Failure: FSI/flap_perp_2D/OpenFOAM-FEniCS

[BenjaminRodenberg]

I just ran into an error, when trying to run FSI/flap_perp/OpenFOAM-FEniCS using the current state of develop (b74fdffb8062d6fd49733244be0d918643ec543c). Everything works for 13f101bb77461ea3f31451e3d22e6d72320c9f3e (before merging precice/fenics-adapter#71).

I used the following revision of the tutorial: ac6a94c (latest commit of #124). Observed the same behavior for 1c891a7 (latest commit of https://github.com/precice/tutorials/tree/develop).

[IshaanDesai]

The error in the FSI/flap_perp/OpenFOAM-FEniCS tutorial when run with the following commits:

• FEniCS-Adapter: precice/fenics-adapter@b74fdff
• OpenFOAM-Adapter (with OpenFOAM 7): precice/openfoam-adapter@30b7af9
• Tutorials repository: 1c891a7

is as follows:

---[precice]  it 4 of 50 | dt# 45 | t 0.44 of 5 | dt 0.01 | max dt 0.01 | ongoing yes | dt complete no | read-iteration-checkpoint | 
---[preciceAdapter] [DEBUG] Reading a checkpoint...
---[preciceAdapter] [DEBUG] Reloaded time value t = 0.440000
---[preciceAdapter] [DEBUG] Moving mesh points to their previous locations...
---[preciceAdapter] [DEBUG] Moved mesh points to their previous locations.
---[preciceAdapter] [DEBUG] Reading a mesh checkpoint...
---[preciceAdapter] [DEBUG] Mesh checkpoint was read. Time = 0.440000
---[preciceAdapter] [DEBUG] Checkpoint was read. Time = 0.440000
---[preciceAdapter] [DEBUG] Reading coupling data...
---[preciceAdapter] [DEBUG] Adjusting the solver's timestep...
---[preciceAdapter] [DEBUG] The solver's timestep is the same as the coupling timestep.
Courant Number mean: 1.4508 max: 1371.95
Time = 0.45

smoothSolver:  Solving for cellDisplacementx, Initial residual = 1, Final residual = 9.99951e-07, No Iterations 631
smoothSolver:  Solving for cellDisplacementy, Initial residual = 1, Final residual = 9.98908e-07, No Iterations 458
#0  Foam::error::printStack(Foam::Ostream&) at ??:?
precice/fenics-adapter#1  Foam::sigFpe::sigHandler(int) at ??:?
precice/fenics-adapter#2  ? in "/lib/x86_64-linux-gnu/libc.so.6"
precice/fenics-adapter#3  Foam::DICPreconditioner::calcReciprocalD(Foam::Field<double>&, Foam::lduMatrix const&) at ??:?
precice/fenics-adapter#4  Foam::DICPreconditioner::DICPreconditioner(Foam::lduMatrix::solver const&, Foam::dictionary const&) at ??:?
precice/fenics-adapter#5  Foam::lduMatrix::preconditioner::addsymMatrixConstructorToTable<Foam::DICPreconditioner>::New(Foam::lduMatrix::solver const&, Foam::dictionary const&) at ??:?
precice/fenics-adapter#6  Foam::lduMatrix::preconditioner::New(Foam::lduMatrix::solver const&, Foam::dictionary const&) at ??:?
precice/fenics-adapter#7  Foam::PCG::solve(Foam::Field<double>&, Foam::Field<double> const&, unsigned char) const at ??:?
precice/fenics-adapter#8  Foam::fvMatrix<double>::solveSegregated(Foam::dictionary const&) at ??:?
precice/fenics-adapter#9  Foam::fvMatrix<double>::solve(Foam::dictionary const&) in "/opt/openfoam7/platforms/linux64GccDPInt32Opt/bin/pimpleFoam"
precice/fenics-adapter#10  Foam::fvMatrix<double>::solve() in "/opt/openfoam7/platforms/linux64GccDPInt32Opt/bin/pimpleFoam"
precice/fenics-adapter#11  ? in "/opt/openfoam7/platforms/linux64GccDPInt32Opt/bin/pimpleFoam"
precice/fenics-adapter#12  ? in "/opt/openfoam7/platforms/linux64GccDPInt32Opt/bin/pimpleFoam"
precice/fenics-adapter#13  __libc_start_main in "/lib/x86_64-linux-gnu/libc.so.6"
precice/fenics-adapter#14  ? in "/opt/openfoam7/platforms/linux64GccDPInt32Opt/bin/pimpleFoam"
./runFluid: line 44: 32037 Floating point exception(core dumped) $solver -case Fluid
Looking for any time directories without results (e.g. stray functionObjectProperties files, see issue precice/fenics-adapter#26 on GitHub)...

