3 d0 d coupling cap #381

dseyler · 2025-05-19T22:07:48Z

Couple struct/ustruct to LPN with virtual cap surfaces

Current situation

This pull request addresses SimVascular/svMultiPhysics/issues/73.

This feature has not been tested yet for LPNs made with svZeroDSolver.

Release Notes

This feature allows users to couple struct and ustruct simulations to an LPN using a "virtual" cap surface to enclose the surface to be coupled to the LPN.

A brief description of changes to the code:

The user must provide a "virtual" cap face that encloses a surface to be coupled to an LPN. This face does not lie on the volume mesh.
BC info for the capped surface is copied and a new BC is created for the virtual cap surface.
The flow rate coupled to the LPN is then calculated as the velocity flux over both the capped face and the cap.
The cap surface is not included in the pressure integral.

Documentation

An example of how to couple an LPN to a face enclosed by a virtual cap face is shown in dseyler/svMultiPhysics/tree/3D0DCouplingCap/tests/cases/struct/LV_NeoHookean_passive_genBC_capped.

Virtual faces are flagged in the solver.xml as follows:

<Add_face name="cap">
    <Face_file_path> mesh/mesh-surfaces/endo_cap.vtp </Face_file_path>
    <Virtual> true </Virtual> <!-- Flag for virtual capping surface -->
</Add_face>

The cap surface associated with a boundary condition is then defined as:

   <Couple_to_genBC type="I">
      <!-- Pressure returned by genBC is in dyne/cm^2 -->
      <ZeroD_code_file_path> genBC_svMultiPhysics/genBC.exe </ZeroD_code_file_path>
   </Couple_to_genBC>

   <Add_BC name="endo" > 
      <Type> Neu </Type> 
      <Time_dependence> Coupled </Time_dependence> 
      <Follower_pressure_load> true </Follower_pressure_load>
      <Capping_face> cap </Capping_face>    <!-- Specify capping surface for dVdt calculation on endo surface -->
   </Add_BC>

Once this pull request is approved, I will update the cardiac mechanics example in the user guide to include this feature.

Testing

Please ensure that the PR meets the testing requirements set by GitHub Actions.

In addition, please ensure all modified and new code is covered by tests.

Code of Conduct & Contributing Guidelines

I agree to follow the Code of Conduct and Contributing Guidelines.

virtual surfaces. Updated add_bc_mul.cpp to include capping surfaces in the flowrate integral but not in the pressure integral.

Merge main branch with 3D0DCouplingCap

ktbolt

We need to do some rewrite here.

And remove some of inline commenting that clutters the code.

Code/Source/liner_solver/add_bc_mul.cpp

Code/Source/solver/read_files.cpp

ktbolt · 2025-05-20T01:39:57Z

Code/Source/solver/set_bc.cpp

+        // Set resistance for capping BC if it exists for this BC.
+        // Set it to be the same resistance as the capped BC.
+        auto& iCapBC = bc.iCapBC;
+        if (iCapBC != -1) {


Once again we have an implicit meaning for -1 which is set somewhere. It is much better to have bool member data to this so then you can write

if (bc.is_coupling_cap) { }

and you don't need a comment and you can search to see who sets this.

ktbolt · 2025-05-20T01:45:15Z

Code/Source/solver/vtk_xml_parser.cpp

 {
  auto elem_ids = vtkIntArray::SafeDownCast(vtk_ugrid->GetCellData()->GetArray(ELEMENT_IDS_NAME.c_str()));
-  if (elem_ids == nullptr) { 
+  if (elem_ids == nullptr && !mesh.vrtual) { 


This is why if we treat a virtal face as not actually used like other faces we have to add all of these checks.

ktbolt · 2025-05-20T01:47:32Z

build_test.sh

@@ -0,0 +1,64 @@
+#!/bin/bash


What is all this data and scripts stuff ?

Ah this was a script that I wrote to test compiling and running the code quickly. Will remove before merging.

Copilot

Pull Request Overview

This PR introduces support for “virtual” cap surfaces to couple structural/ustru ct simulations to 0D LPN models.

