Solids4foam-openfoam FSI coupling: can't use stress as the exchange data

zyx · October 5, 2025, 5:45am

Hello everyone！ I am using solids4foam and OpenFOAM to simulate an FSI problem. I want to use Stress as the exchange date instead of Force. However, solids4foam displayed error messages as：

[31mError (deferred - will exit later) in the preCICE adapter: e[0m Reading forces is not supported.

The following are my config files:
precice-config.xml (3.0 KB)
preciceDict-fluid.txt (1.1 KB)
preciceDict-solid.txt (1.0 KB)
Could you please help me check them for errors?
The following are the log files after running:
fluid.log (30.5 KB)
slurm-solid.out (4.8 KB)
solid.log (13.8 KB)

Makis · October 8, 2025, 9:11am

Reading forces is supported since v1.2.0 of the adapter, and you do run v1.2.1 (which is good).

Have a look at our tutorial case with solids4foam:

github.com/precice/tutorials

perpendicular-flap/solid-solids4foam/system/preciceDict

develop

FoamFile
{
    version     2.0;
    format      ascii;
    class       dictionary;
    location    "system";
    object      preciceDict;
}

preciceConfig "../precice-config.xml";

participant Solid;

modules (FSI);

interfaces
{
    Interface1
    {
        mesh              Solid-Mesh;

This file has been truncated. show original

The error message is indeed confusing and we should refine it with suggestions:

github.com/precice/openfoam-adapter

FSI/ForceBase.C

v1.2.1


      
          }
          
          void preciceAdapter::FSI::ForceBase::readFromBuffer(double* buffer) const
          {
              /* TODO: Implement
               * We need two nested for-loops for each patch,
               * the outer for the locations and the inner for the dimensions.
               * See the preCICE readBlockVectorData() implementation.
               */
              FatalErrorInFunction
                  << "Reading forces is not supported."
                  << exit(FatalError);
          }

It is also a bit strange that this is triggered (I need to look deeper into it). The right code should be:

github.com/precice/openfoam-adapter

FSI/Force.C

v1.2.1


      
          
                  Force_ = ForceOwning_.get();
              }
          }
          
          void preciceAdapter::FSI::Force::write(double* buffer, bool meshConnectivity, const unsigned int dim)
          {
              this->writeToBuffer(buffer, *Force_, dim);
          }
          
          void preciceAdapter::FSI::Force::read(double* buffer, const unsigned int dim)
          {
              // Copy the force field from the buffer to OpenFOAM
          
              // Here we assume that a force volVectorField exists, which is used by
              // the OpenFOAM solver
          
              // Set boundary forces
              for (unsigned int j = 0; j < patchIDs_.size(); j++)
              {
                  // Get the ID of the current patch

Could you tell me how to reproduce the issue starting from the tutorial?

However, your configs (preciceDict) show stresses, not forces. Could you please check again?

There is a limitation that only solid participants can read Force and Stress, and it is important to use the exact words (or start with them).

zyx · October 11, 2025, 3:05am

Sorry for being late to reply!
I used a tutorial case of “perpendicular-flap” to reproduce the issue. I changed the Force to Stress in the config files. The error also displayed as “Reading forces is not supported.”
The following are the config files:
precice-config.xml (2.8 KB)
preciceDict-fluid.txt (661 Bytes)
preciceDict-solid.txt (942 Bytes)
The following are the output log files:
fluid.log (6.4 KB)
solid.log (5.0 KB)
The following is the modfiled perpendicular-flap case:
flap-stress.zip (113.0 KB)
Could you please check where they are set incorrectly? Thank you very much!

Makis · October 13, 2025, 9:22am

Now I see what you mean and what the issue is. Yes, reading stresses is not yet supported, but should be easy to add:

github.com/precice/openfoam-adapter

Reading stress from solid solvers

opened 05:19PM - 08 Aug 22 UTC

MakisH

enhancement good first issue FSI

https://github.com/precice/openfoam-adapter/pull/236 added the possibility to re…ad forces, effectively enabling structure solvers (such as solidDisplacementFoam or solids4Foam) to use the adapter. It would still be nice if the adapter could also read stresses (`Stress`). This should hopefully not be so complicated to implement, assuming it is relatively similar to the writing part.

See the list of supported read/write fields:

The error is indeed misleading, as you are asking for stresses, but get an error about forces. Assume “stresses” in the error message (I fixed the error message in Fix error message for reading stresses by MakisH · Pull Request #379 · precice/openfoam-adapter · GitHub).

zyx · October 15, 2025, 12:29am

Thank you for your reply!
I am wondering if solids4foam can only write the date “displacement” and read data “Force”? So far, it doesn’t support exchanging other types of dates, such as reading “stress”, writing “displacementDelta”, or writing “velocity”, right?

zyx · October 15, 2025, 3:13am

Could I ask some more questions about exchanging data in an FSI simulation? Question 1:OpenFOAM fluid simulation requires ‘velocity’ data on the interface boundary, but solids4faom solid simulation can only provide ‘displacement’. Does Precice have the function to calculate point velocity using the displacement on the interface boundary and then provide the velocity to OpenFOAM?
Question 2:Furthermore, as I understand it, fluid velocity should equal solid velocity at the interface boundary. In the tutorial, many cases only exchange displacement data. Why can velocity be ignored? Could you share your thoughts on this?
Question 3:Also, Precice exchanges displacement and force, which is the Dirichlet–Neumann boundary condition. How is the Robin-Neumann boundary condition used, where fluid takes a combination of kinematic and dynamic interface conditions as the boundary condition? How can this be achieved?
I would be very grateful if you could clear up my confusion!

Makis · October 30, 2025, 3:20pm

No, the adapter does not handle this explicitly.

Since the velocity depends directly on the displacement (derivative), we don’t impose it. One could extend the adapter to exchange the velocity of the interface as well (in fact, we had this in first implementations), but so far we had several successful applications exchanging only displacement and force/stress.

We don’t support such coupling for FSI in the OpenFOAM adapter at the moment, but one could extend it. In terms of preCICE, this is possible, and we already have examples of Robin coupling in CHT applications (see heat exchanger).

zyx · October 31, 2025, 12:47am

Thank you very much for your response, which solves my question.

system · November 3, 2025, 12:47am

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.

Topic		Replies	Views
Openfoam-Solids4foam FSI: how to set Stress as exchange data instead of Force Using preCICE openfoam , fsi	2	72	August 6, 2024
OpenFOAM-OpenFOAM FSI coupling Official adapters and tutorials openfoam , adapters , configuration , fsi , inactive	51	8494	February 22, 2021
How to transfer Pressure instead of Force in FSI? Using preCICE openfoam , adapters , fsi	12	990	August 24, 2021
A few questions about coupling OpenFOAM with in-house solver Using preCICE openfoam , fsi , multiphase	5	661	April 18, 2021
ERROR when reading "Velocity" in FSI Using preCICE openfoam , adapters , fsi	12	436	September 17, 2021

Solids4foam-openfoam FSI coupling: can't use stress as the exchange data

Related topics