Minimum working example Utah brain

[hovey]

Hi @tripaanu do you want to work together at all on this item this week (17-20 Oct 2022)?

Background and links to the two FE meshes.

Action - Perform simulation and analysis on the two meshes:

• 32,248 elements (coarse)
• 244,136 elements (refined)

[tripaanu]

Hi Chad,

Sorry, couldn't respond sooner- I will work on it this week.

Best Anu

On Mon, Oct 17, 2022 at 4:40 PM Chad B. Hovey @.***> wrote:

Assigned #19 https://github.com/autotwin/mesh/issues/19 to @tripaanu https://github.com/tripaanu.

— Reply to this email directly, view it on GitHub https://github.com/autotwin/mesh/issues/19#event-7605727600, or unsubscribe https://github.com/notifications/unsubscribe-auth/A3RIEEP6FDR6CQQXDBNSK5LWDW2S5ANCNFSM6AAAAAARHNYX3Q . You are receiving this because you were assigned.Message ID: @.***>

[hovey]

@tripaanu

Confirmation and TODO

• Correct:
  • the two models are cell_050_050_050.inp and cell_100_100_100.inp
  • (assumed correct for now) models are in millimeters length scale
  • one material (i.e., one "section"), and assigning as gray matter for now, homogeneous, isotropic
• To do
  • I can get you the "skull" of both of these models this afternoon (on 18 Oct 2022)

Q and A

You asked "Please let me know if I am using the wrong files or missing something."

• Answer: You are using the correct files. I don't see anything that is missing, aside from the need for the skulls, which I will get to you as soon as I am able this afternoon.

Regarding, "Also, I can scale the models in meters myself but it might be easier to just scale it right when you are creating the mesh file. Let me know if that works for you - I am good either way."

• Answer: I agree, scaling should be perfect upstream, at the mesh representation. We will refine this issue moving forward. For now, the main goal is to automate the process. Refinements later integrate in a trivial manner once our process is automated.

[hovey]

@tripaanu What is your requirement on the interface between the interior of the skull and the exterior of the brain?

Specifically, are you expecting node-to-node contact, wherein the interior surface nodes tie 1-to-1 to the exterior nodes of the brain?

[tripaanu]

Hi Chad,

Yes, I do need node-to-node contact between the brain and the skull.

Best Anu

On Tue, Oct 18, 2022 at 5:17 PM Chad B. Hovey @.***> wrote:

@tripaanu https://github.com/tripaanu What is your requirement on the interface between the interior of the skull and the exterior of the brain?

Specifically, are you expecting node-to-node contact, wherein the interior surface nodes tie 1-to-1 to the exterior nodes of the brain?

— Reply to this email directly, view it on GitHub https://github.com/autotwin/mesh/issues/19#issuecomment-1283014760, or unsubscribe https://github.com/notifications/unsubscribe-auth/A3RIEEJVHKSHCCZPJ65ROZ3WD4HWRANCNFSM6AAAAAARHNYX3Q . You are receiving this because you were mentioned.Message ID: @.***>

[hovey]

@tripaanu Afternoon update: It is a new requirement (which is of course fine bc this is work in progress) to have node-to-node contact between interfacing materials. Prior to today, I had envisioned node-to-surface (a more general interface). So, I will need probably another day to rework our sculpt_stl_to_inp.py script.

@calehma I will update the sculpt_stl_to_inp.py on a branch separated from the main branch. So, my dev work won't interfere with your current IGEA case study.

[hovey]

@tripaanu here is a preview of how things will be coming to you. Attached are two images that were done with Sculpt via the Cubit GUI. I want to implement this manual workflow into the automated sculpt_stl_to_inp.py, so I need some time for additional (multi-material) implementation.

My Q to you is this: If the meshes (brain and outer) came to as shown in these images, does this meet a specification you can simulate with? Thank you.

Zoomed image with brain highlighted in yellow and outer not highlighted.

[hovey]

Hi @tripaanu I uploaded two .inp files, details here (on a branch called multi-matl): https://github.com/autotwin/mesh/tree/multi-matl/doc/T1_Utah_SCI_brain

[tripaanu]

Hi Chad,

I will work on it today and let you know if I run into any roadblocks- so far it looks good.

