Boco interpolation problem

sufjanst · August 15, 2017, 14:08

Hi,

I try to simulate a combustion problem with profiles as inlets. I use a Hexa-Mesh, the profile is from a Tetra-Mesh. I wrote the profile and loaded it into the new mesh. The coordinates are the same. The values in the profile look good. But the interpolation won't work. A few cells have very high velocities and I can't figure out why. The cells have about 20 m/s, but a few cells have over 10.000 m/s. That makes the flow extremely unstable.
Does anybody have an idea?

CeesH · August 16, 2017, 03:33

Hi,

Interpolation always comes with some risk in wall regions or regions with poor cells, where gradients may be very steep. You will anyway have to run a couple of iterations to reconverge your flow; perhaps the simplest option is to patch the poor zones. If you can mark the cells with the high velocities (don't have FLUENT open now, so can't give you the exact names of what to click on) and then use patch (under initialization) to set the velocities in these cells to some more realistic value, that may stabilize your solution.

Good luck!

sufjanst · August 16, 2017, 18:05

I created an o-grid for the inlet region with a very good quality (over 0.8). But it didn't work... It seems like the interpolation just won't work.

https://picload.org/view/rwlowdcl/profile.png.html

I uploaded a screenshot of the inlet. This is the velocity magnitude from 0-50 m/s. It looks just like the profile from the old simulation. But there are a few cells with very high velocities (transparent regions). Do you have any idea?

CeesH · August 17, 2017, 05:05

Did you use an inflation layer/very small cells near the boundary? I can imagine the gradients there are very steep, which may cause (presumably?) linear interpolation to cause huge overshoots. Anyway, the most pragmatic solution seems to me to use some form of clipping/patching to remove excess velocities, as stated before.

sufjanst · August 18, 2017, 09:26

I already tried patching the inlet, but after convergence it got to the old state with the high velocities. I also tried to write the profile from a plane 5mm above, but that didn't work out as well. Is there a possibility to block the high velocities with an UDF?
Do I need an inflation layer? I use a Hexa-Mesh. The elements are parallel anyway...
Thank you for your replies.

August 15, 2017, 14:08	Boco interpolation problem	#1
sufjanst Member Join Date: Mar 2016 Posts: 73 Rep Power: 10	Hi, I try to simulate a combustion problem with profiles as inlets. I use a Hexa-Mesh, the profile is from a Tetra-Mesh. I wrote the profile and loaded it into the new mesh. The coordinates are the same. The values in the profile look good. But the interpolation won't work. A few cells have very high velocities and I can't figure out why. The cells have about 20 m/s, but a few cells have over 10.000 m/s. That makes the flow extremely unstable. Does anybody have an idea?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
area does not match neighbour by ... % -- possible face ordering problem	St.Pacholak	OpenFOAM	11	September 4, 2024 05:28
Mapping (interpolation) between two meshes	Ebrahim	OpenFOAM Programming & Development	2	June 15, 2020 14:31
Heat Transfert, mixing problem	Wazdq	CFX	8	November 7, 2016 06:44
Surface interpolation schemes and parallelization	jutta	OpenFOAM Running, Solving & CFD	0	February 25, 2010 15:32
Interpolation problem!	Severin	FLUENT	1	July 28, 2006 07:10

August 16, 2017, 03:33		#2
CeesH Senior Member Cees Haringa Join Date: May 2013 Location: Delft Posts: 607 Rep Power: 0	Hi, Interpolation always comes with some risk in wall regions or regions with poor cells, where gradients may be very steep. You will anyway have to run a couple of iterations to reconverge your flow; perhaps the simplest option is to patch the poor zones. If you can mark the cells with the high velocities (don't have FLUENT open now, so can't give you the exact names of what to click on) and then use patch (under initialization) to set the velocities in these cells to some more realistic value, that may stabilize your solution. Good luck!

August 16, 2017, 18:05		#3
sufjanst Member Join Date: Mar 2016 Posts: 73 Rep Power: 10	I created an o-grid for the inlet region with a very good quality (over 0.8). But it didn't work... It seems like the interpolation just won't work. https://picload.org/view/rwlowdcl/profile.png.html I uploaded a screenshot of the inlet. This is the velocity magnitude from 0-50 m/s. It looks just like the profile from the old simulation. But there are a few cells with very high velocities (transparent regions). Do you have any idea?

August 17, 2017, 05:05		#4
CeesH Senior Member Cees Haringa Join Date: May 2013 Location: Delft Posts: 607 Rep Power: 0	Did you use an inflation layer/very small cells near the boundary? I can imagine the gradients there are very steep, which may cause (presumably?) linear interpolation to cause huge overshoots. Anyway, the most pragmatic solution seems to me to use some form of clipping/patching to remove excess velocities, as stated before.

August 18, 2017, 09:26		#5
sufjanst Member Join Date: Mar 2016 Posts: 73 Rep Power: 10	I already tried patching the inlet, but after convergence it got to the old state with the high velocities. I also tried to write the profile from a plane 5mm above, but that didn't work out as well. Is there a possibility to block the high velocities with an UDF? Do I need an inflation layer? I use a Hexa-Mesh. The elements are parallel anyway... Thank you for your replies.