 2016, 11:36	Question about rhoCentralFoam and rhoPimpleFoam?	#1
iLyrics New Member Lee Yachao Join Date: Dec 2015 Location: Beijing,China Posts: 12 Rep Power: 10	Hello foamers. I'm now making an effort to generating turbulence inflow in for a flat in the supersonic flow. At the very first, I was doing the test case by the rhoPimpleFoam. The results were acceptable. However, when I turned to the rhoCentralFoam, I cann't reproduce the same results. The puzzle troubled me a lot. I just can't find out where is going wrong. In fact, I just modified the boundary condition of the inlet plane, namely U, T and rho field. Any suggestions?

September 4, 2016, 11:28		#3
iLyrics New Member Lee Yachao Join Date: Dec 2015 Location: Beijing,China Posts: 12 Rep Power: 10	Let's answer myself. As to my experience, the rhoCentralFoam solver is adopting a dissipative scheme(TVD/NVD), while the the rhoPimpleFoam is adoping a non-dissipative scheme(linear scheme). So you can use the rhoCentralFoam solver to deal with the discountinuity problem, but you can't expect it to generate turbulence. The rhoPimpleFoam solver is just the opposite side. aerosjc, olegrog, febriyan91 and 6 others like this.

December 10, 2016, 04:28	Suggestions	#6
iLyrics New Member Lee Yachao Join Date: Dec 2015 Location: Beijing,China Posts: 12 Rep Power: 10	The key point is that you should adopt a low dissipation scheme, which could deal with shock wave and resolve the turbulence deep into the wall boundary layer. Most importantly, the grid should meet some requirements, especially combined with LES methods. So realization of turbulence inflow generation within the OpenFOAM frame is not an easy case. febriyan91, hogsonik, guanjiang.chen and 1 others like this. Last edited by iLyrics; December 10, 2016 at 04:31. Reason: wrong words

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
how to make rhoCentralFoam to write continuity residuals?	immortality	OpenFOAM Running, Solving & CFD	6	April 18, 2018 03:56
rhoCentralFoam flat plate boundary layer issues	laurensvd	OpenFOAM Running, Solving & CFD	6	September 13, 2013 03:10
momentum equation in rhoCentralFoam	immortality	OpenFOAM Running, Solving & CFD	0	June 12, 2013 06:22
how to modify rhoCentralFoam to write continuity residuals?	immortality	OpenFOAM Programming & Development	0	May 1, 2013 12:44
is sonicFoam better in convergence than rhoPimplefoam or rhoCentralFoam?	immortality	OpenFOAM Running, Solving & CFD	1	February 11, 2013 22:26

June 30, 2016, 11:38		#2
iLyrics New Member Lee Yachao Join Date: Dec 2015 Location: Beijing,China Posts: 12 Rep Power: 10	How to attatch the picture?

December 10, 2016, 07:57		#7
usv001 Senior Member Join Date: Sep 2015 Location: Singapore Posts: 102 Rep Power: 10	Hello Lee, Thanks for the quick reply. So, do you think that if I were to use higher order schemes with proper shock capturing (e.g. ENO, WENO), this problem shouldn't arise? The mesh I am using now is close to DNS requirements, actually. So, I believe mesh shouldn't be a problem. Regards, USV