Best Anu

On Wed, Oct 19, 2022, 7:39 PM Chad B. Hovey @.***> wrote:

Hi @tripaanu https://github.com/tripaanu I uploaded two .inp files, details here (on a branch called multi-matl): https://github.com/autotwin/mesh/tree/multi-matl/doc/T1_Utah_SCI_brain

— Reply to this email directly, view it on GitHub https://github.com/autotwin/mesh/issues/19#issuecomment-1284705274, or unsubscribe https://github.com/notifications/unsubscribe-auth/A3RIEEIZWAPPHUCR6CQA6NTWECBC7ANCNFSM6AAAAAARHNYX3Q . You are receiving this because you were mentioned.Message ID: @.***>

[hovey]

@tripaanu I am interested if you would like to have a Teams discussion on this item this week?

[hovey]

@tripaanu with cc to @carlsenrw @elejeune11

I have nodesets now created on the blocks (finite element meshes) automatically from within our Python script that runs Sculpt on isosurfaces (see 1bb036a5ee8ae71bb9085e7ff307bdf4740abffa).

Also, I have created an examplar problem called two_spheres that is exactly the same as the Utah brain model (two materials, inner and outer), but a simplified verison of the Utah brain model.

I would like to see if @tripaanu can use the two_spheres (see https://github.com/autotwin/mesh/tree/main/doc/two_spheres) example to

1. Define the outer block as rigid skull and the inner block as brain, does the same ABAQUS error exist as with the Utah model?
2. Use the nodeset on block 2 alone (without the outer block), and slosh the inner brain using boundary conditions alone.

Do you have time to work on this item this week (21, 22, 23 Nov 2022)? We could do a Teams screenshare.

Thanks.

[tripaanu]

Hi Chad,

Yes, I am back to work now and I will work on this item this week. I will update you with any progress soon.

Best Anu

On Mon, Nov 21, 2022, 6:12 PM Chad B. Hovey @.***> wrote:

@tripaanu https://github.com/tripaanu with cc to @carlsenrw https://github.com/carlsenrw @elejeune11 https://github.com/elejeune11

I have nodesets now created on the blocks (finite element meshes) automatically from within our Python scrip that runs Sculpt on isosurfaces (see 1bb036a https://github.com/autotwin/mesh/commit/1bb036a5ee8ae71bb9085e7ff307bdf4740abffa ).

Also, I have created an examplar problem called two_spheres that is exactly the same as the Utah brain model (two materials, inner and outer), but a simplified verison of the Utah brain model.

I would like to see if @tripaanu https://github.com/tripaanu can use the two_spheres (see https://github.com/autotwin/mesh/tree/main/doc/two_spheres) example to

1. Define the outer block as rigid skull and the inner block as brain, does the same ABAQUS error exist as with the Utah model?
2. Use the nodeset on block 2 alone (without the outer block), and slosh the inner brain using boundary conditions alone.

Do you have time to work on this item this week (21, 22, 23 Nov 2022)? We could do a Teams screenshare.

Thanks.

— Reply to this email directly, view it on GitHub https://github.com/autotwin/mesh/issues/19#issuecomment-1322780405, or unsubscribe https://github.com/notifications/unsubscribe-auth/A3RIEEKBBX5XXOEQ6CTZLXDWJP6UZANCNFSM6AAAAAARHNYX3Q . You are receiving this because you were mentioned.Message ID: @.***>

[hovey]

Hi @tripaanu

Do you have availability 11/30 or 12/01 to discuss? Here are discussion points:

• brain only and brain + outer both not working for coarse and fine resolutions: Hex element count reference:

file	brain	outer	total
cell_size_4.inp	34,341	60,973	95,314
cell_size_2.inp	244,105	403,494	647,599

• two_spheres example works reference:

file	inner	outer	total
two_spheres.inp	152	1,048	1,200

[hovey]

Hi @tripaanu - I just discovered an error and trying to understand it. So, can we pause any plans of the two of us meeting until I figure this error out? Thanks

[hovey]

Using the T1_Utah_SCI_brain.stl input file:

and then manually (e.g., with the Cubit GUI) meshing just the brain with cell_size = 8 gives a floating element:

cell_size = 4 also produces floaters:

cell_size = 2 same:

[hovey]

@tripaanu with cc to @carlsenrw and @elejeune11

I have redone the "brain only" models at cell_size = 2 mm (fine) and cell_size = 4 mm (coarse) refinement levels, now including a nodeset on the surface of the brain.

