sliding mesh for valves & sonicTurbDyMEngineFoam
 

Hi foamers ,

I am simulating an engine with sonicTurbDyMEngineFoam and simpleEngineTopoFVmesh library.
I have a question about using sliding mesh for valves .

In a 2D case mesh generated with blockMesh and two blocks generated in the valve curtain and they are separated by using different vertices . My problem comes when I want to use gambit for mesh generation (Actually I'm testing gambit for 3D simulation)
In gambit I splite the face in valvecurtain and I created two edges at valvecurtain.But when I imported mesh in openfoam and I ran the case I had an error :

I think this error comes from sliding mesh!!! I confused....
please anyone help me...any idea??How can I solve this problem??

Thanks and best regards,
Sasan.

I appreciate any help from you....

Hi sasan

I have reached the same problem as you. The problem is in the mesh generation.
I suppose that you have generated whole geometry at once in gambit. The problem is truly in the sliding interface. In the case of whole geometry generation in some "mesher" (I use Ansa) the stitchMesh (basically static sliding interface) needs the -perfect flag since the points on Cyl and Port are same (within some tolerance). But this is impossible in the case of Curtain in ICE, because of the valve inside. I have evaded this problem by generating all the parts separately (ports, cylinder, intakePorts, exHaustPorts). Then I used mergeMeshes to put them together. In this case the topological errors disappeared. But I must warn you. I have not been able to force the valves to move.
The motion of piston works fine, but when the valve should move the solver just stops with any error message. I have tracked the error and it show up, that is something wrong in the line

phi = fvc::interpolate(rho)*((fvc::interpolate(U) & mesh.Sf()) - fvc::meshPhi(rho, U));

but there my knowledge of OF ends.
Does anyone know, where is the dog buried?

Hi Martin ,

Can you explain what does stitchMesh do exactly?
Is it necessary for our simulation???

thanks,
Sasan.

StitchMesh stitch the mesh. It is not necessary for the simulation, but it is a good pointer on the validity of the mesh. If stitchMesh works correctly, then the mesh is valid for the sliding interface, i.e the sliding interface is possible.

Hi Martin

I have an error when I use stitchMesh...

--> FOAM FATAL ERROR:
Face 2229 reduced to less than 3 points. Topological/cutting error B.
Old face: 2(22 1432) new face: 2(22 1432)
old points: 2((-0.01 -0.003 0.00107703) (-0.01 -0.003 -0.00107703))

From function void slidingInterface::coupleInterface(polyTopoChange& ref) const
in file polyMeshModifiers/slidingInterface/coupleSlidingInterface.C at line 1765.

FOAM aborting

Aborted (core dumped)

It means that the slave patch and master patch have same vertices????
please correct me if I am wrong...
In Gambit I splite my geometry with face and connect selector switch was deactive and I created two edges in the interface (2D case)...where is problem????

thanks,
Sasan.

please help me for solving this problem......How can I remove this error???

Make sure your sliding interface patches are absolutely flat and that they lie in the same plane. The mesher may not do this sufficiently well. Check the cells located at coordinate (-0.01 -0.003 0.00107703). If they are awkward, remesh the whole thing and get rid of ugly cells where you use sliding interface.

K

Sorry, I have been very busy recently. Does the Kalle's solution help?

Hi Martin ,

I created the mesh again and I could overcome that error(thanks for karl's suggestion)...But I have a new problem in the 2D simulation that I described it in the below thread :
http://www.cfd-online.com/Forums/ope...imulation.html

And unfortunately I have another error for 3D simulation :
http://www.cfd-online.com/Forums/ope...ion-error.html

Can you help me? please help me..
I appreciate any help from you.

Thanks and best regards,
Sasan.