[Flo_Ha]

Hi all!

I´m doing a simulation with overset meshes. I ran already the tutorials that come with the new version v1706. Till some of the first time steps it seemes alright, but then i get a problem with the overset analysis( see part of the log-file)

logfile:

PIMPLE: iteration 1
DILUPBiCGStab: Solving for p, Initial residual = 0.02883, Final residual = 9.8924e-07, No Iterations 101
time step continuity errors : sum local = 1.9287e-11, global = 5.2236e-12, cumulative = -1.9194e-10
DILUPBiCGStab: Solving for p, Initial residual = 0.0030687, Final residual = 9.4395e-07, No Iterations 83
time step continuity errors : sum local = 1.7254e-11, global = -1.2876e-11, cumulative = -2.0482e-10
smoothSolver: Solving for epsilon, Initial residual = 0.044259, Final residual = 7.0462e-07, No Iterations 1
bounding epsilon, min: -224.35 max: 55707 average: 130.87
smoothSolver: Solving for k, Initial residual = 0.026315, Final residual = 4.6123e-07, No Iterations 1
bounding k, min: -0.058255 max: 39.248 average: 0.17978
ExecutionTime = 265.99 s ClockTime = 269 s

Courant Number mean: 0.0056557 max: 10.001
deltaT = 3.3825e-05
Time = 0.000996883

inverseDistance : detected 2 mesh regions
zone:0 nCells:40476 voxels57 57 57) bb-0.66432 0.2355 -0.645) (1.3357 1.2355 -0.145)
zone:1 nCells:332945 voxels57 57 57) bb-0.15859 0.4506 -0.46) (0.36683 0.74074 -0.33)
Overset analysis : nCells : 373421
calculated : 363413
interpolated : 9264 (interpolated from local:3364 mixed local/remote:29 remote:5871)
hole : 744

PIMPLE: iteration 1
DILUPBiCGStab: Solving for p, Initial residual = 0.024191, Final residual = 8.6156e-07, No Iterations 85
time step continuity errors : sum local = 1.6103e-11, global = -7.3635e-13, cumulative = -2.0555e-10
DILUPBiCGStab: Solving for p, Initial residual = 0.0023978, Final residual = 8.8674e-07, No Iterations 45
time step continuity errors : sum local = 1.6293e-11, global = -1.2992e-11, cumulative = -2.1854e-10
smoothSolver: Solving for epsilon, Initial residual = 0.042515, Final residual = 6.8282e-07, No Iterations 1
bounding epsilon, min: -159.62 max: 66858 average: 129.59
smoothSolver: Solving for k, Initial residual = 0.025049, Final residual = 4.4227e-07, No Iterations 1
bounding k, min: -0.089723 max: 40.213 average: 0.17932
ExecutionTime = 277.5 s ClockTime = 281 s

Courant Number mean: 0.0056548 max: 10.003
deltaT = 3.3815e-05
Time = 0.0010307

inverseDistance : detected 2 mesh regions
zone:0 nCells:40476 voxels57 57 57) bb-0.66432 0.2355 -0.645) (1.3357 1.2355 -0.145)
zone:1 nCells:332945 voxels57 57 57) bb-0.1581 0.45058 -0.46) (0.367 0.741 -0.33)
Overset analysis : nCells : 373421
calculated : 325906
interpolated : 10793 (interpolated from local:3937 mixed local/remote:79 remote:6777)
hole : 36722

As you can see, in the beginning the number of cells defined as "hole" is around 750, but then it switches immediately to 36700. That means all the cells in the surrounding mesh are also defined as hole. Does anybody have an idea what the reason could be?

Best Regards
Flo

[Flo_Ha]

It seems to be a 2d->3d problem.

I did the same tutorials, which are 2d cases, as a 3d case. And now i have the same problem also in the tutorials.....

does anyone have an idea, what could be wrong?

Regards
Flo

[Flo_Ha]

I have solved the problem.

First mistake: the overset patch had some overlapping elements.
Second mistake: The background mesh was to coarse.

Now it runs fine

Best Regards
Flo

[epc1]

Hello:

I am trying out the overset implementation in the same openfoam release as you
posted.

I am trying to model a 2D cilynder-piston assembly but so far I have not been
able to figure out how to properly define the "hole " in my case.

The problem is that my overset mesh is defined in such a way that
no holes need to be cut for it. Rather, the hole needs to be cut
for the background mesh.

In the attached figure the grid with red cells is supposed to be the boundary
fitted overset mesh for the piston. With the moving piston wall on the right hand side
of the domain.

The hole then need to be defined as the region between this moving piston wall, and
the background blue grid.

So far I have tried the using the "simplerotor" tutorial example but the hole is not
defined correctly as you can see from the "celltypes" field in the attached image.

Could you take a look an see if you can help setting up the case?

Thanks in advance

[Flo_Ha]

Sorry for the late response .....

