Welcome to blastAMR Discussions!

[FoamScience]

👋 Welcome!

We’re using Discussions as a place to connect with other members of our community. We hope that you:

• Ask questions you’re wondering about.
• Share ideas.
• Engage with other community members.
• Welcome others and are open-minded. Remember that this is a community we
  build together 💪.

To get started, comment below with an introduction of yourself and tell us about what you do with this community.

[MahadevP2001]

What is the correct way to restart simulations from a particular time? Will the cells created because of AMR previously be considered candidates for unrefinement after restarting the simulation?

I am trying to integrate this library with my own solver and I notice that when I restart the simulation, the number or split cells considered for unrefinement starts again from zero which makes me think that it not considering the refinement levels at that time and treating all the cells from level zero. I am using hexRefiner

Would greatly appreciate your help on this.

[FoamScience]

If the restart doesn't consider blastAMR's own refinement history, it's a bug

If your simulation was running with another refinement engine before, history support is limited and young (#27). Make sure cellLevel and pointLevel files are emitted in the time folder so they can be read. These are usually labelIOList files (not full fields) and OpenFOAM may be a little inconsistent as to when to write them.

If you think this is a bug, please open an issue, ideally reproducible with one of the tutorials - else you may share more info (OF version, preprocessing setup) on your situation.

[MahadevP2001]

Thank you for the response. I do not know if It is because I set up the case incorrectly, but it was not writing a refinementHistory file to the time step folder like openFoams own AMR does. Is it supposed to do that? I have yet to try restarting a case with the tutorial but I will try that. I was not using any other refinement engine either.

[FoamScience]

Thank you for the response. I do not know if It is because I set up the case incorrectly, but it was not writing a refinementHistory file to the time step folder like openFoams own AMR does. Is it supposed to do that? I have yet to try restarting a case with the tutorial but I will try that. I was not using any other refinement engine either.

Yes you should find a polyMesh/refinementHistory if there is some history from other tools. blastAMR itself doesn't write this.
what's more important is the polyMesh/cellLevel and polyMesh/pointLevel

This does seem like a bug to me, will investigate it further

[MahadevP2001]

I think the restart issue is a problem with hexRefiner. I restarted the damBreak2D case from 0.4s using hexRefiner instead of polyRefiner and changing the refine and unrefine interval to 1. Replaced the last command in allRun with :

mpirun -np 4 interFoam -parallel 2>&1 | tee log

Time = 0.4

PIMPLE: iteration 1
---->! custom error estimator !<----
---->! end error estimator !<----
Protecting 0 cells next to requested boundary patches.
Selected 10 cells for refinement out of 14538.
Refined from 14538 to 14568 cells.
Expired sampledSurface in surfaceFieldValue: nonWeightedAreaAverage
Selected 40 split edges out of a possible 2284.
Unrefined from 14568 to 14448 cells.

deltaT = 0.000294118
Time = 0.400294

PIMPLE: iteration 1
---->! custom error estimator !<----
---->! end error estimator !<----
Protecting 0 cells next to requested boundary patches.
Selected 2 cells for refinement out of 14448.
Refined from 14448 to 14454 cells.
Expired sampledSurface in surfaceFieldValue: nonWeightedAreaAverage
Selected 0 split edges out of a possible 2.

Upon restarting it is only identifying 2 split edges when there used to be 2284 before. Isn't that an error? I do not face this error with polyRefiner. Infact even the cells created during updateMesh are retained and not unrefined with hexRefiner. I am not sure if I have made some error in installing the library so I would be very grateful if you can also attempt it. Do you think it could be because refinementHistory is not being written with hexRefiner?

hexRefineError.pdf

[MahadevP2001]

This does seem like a bug to me, will investigate it further

Was this reproducible by you too? To rule out any setup or installation error on my part.

[MahadevP2001]

I have a question regarding lowerRefineLevel , upperRefineLevel which are required entries and their unrefine counterparts. Do they matter when the errorEstimator is coded? Or does it only matter when errorEstimator is delta?

[FoamScience]

I have a question regarding lowerRefineLevel , upperRefineLevel which are required entries and their unrefine counterparts. Do they matter when the errorEstimator is coded? Or does it only matter when errorEstimator is delta?

they matter when normalize() is called in the coded one; all other error estimators call it too
if they feel confusing, just the rror field to -1 (unrefine) 0 (leave alone) and 1 (refine) manually

blastAMR/tutorials/damBreak3D/constant/dynamicMeshDict

Line 62 in f1a9851

if (scale) normalize(error_);

[MahadevP2001]

Even if a cell has error field set to 0 in the coded section, it will still be refined if it lies within in the buffer region of a cell which is selected for refinement, correct? If so are there any ways to prevent that ?

[FoamScience]

Even if a cell has error field set to 0 in the coded section, it will still be refined if it lies within in the buffer region of a cell which is selected for refinement, correct? If so are there any ways to prevent that ?

There is no way to prevent that, and we shouldn't prevent that. the error estimators only define refinement/unrefinement candidates. The algorithm has to figure out whether to actually refine the cell on its own. there is a lot of constraints that need to be satisfied to get a valid FVM mesh, and then there are some OpenFOAM-specific constraints (eg. not being able to share a potion of a face between two adjacent cells)

[MahadevP2001]

Are the cells created by refining and unrefining using the polyRefiner hexahedral in case of an initially hexahedral mesh?

[FoamScience]

Not reliably. The general refinement algorithm would theoretically result in a hex mesh but there is more to the unrefinement (and load-balancer) algorithms that may break that (also there is patch and cell zone protections ... etc) - so not guaranteed to generate a full hex mesh.