This error occurs on the OpenFOAM side and is not related to any new functionality added in the FEniCS-Adapter.
Keeping this issue open for future resolution.

[uekerman]

Why was this issue closed? Did we resolve it or open it somewhere else? In the latter case, please link.

[IshaanDesai]

@MakisH need your perspective on this error. The conclusion is that nothing is wrong on the FEniCS side. Is this a reasonable assumption?

[MakisH]

I see that it does solve a few iterations. How do the intermediate results look like visually? I guess that it either computes wrong forces or it receives wrong displacements (if the rest is the same as in our other tutorials).

@IshaanDesai would it also be easy for you to try with OpenFOAM v2006 and the adapter develop and document if you get the same error? Do you get any error with OpenFOAM 7 and any other FSI tutorial?

[IshaanDesai]

I see that it does solve a few iterations. How do the intermediate results look like visually?

This is the fluid domain just before the crash, it looks reasonable.

@IshaanDesai would it also be easy for you to try with OpenFOAM v2006 and the adapter develop and document if you get the same error? Do you get any error with OpenFOAM 7 and any other FSI tutorial?

I checked with OpenFOAM v2006 and the latest develop branch of the adapter and sadly the error is the same and it occurs slightly sooner than the error with OpenFOAM 7. The error log is similar:

---[precice]  it 2 of 50 | dt# 43 | t 0.42 of 5 | dt 0.01 | max dt 0.01 | ongoing yes | dt complete no | read-iteration-checkpoint | 
---[preciceAdapter] [DEBUG] Reading a checkpoint...
---[preciceAdapter] [DEBUG] Reloaded time value t = 0.420000
---[preciceAdapter] [DEBUG] Moving mesh points to their previous locations...
---[preciceAdapter] [DEBUG] Moved mesh points to their previous locations.
---[preciceAdapter] [DEBUG] Reading a mesh checkpoint...
---[preciceAdapter] [DEBUG] Mesh checkpoint was read. Time = 0.420000
---[preciceAdapter] [DEBUG] Checkpoint was read. Time = 0.420000
---[preciceAdapter] [DEBUG] Reading coupling data...
---[preciceAdapter] [DEBUG] Adjusting the solver's timestep...
---[preciceAdapter] [DEBUG] The solver's timestep is the same as the coupling timestep.
Courant Number mean: 1.42724 max: 16.6107
Time = 0.43

PIMPLE: iteration 1
smoothSolver:  Solving for cellDisplacementx, Initial residual = 0.998497, Final residual = 1.4967e-18, No Iterations 2
smoothSolver:  Solving for cellDisplacementy, Initial residual = 0.998802, Final residual = 8.10127e-18, No Iterations 2
DICPCG:  Solving for pcorr, Initial residual = 1, Final residual = 0.000880973, No Iterations 69
DICPCG:  Solving for pcorr, Initial residual = 0.515516, Final residual = 6.5027e-09, No Iterations 96
time step continuity errors : sum local = -8.16676e-08, global = -9.24857e-09, cumulative = 0.022902
#0  Foam::error::printStack(Foam::Ostream&) at ??:?
#1  Foam::sigFpe::sigHandler(int) at ??:?
#2  ? in /lib/x86_64-linux-gnu/libpthread.so.0
#3  Foam::smoothSolver::solve(Foam::Field<double>&, Foam::Field<double> const&, unsigned char) const at ??:?
#4  Foam::fvMatrix<Foam::Vector<double> >::solveSegregated(Foam::dictionary const&) at ??:?
#5  Foam::fvMatrix<Foam::Vector<double> >::solveSegregatedOrCoupled(Foam::dictionary const&) at ??:?
#6  Foam::fvMesh::solve(Foam::fvMatrix<Foam::Vector<double> >&, Foam::dictionary const&) const at ??:?
#7  ? in /usr/lib/openfoam/openfoam2006/platforms/linux64GccDPInt32Opt/bin/pimpleFoam
#8  __libc_start_main in /lib/x86_64-linux-gnu/libc.so.6
#9  ? in /usr/lib/openfoam/openfoam2006/platforms/linux64GccDPInt32Opt/bin/pimpleFoam
./runFluid: line 44:  6207 Floating point exception(core dumped) $solver -case Fluid
Looking for any time directories without results (e.g. stray functionObjectProperties files, see issue #26 on GitHub)...

[IshaanDesai]

The above mentioned error does not occur when the case is run with a previous version of the fenics-adapter: precice/fenics-adapter@d551762.
This clearly means something went wrong in the redesigning of the FEniCS-Adapter: precice/fenics-adapter#71