Did you already solve your problem?

CellType 0..... calculated
CellType 1..... interpolated
CellType 2..... hole

So in your picture i´m not missing the holes. I´m missing the interpolated cells.

Best Regards
Flo

[nixiaonan]

I have the same problem of this.

logfile:

IMPLE: iteration 1
DICPCG: Solving for cellDisplacementx, Initial residual = 1, Final residual = 3.70390364717e-07, No Iterations 19
DICPCG: Solving for cellDisplacementy, Initial residual = 0, Final residual = 0, No Iterations 0
inverseDistance : detected 2 mesh regions
zone:0 nCells:600 voxels30 30 1) bb-2.69443871706e-06 -2.69443871706e-06 -2.69443871706e-06) (2.50000269444 1.00000269444 0.100002694439)
zone:1 nCells:300 voxels30 30 1) bb0.249999285857 0.249999285857 -7.14142842854e-07) (0.750000714143 0.750000714143 0.100000714143)
Overset analysis : nCells : 900
calculated : 600
interpolated : 300 (interpolated from local:300 mixed local/remote:0 remote:0)
hole : 0

As you can see, why my hole's number is “0” ?

Anyone can help me? any suggestion would be appreciated.

[epc1]

Originally Posted by Flo_Ha

Sorry for the late response .....

Did you already solve your problem?

CellType 0..... calculated
CellType 1..... interpolated
CellType 2..... hole

So in your picture i´m not missing the holes. I´m missing the interpolated cells.

Best Regards
Flo

Hello,

Yes, i already solved it. The problem was not the missig the overset patches,
they were setup correctly. The problem was setting up the hole patches correctly.
Thanks.

[epc1]

Originally Posted by nixiaonan

I have the same problem of this.

logfile:

IMPLE: iteration 1
DICPCG: Solving for cellDisplacementx, Initial residual = 1, Final residual = 3.70390364717e-07, No Iterations 19
DICPCG: Solving for cellDisplacementy, Initial residual = 0, Final residual = 0, No Iterations 0
inverseDistance : detected 2 mesh regions
zone:0 nCells:600 voxels30 30 1) bb-2.69443871706e-06 -2.69443871706e-06 -2.69443871706e-06) (2.50000269444 1.00000269444 0.100002694439)
zone:1 nCells:300 voxels30 30 1) bb0.249999285857 0.249999285857 -7.14142842854e-07) (0.750000714143 0.750000714143 0.100000714143)
Overset analysis : nCells : 900
calculated : 600
interpolated : 300 (interpolated from local:300 mixed local/remote:0 remote:0)
hole : 0

As you can see, why my hole's number is “0” ?

Anyone can help me? any suggestion would be appreciated.

Hello. I could have a look but i'd need pictures of the grids and a description
of the intended overset/hole patches.

If you follow the tutorials you can quickly realize that the trick is defining correctly
the hole patches which is done via topoSet

[nixiaonan]

Originally Posted by epc1

Hello. I could have a look but i'd need pictures of the grids and a description
of the intended overset/hole patches.

If you follow the tutorials you can quickly realize that the trick is defining correctly
the hole patches which is done via topoSet

thank you very much. can you help me have a look at the case?

I have attach my case named tut-2.zip . at next post.

[nixiaonan]

I have the same problem of this.

logfile:

IMPLE: iteration 1
DICPCG: Solving for cellDisplacementx, Initial residual = 1, Final residual = 3.70390364717e-07, No Iterations 19
DICPCG: Solving for cellDisplacementy, Initial residual = 0, Final residual = 0, No Iterations 0
inverseDistance : detected 2 mesh regions
zone:0 nCells:600 voxels30 30 1) bb-2.69443871706e-06 -2.69443871706e-06 -2.69443871706e-06) (2.50000269444 1.00000269444 0.100002694439)
zone:1 nCells:300 voxels30 30 1) bb0.249999285857 0.249999285857 -7.14142842854e-07) (0.750000714143 0.750000714143 0.100000714143)
Overset analysis : nCells : 900
calculated : 600
interpolated : 300 (interpolated from local:300 mixed local/remote:0 remote:0)
hole : 0

As you can see, why my hole's number is “0” ?

Anyone can help me? any suggestion would be appreciated.

[Flo_Ha]

@nixiaonan: Is your "hole" patch a closed surface? it seems the sover doesn´t recognize it to cut it out. can you post a picture?


I´m having a new problem now

My case runs just fine, and then after some random time it crashes without any reason i could imagine.

Did anyone already had that problem?

see attached logfile

Best Regards
Flo

[nixiaonan]

Originally Posted by Flo_Ha

@nixiaonan: Is your "hole" patch a closed surface? it seems the sover doesn´t recognize it to cut it out. can you post a picture?


I´m having a new problem now

My case runs just fine, and then after some random time it crashes without any reason i could imagine.