Adds gnnbv and updates gnnb to compute normals on virtual faces via master-only gathers
Extends mesh loading, partitioning, and integration routines with vrtual flags and GatherMasterS/V helpers
Enhances FSILS (linear solver interface) to include capping surfaces in flow‐rate integrals

Reviewed Changes

Copilot reviewed 128 out of 128 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
Code/Source/solver/nn.cpp	New `gnnbv` for virtual face normals + updated calls
Code/Source/solver/load_msh.cpp	Read and initialize `face.vrtual`, `capID`
Code/Source/solver/distribute.cpp	Added `part_faceV` for virtual face partitioning
Code/Source/solver/all_fun.h/.cpp	New `GatherMasterV/S` routines for master gathers
Code/Source/liner_solver/*	Extended FSILS APIs to mark and skip virtual faces

Comments suppressed due to low confidence (2)

Code/Source/solver/all_fun.cpp:70

New virtual-face integration in GatherMasterV/S is not covered by existing tests. Please add unit tests that verify correct gathering behavior for owned and non‐owned nodes.

void GatherMasterV(const ComMod& com_mod, const CmMod& cm_mod, const Array<double>& s, const int Ac, Vector<double>& snode)

Code/Source/solver/ComMod.h:537

[nitpick] The field name vrtual is a misspelling of ‘virtual’. Consider renaming to isVirtual or virtualFace for clarity.

bool vrtual = false;

Code/Source/solver/distribute.cpp

Code/Source/solver/all_fun.cpp

Code/Source/solver/distribute.cpp

dseyler · 2025-05-30T23:19:49Z

@ktbolt As we discussed last week, it seems like this branch might need some pretty big structural revisions. I'm pretty busy for the next few weeks getting ready for SB3C, so will wait until after that to start working on this again. Let's discuss this as a group sometime after June 26 to figure out a plan for what changes would be most important to address.

@aabrown100-git I know you mentioned you might need to use the capping feature, so in the meantime, feel free to use my branch linked in the issue for this feature. Let me know if that works!

ktbolt · 2025-05-30T23:47:46Z

@dseyler @aabrown100-git I would say just merge the code and refactor later but this PR has failing CI tests.

It is not good to leave a PR open for so long and your branch is out-of-date.

ktbolt · 2025-07-29T21:18:00Z

@dseyler What's the status of this PR ?

dseyler · 2025-07-29T22:09:07Z

@ktbolt Currently addressing the smaller comments like decluttering the comments and renaming variables to be more descriptive. I also realized this feature broke something related to DomainID's so I'm working on fixing that too. If it's ok with you, I'll save for later any major refactoring like making virtual faces their own class.

…sed vtk files to .gitignore

Code/Source/solver/all_fun.cpp

Code/Source/solver/set_bc.cpp

ktbolt · 2025-10-02T16:23:19Z

@dseyler @aabrown100-git @dcodoni I am wondering if all of these changes to lots of code are going to destabilize the solver.

Maybe it would be better to close this PR and reimplement this functionality with an object-oriented design that isolates its connection with the rest of the code.

dseyler · 2025-10-02T16:46:55Z

@dseyler @aabrown100-git @dcodoni I am wondering if all of these changes to lots of code are going to destabilize the solver.

Maybe it would be better to close this PR and reimplement this functionality with an object-oriented design that isolates its connection with the rest of the code.

That may be a good idea. Current status: Since the connectivity array can contain -1 for cap faces, this causes indexing errors throughout the code. I added checks in ~10 places to try to avoid indexing out of bounds, but it broke several test cases. Was there previously any reason why the connectivity array would contain -1?

dcodoni · 2025-10-02T17:14:14Z

@ktbolt many changes can break the solver, especially when they involve connectivities, matrix assembly, rhs assembly and boundary conditions. I think maybe it would be better should rethink this implementation that may have reached a point where is difficult to debug

ktbolt · 2025-10-02T17:24:14Z

@dseyler The connectivity array containing -1 was a major source of pain for me when converting the code. I'm not sure what -1 is for.

aabrown100-git · 2025-10-02T20:37:31Z

@dseyler I thought this feature was working a few months ago, you just wanted to clean up some things. Is it totally broken now?