You can get the two ABAQUS input files from two links on this page (circled in yellow on the screenshot below):

https://github.com/autotwin/mesh/blob/main/doc/T1_Utah_SCI_brain/README.md

Here is a screenshot of that page, with the links cell_050_050_050_v02.inp and cell_100_100_100_v02.inp shown in blue.

Can you please try the brain slosh analysis again with these input files? Thank you.

[tripaanu]

Hi Chad,

Thank you for fixing this! I will run the jobs this week and update.

Best Anu

On Thu, Dec 1, 2022 at 10:34 AM Chad B. Hovey @.***> wrote:

@tripaanu https://github.com/tripaanu with cc to @carlsenrw https://github.com/carlsenrw and @elejeune11 https://github.com/elejeune11

I have redone the "brain only" models at cell_size = 2 mm (fine) and cell_size = 4 mm (coarse) refinement levels, now including a nodeset on the surface of the brain.

You can get the two ABAQUS input files from two links on this page (circled in yellow on the screenshot below):

https://github.com/autotwin/mesh/blob/main/doc/T1_Utah_SCI_brain/README.md

Here is a screenshot of that page, with the links cell_050_050_050_v02.inp and cell_100_100_100_v02.inp shown in blue.

[image: Screen Shot 2022-12-01 at 8 24 23 AM] https://user-images.githubusercontent.com/1473854/205093634-15b1c845-6c45-42c3-b50f-3dfb0d1e20a8.png

Can you please try the brain slosh analysis again with these input files? Thank you.

— Reply to this email directly, view it on GitHub https://github.com/autotwin/mesh/issues/19#issuecomment-1333948113, or unsubscribe https://github.com/notifications/unsubscribe-auth/A3RIEEPKHU7P7MEEPKNR4FTWLDAQBANCNFSM6AAAAAARHNYX3Q . You are receiving this because you were mentioned.Message ID: @.***>

[hovey]

Hi @tripaanu (cc @elejeune11 @wangyrmu @carlsenrw )

Briefly, I have been working this week to get adaptivity working. Adaptivity in Sculpt is a strategy to create small sized element sizes near features, esp with curvature (e.g., surface gyri/sulci), and then transition to larger elements toward the interior of the domain (brain). But, many of the trials either had either 10s of millions of elements (too many to start), or had negative MSJ values. I will continue to investigate. The levers/settings that can be used with Sculpt is large, and the process quickly becomes combinatoril!

Ball still in my court. Have to pause an attend to other deliverables on other projects for a few days.

ttyl. cbh

[hovey]

@tripaanu (cc @carlsenrw @wangyrmu @CompMechPitt )

For your reference, here is the documentation for how I created the new candidates meshes of cell size 8, 4, and 2.

Please find the .inp files here: Dropbox/scratch/UTAH_SCI_brain/

cell size	link to .inp file on Dropbox
8	https://www.dropbox.com/s/1be50nw9nq6hofb/cell_size_8.0_2023-05-22_UTC_20_50_44_794153.inp?dl=0
4	https://www.dropbox.com/s/xxn7fsq772qxfjv/cell_size_4.0_2023-05-22_UTC_20_50_44_795467.inp?dl=0
2	https://www.dropbox.com/s/lnn6a0u2ji2jozo/cell_size_2.0_2023-05-22_UTC_20_50_44_796321.inp?dl=0

These .inp files are generated from @elejeune11 new .stl file T1_Utah_SCI_brain_5_16_2023_20_25.stl

[hovey]

Notes on @tripaanu email update of 2023-06-07-0903 (copy to @carlsenrw and @wangyrmu ):

• Cell size 4 and 2 also failed at early time similar to cell size 8 simulations (with RMU material properties and relaxation time from 0.0014 s [RMU] to 0.0250 s [Terpsma et al.]).
• Cell size 8 with (1/ bulk modulus)
  • from D = 9.132e-10 Pa^-1; K = 1.095e+09 Pa [RMU]
  • to D = 2.28e-10 Pa^-1; K = 4.386e+09 Pa
  • does complete (but strain distribution is different from RMU mesh model All_Hex_dec.inp).
• Impression: Mesh is fine, but material properties may need updates.

Chad's impression: a fully working image to analysis workflow is a priority, thus simulation completion is a priority. Material properties research is a secondary priority.