Did anyone already had that problem?

see attached logfile

Best Regards
Flo

hello Flo, my problem had benn solved,it is the reason of 2D geometry, but now here have a other problem, please see my picture,why my picture has the sawtooth shape when the overset moving ?

[nixiaonan]

Originally Posted by Flo_Ha

@nixiaonan: Is your "hole" patch a closed surface? it seems the sover doesn´t recognize it to cut it out. can you post a picture?


I´m having a new problem now

My case runs just fine, and then after some random time it crashes without any reason i could imagine.

Did anyone already had that problem?

see attached logfile

Best Regards
Flo

hello Flo, my problem had benn solved,it is the reason of 2D geometry, but now here have a other problem, please see my picture,why my picture has the sawtooth shape when the overset moving ?

[nixiaonan]

Originally Posted by Flo_Ha

@nixiaonan: Is your "hole" patch a closed surface? it seems the sover doesn´t recognize it to cut it out. can you post a picture?


I´m having a new problem now

My case runs just fine, and then after some random time it crashes without any reason i could imagine.

Did anyone already had that problem?

see attached logfile

Best Regards
Flo

There seems to be some error in the boundary condition setting. You need to check it carefully

[beishannmuyu]

Originally Posted by Flo_Ha

I have solved the problem.

First mistake: the overset patch had some overlapping elements.
Second mistake: The background mesh was to coarse.

Now it runs fine

Best Regards
Flo

Hi flo
I have the same problem in my tutorials.in the beginning the number of cells defined as "hole" is around 2808, But when it comes to 4.8 seconds ,then it switches to 284210.
I want to know (1)what's meaning about the overset patch had some overlapping elements???can you detailed explanation about it??
(2)Can you share your experience with the relation between background mesh and rigid mesh size?
thanks
Best Regrads
Li

[Flo_Ha]

Hi.

My case was not working when the background mesh was to coarse compared to the rotating mesh. But i cannot tell you a ratio which you should have. In the end i chose the same mesh size in that area, where i have my overset mesh. And then it worked. I guess i could have made it a bit coarser, but i didn´t try.

Concerning the overlapping elements, i do not remember anymore. Sorry for that

Best regards Flo

[HARISHKG]

Hi ,

I too have issue with openfoam overset mesh , as i get warning of

master overset patch should be the first patch.. and i dont whether it is affecting my results..

I have attached the images of the result..the interpolation is not happening between the overset boundary and the background..

[alpha23]

Originally Posted by HARISHKG

Hi ,

I too have issue with openfoam overset mesh , as i get warning of

master overset patch should be the first patch.. and i dont whether it is affecting my results..

I have attached the images of the result..the interpolation is not happening between the overset boundary and the background..

I have been experiencing the very same problem, where my background mesh is not updated while only the overset mesh is. I believe it has something to do with how regions are defined.

I basically used the overSimpleFoam tutorial case (aerofoil) and extended it to a 3D setting, defining regions using the topoSet functionality as it is done in that tutorial. But when overSimpleFoam starts it writes cellTypes (0, 1 and 2, that are respectively calculated, interpolated and holes) in a weird way, where the background mesh has all cells of cellType 2 (holes), so I believe this is the reason. But I don't know how to fix this, since no tutorial case is using 3D geometry files as my case, and the only way to define regions I can infer from tutorials is to use topoSet with a regionToCell using the point in the background for the first region, and the other region performing the inverse of the first one.

Does anyone have any advice about how to deal with 3D geometries?

Thanks a lot

[alpha23]

Let me add few details to be more clear:

I tried a very basic setup, extending the tutorial aerofoil for overSimpleFoam from 2D to a 3D setting. It is all explained in the following image: I have a valve plate inside an overset mesh that I want to add to a simple channel background mesh. But what I am finding is that, if I enclose my plate inside a STL cube (or blockmesh base) to delimit the external boundary of the overset mesh, everything goes well, while if I enclose my plate in a STL cylinder to delimit with it the external boundary of the overset mesh, when I run overSimpleFoam it wrongly deduces the cellType, it basically sees an additional wall at the overset Mesh external limit. The image explains it quite well.

Few notes:
1) If I use the STL cube to delimit the overset mesh or simply blockmesh it goes well
2) It is not a matter of snapping, the cube is snapped and I also tried to use castellated mesh for the cylinder but it still has the same problem
3) I need to use a STL defined external boundary for the overset mesh because it has to fit my internal geometries
4) I have been setting ZoneID using topoSet+setFields and it works correctly (apparently), in the sense that the two zones appears to be correctly separated

See the image (also in the attached pdf file)


I believe it is either a stupid mistake or a non-intuitive (wrong) behavior of the code...

[Ron.CC]

Hi Alessandro,

I'm into a case with 3D overset mesh and having the same problem. The celltypes are automatically detected in a wrong way though the zoneID are set correctly.

Have you figured out any solutions? Any help will be appreciated!