I'm a fan of reimplementing this in a more object-oriented fashion. But it would be nice to have this feature available sooner (for my own research needs). Is it possible to get this feature working again and just keep it on your fork? Then you can work on refactoring it into a class?

For the refactoring, maybe you can work with @zasexton to convert the faceType class into a new Face class, and have this capping be a feature or a derived class of Face.

…functions

zasexton · 2025-10-03T16:39:16Z

Yes, we can meet to talk about some potential templates for a Face class. I have availability later next week @dseyler @aabrown100-git if y'all also have any free time.

dseyler · 2025-10-03T17:29:16Z

@dseyler I thought this feature was working a few months ago, you just wanted to clean up some things. Is it totally broken now?

I'm a fan of reimplementing this in a more object-oriented fashion. But it would be nice to have this feature available sooner (for my own research needs). Is it possible to get this feature working again and just keep it on your fork? Then you can work on refactoring it into a class?

For the refactoring, maybe you can work with @zasexton to convert the faceType class into a new Face class, and have this capping be a feature or a derived class of Face.

It was working a few months ago and then broke when I merged the most recent version of svMultiphysics with it. I believe the latest commit should now be working both for 3D-0D Coupling and for the other test cases. @aabrown100-git give it a try and let me know if it works. Admittedly, the implementation is still a little messy and might be confusing for users.

@zasexton Let's meet sometime soon to discuss a cleaner way to implement it. Next week should also work for me.

dseyler · 2025-10-04T01:12:02Z

Trying to figure out how cap surfaces should interface with svZeroD. is it possible to have multiple surfaces coupled to one svZeroD block? Or should I create a 2nd block in the json file for the cap surface? @ktbolt @ncdorn. Both would require a few changes to the svZeroD subroutines

…Solver (incomplete)

dseyler · 2025-10-07T00:56:23Z

I believe this branch now works for 3D-0D coupling cap surfaces with GenBC as well as all current test cases. cap faces are still not supported for svZeroDSolver. After talking to @aabrown100-git, it may make sense to wait until after cap surfaces have been refactored as their own class to tackle svZeroDSolver.

@ktbolt Should we freeze this branch for internal lab use, and wait until later to reopen a pull request for the refactored version? Let me know your thoughts.

dseyler · 2025-10-09T21:35:08Z

Discussed with @zasexton and @ncdorn today about a proposal for a new cap coupling implementations that isolates changes from the rest of the code. We decided that creating a new CapFace class, where a faceType Object can optionally contain a CapFace object rather than treating cap faces as regular faces with a cap flag. Only one bcType object would then need to be created for the face/cap complex.

This proposal has the advantages:

No boolean checks and -1 sentinel values for cap faces cluttering the code
Cap faces cannot be created without a corresponding regular face
Cap processing is isolated from regular face processing
No duplicate BCs for cap faces (unintuitive and difficult to interface with svZeroDSolver)

We also discussed coming up with a better name for this feature besides "cap surface" since "cap surface" could also refer to the end cap on a vessel lumen. Current idea is "pseudo-surface" but I'm open to other suggestions.

@ktbolt I know this discussion is getting quite long for a pull request so can move to https://github.com/SimVascular/svMultiPhysics/issues/73 if that would be a better place to post.

ktbolt · 2025-10-14T17:34:29Z

@dseyler Please move the new cap coupling implementation ideas to the Issue and close this PR.

dseyler and others added 16 commits November 6, 2024 14:30

Updated fils_struct.hpp to include a flag for

262c857

virtual surfaces. Updated add_bc_mul.cpp to include capping surfaces in the flowrate integral but not in the pressure integral.

