Handling lattice regions in mesh

[rnpratoori]

I have lattices produced by LatticeSmoothCircleIC or any other method. I want to be able to separately identify each of the lattice circles produced - either as a subdomain or some inidcator. Is there a way to do it? I am using CoupledVarThresholdElementSubdomainModifier but it only gives 1 subdomain id for all the lattices.
In the below image, the red circles are the lattices.

[GiudGiud]

Hello

Yes you can do this with the FeatureFloodCount objects, and the auxkernel associated
https://mooseframework.inl.gov/source/postprocessors/FeatureFloodCount.html

Guillaume

[rnpratoori]

I used this, but how do I assign subdomainIDs to each lattice? Or how do I access/isolate each of them?
What I want to achieve ultimately is - identify if a set of equations need to be solved in each lattice.

[GiudGiud]

identify if a set of equations need to be solved in each lattice.

This means setting the subdomain ID then. Equations are created using kernels that are restricted on subdomains.

I used this, but how do I assign subdomainIDs to each lattice

You ll likely need to write a MeshModifier, derived from the ElementModifierBase most likely, that assigns the featureID as the subdomain ID

[rnpratoori]

How do I visualize the FeatureFlood in paraview with featureIDs?
I used the following block at gives nothing -

[voids]
        type = FeatureFloodCount
        variable = eta
    []

[GiudGiud]

use theFeatureFloodAux to output the features to an auxvariable

[rnpratoori]

@GiudGiud Thank you for you pointers. FeatureFloodCountAux works well.

What I want to achieve -
When the value of an auxvariable (say c) becomes non-zero on the boundary (inside block 0) of the lattice object, I want to change the subdomain id of the lattice object (initially block 1) to block 0. This is so that the block 0 restricted equations can be solved in this region.

These lattice objects are identified by eta=1. So, I want to be able to change it to eta=0. This would trigger the CoupledVarThresholdElementSubdomainModifier to change the subdomain id to 0.

To do this, I am thinking of writing a custom GeneralUserObject. This would look for the "bounding nodes" of FeatureFlood objects and monitor the value of c at TIMESTEP_BEGIN. Is that the best way? Or is there any other better alternative?

[GiudGiud]

To do this, I am thinking of writing a custom GeneralUserObject. This would look for the "bounding nodes" of FeatureFlood objects and monitor the value of c at TIMESTEP_BEGIN. Is that the best way? Or is there any other better alternative?

This looks like a viable plan to me.

Alternatively, keeping with this

These lattice objects are identified by eta=1. So, I want to be able to change it to eta=0. This would trigger the CoupledVarThresholdElementSubdomainModifier to change the subdomain id to 0.

we just need to make eta=1 happen `when the value of an auxvariable (say c) becomes non-zero on the boundary (inside block 0) of the lattice object, I want to change the subdomain id of the lattice object (initially block 1) to block 0. This is so that the block 0 restricted equations can be solved in this region.

becomes non-zero on the boundary

Is this boundary an actual sideset? If we had a postprocessor that computes this boundary value we could do something like this in a ParsedAux for eta:

[AuxKernels]
  [set_inside_feature_i]
    type = ParsedAux
    variable = eta
    expression = 'if(feature_id == i, if(boundary_value_indicator > 0, 1, eta), eta)'

with :

• feature_id the variable that is set by the FeatureFloodCountAux
• boundary_value_indicator: a PP that detects becomes non-zero on the boundary by for example taking the integral of the value on that boundary

[rnpratoori]

Is this boundary an actual sideset? If we had a postprocessor that computes this boundary value we could do something like this in a ParsedAux for eta:

I am not sure how to create a side set for this. But since you mentioned boundary value - can I do something by calling the eta=1 regions as separate grains and calculate their boundary? I have never done any grain simulations too, so I am completely unaware of its capabilities.

[GiudGiud]

But since you mentioned boundary value - can I do something by calling the eta=1 regions as separate grains and calculate their boundary?

The grain tracker lets you find the separate grains.
However, the boundary is the region with 0 < eta < 1 around these grains, and currently we do not define a boundary sideset around them.

IMO, what you should do if you go that patch is create a "SideSetAroundGrainUpdater" MeshModifier, modeled after the SideSetAroundSubdomainUpdater. Then you can use a postprocessor on this sideset that will follow the grain boundary