[Eloise]

Hi Bruno,
I finally had time to work again on this setup, at least on the cylinder case. The cyclic AMI condition works great! I'll try in short future to apply it to the sphere, and then to apply a rotation in between each computation to speed up the convergence.
Thanks for the valuable help!
Regards,
Eloïse

[esujby]

Hello,

Please i am wondering, where do i get the tolerancedict to add to the constant folder for the multi patch utility?

thanks

[Eloise]

Hi,

I couldnt find it in 2.4.x distribution. But it seems the option -toleranceDict is still available in stitchMesh. Here is what the file should contain with the default values for all parameters.

Regards,
Eloise

FoamFile
{
    version     2.0;
    format      ascii;
    class       dictionary;
    object      toleranceDict;
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 
tolerances
    {
pointMergeTol            0.05;
edgeMergeTol             0.01;
nFacesPerSlaveEdge       5   ;
edgeFaceEscapeLimit      10  ;
integralAdjTol           0.05;
edgeMasterCatchFraction  0.4 ;
edgeCoPlanarTol          0.8 ;
edgeEndCutoffTol         0.0001;
    }

[Antimony]

Hi All,

I am not sure how useful this post would be coming as it does about 2 years since the last post.

Nevertheless, I thought it would be useful for anyone, who like me, has been wondering how to solve the issue with stitchMesh.

Just a background on the problem I was trying to solve: It was the case that Bruno had mentioned earlier - stitching together meshes to the north, south, east and west of a given mesh. Any two would work, after which it would invariably cause problems.

I tried the utility that Bruno had put up as well - stitchMeshMultiPatch. My problem didn't get solved completely - any two would work, this time better than using stitchMesh individually, but still the problem persisted.

I read in the forums that cyclicAMI would be a better option and that is what I tried. Initially, I ran into some problems in it as well. Not with joining the mesh, but with running a simulation after.

Digging further showed that an important parameter called "transform" could be specified in the "boundary" file. I initially tried the first one I found "noOrdering" but it lead me nowhere.

Using the "banana" method, I found that there is one called "coincidentFullMatch" and gave a shot with that.

Presto, it worked!

So, at the end of this long story, for anyone looking to stitch meshes like the one I just mentioned, use the "cyclicAMI" type and "coincidentFullMatch" for the "transform" value.

Hope this proves to be useful to someone in the future!

Cheers,
Antimony

[Antimony]

Hi All,

Just had time to relook into stitchMesh and found this workaround, which seems to work.

Note that I have not tested it extensively, but I did try it in a couple of cases and it seemed to work. You have been warned!

Essentially what I found was that you can stitch two meshes well only a couple of times. Beyond that there is going to be a problem. However, if you change the way the merging and stitching is done, then you might not have the usual problem with stitchMesh.

Consider a 3x3 grid of meshes (Denote as L1,L2,L3,M1,M2,M3 and U1,U2,U3), with all of them having different mesh resolutions. The point I was trying to make was this:

1. Merge L1 and L2. And stitch using right patch of L1 and left patch of L2
2. From constant/polyMesh remove the meshModifiers file.
3. Merge L1L2 and L3. Stitch using right patch of L1L2 and left patch of L3.
4. Repeat step 2
5. Using createPatch, combine the front faces in L1L2L3 into a single patch/face (and optionally a single front face)
6. Repeat 1 through 4 for the middle meshes (M1,M2,M3) and the top meshes (U1,U2,U3)
7. For the middle meshes, use createPatch to create a single front face and a single back face.
8. For the upper meshes, use createPatch to create a single back face (and optionally a single front face).
9. Merge L1L2L3 and M1M2M3. Stitch using the front patch of L1L2L3 and back patch of M1M2M3.
10. Repeat step 2.
11. Merge L1L2L3M1M2M3 and U1U2U3. Stitch using the front patch of L1L2L3M1M2M3 and back patch of U1U2U3.
12. Repeat step 2.

Hopefully, at the end of this you will have a single mesh, with no errors in the process of stitching them.

It is a tedious process, but it seems to work. Again, I haven't tested it in a variety of configurations, and have had some trouble when trying to stitch cylindrical patches.

Hope this is useful!

Cheers,
Antimony

[Bashar]

Originally Posted by Antimony

1. Merge L1 and L2. And stitch using right patch of L1 and left patch of L2
2. From constant/polyMesh remove the meshModifiers file.
3. Merge L1L2 and L3. Stitch using right patch of L1L2 and left patch of L3.
4. Repeat step 2
5. Using createPatch, combine the front faces in L1L2L3 into a single patch/face (and optionally a single front face)
6. Repeat 1 through 4 for the middle meshes (M1,M2,M3) and the top meshes (U1,U2,U3)
7. For the middle meshes, use createPatch to create a single front face and a single back face.
8. For the upper meshes, use createPatch to create a single back face (and optionally a single front face).
9. Merge L1L2L3 and M1M2M3. Stitch using the front patch of L1L2L3 and back patch of M1M2M3.
10. Repeat step 2.
11. Merge L1L2L3M1M2M3 and U1U2U3. Stitch using the front patch of L1L2L3M1M2M3 and back patch of U1U2U3.
12. Repeat step 2.

Hi,
Thanks for sharing these instructions! I am trying your approach but I think I didn't quite get it. I attached my two meshes that I want to stitch the interfaces between them from all direction NS and EW . The small mesh should be inside the larger meshes that is why you will see that the interfaces are names a_l b_l ... and a_s b_s ... Do you think this is doable using your approach?

[Bashar]

Hi,
I managed to make the two meshes works. First, I stitch the patches with the flow stream since its essential in the solution. So , I stitched the left two patches (a_s and a_l) and the two to the right (c_s and c_l) . Next step was to use the cyclicAMI interface for the patches at the North and South , so I combined (d_l and d_s) and (b_l and b_s). I checked the mesh using checkMesh and visually and its seems to be very good. I hope it will work in the simulation without any issues.

Best wishes!

Bashar

[itsme_kit]

Originally Posted by Eloise

Hello Bruno,

Here are some test cases:
1) for a Cylinder: run .\Allrun_Cylinder. You can see that is works by looking at the log.checkMesh where the two internal patches have no faces.
2) for a Sphere: run .\Allrun_Sphere. The stitchMesh ends with an error.
3) for a Block: no test case, but I'd suggest to simply replace the sphere by a box in the Sphere test case.

Thanks for your interest in this case
Eloïse

Hi Eloise

I'm wondering if you have figured out why those two internal patches have no face after stitchMesh.

Looking forward to hearing from you.

Best Regards,

Kit

[seyfigirgin]

Hi everyone,

I am new to stitchMesh and want to learn about it. I have two patches, they perfectly match each other. I want to combine them and so that I will create 1 patch.

What is the way of using stitchMesh command. Could someone describe this to me? Lets say I have patch1.msh and patch2.msh

What would be the entire command line be like?

Regards,