Added additional member variables associated with virtual faces

338486e

Updated fsils_bc_create to support virtual surface

dea38e0

Bug Fixes (NOT COMPLETE)

5fbf4fb

Updated several files to accomodate virtual surfaces

73e71f6

Added new test case for virtual surface cap

dc877cd

Updated README.md for LV_NeoHookean_passive_sv0D_capped

7c4ab5f

Merge branch '3D0DCouplingCap'

59ed2da

Merge pull request #1 from dseyler/main

31af55e

Merge main branch with 3D0DCouplingCap

Added a few more functions for virtual capping

168a140

Fixed compilation errors in the code

fca7f47

Co-authored-by: Aaron Brown <[email protected]>

2665879

Merge branch 'SimVascular:main' into 3D0DCouplingCap

e0340ac

Merge branch 'SimVascular:main' into 3D0DCouplingCap

c7797cd

Merge branch 'SimVascular:main' into 3D0DCouplingCap

91802e1

Removed debugging print statements from add_bc_mul.cpp

f944e98

aabrown100-git requested a review from ktbolt May 19, 2025 22:25

ktbolt reviewed May 20, 2025

View reviewed changes

dseyler requested review from aabrown100-git, Copilot and dcodoni May 20, 2025 20:43

Copilot AI reviewed May 20, 2025

View reviewed changes

Code/Source/solver/distribute.cpp Show resolved Hide resolved

Code/Source/solver/all_fun.cpp Outdated Show resolved Hide resolved

Code/Source/solver/all_fun.cpp Show resolved Hide resolved

Code/Source/solver/distribute.cpp Outdated Show resolved Hide resolved

Merge branch 'SimVascular:main' into 3D0DCouplingCap

55af584

dseyler and others added 3 commits August 22, 2025 09:48

Merge main into 3D0DCouplingCap - resolved conflict in ComMod.cpp

7945d6e

Merge branch 'SimVascular:main' into 3D0DCouplingCap

2bfc8b7

merged main with this branch

944b0c8

merged main into 3D0D Coupling branch and added some test case proces…

fa7f300

…sed vtk files to .gitignore

ktbolt reviewed Sep 29, 2025

View reviewed changes

Code/Source/solver/all_fun.cpp Show resolved Hide resolved

ktbolt reviewed Sep 29, 2025

View reviewed changes

Code/Source/solver/set_bc.cpp Show resolved Hide resolved

Fixed several bugs, decluttered commenting, and renamed GatherMaster …

4be201c

…functions

dseyler added 2 commits October 6, 2025 17:29

Changed "virtual surface" to "cap surface." Added support for svZeroD…

276749c

…Solver (incomplete)

Fixed compilation error

18c5611

Merge branch 'SimVascular:main' into 3D0DCouplingCap

5b06340

dseyler closed this Oct 14, 2025

3 d0 d coupling cap #381

3 d0 d coupling cap #381

Uh oh!

Conversation

dseyler commented May 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Current situation

Release Notes

Documentation

Testing

Code of Conduct & Contributing Guidelines

Uh oh!

ktbolt left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

ktbolt May 20, 2025

Choose a reason for hiding this comment

Uh oh!

ktbolt May 20, 2025

Choose a reason for hiding this comment

Uh oh!

ktbolt May 20, 2025

Choose a reason for hiding this comment

Uh oh!

dseyler May 20, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

dseyler commented May 30, 2025

Uh oh!

ktbolt commented May 30, 2025

Uh oh!

ktbolt commented Jul 29, 2025

Uh oh!

dseyler commented Jul 29, 2025

Uh oh!

Uh oh!

Uh oh!

ktbolt commented Oct 2, 2025

Uh oh!

dseyler commented Oct 2, 2025

Uh oh!

dcodoni commented Oct 2, 2025

Uh oh!

ktbolt commented Oct 2, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

aabrown100-git commented Oct 2, 2025

Uh oh!

zasexton commented Oct 3, 2025

Uh oh!

dseyler commented Oct 3, 2025

Uh oh!

dseyler commented Oct 4, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

dseyler commented Oct 7, 2025

Uh oh!

dseyler commented Oct 9, 2025

Uh oh!

ktbolt commented Oct 14, 2025

Uh oh!

Reviewers

Assignees

Labels

dseyler commented May 19, 2025 •

edited

Loading

ktbolt commented Oct 2, 2025 •

edited

Loading

dseyler commented Oct 4, 2025 •

edited

Loading