[habaw]

Hello

I have been utilizing the solids4Foam toolbox to simulate turbulent flow(Water) through a 2D channel with flexible obstructions using RANS model Kepsilon and sstKomega. Specifically, I have successfully loaded solids4Foam version 2 in conjunction with OpenFOAM/9

I am using the perpendicular flap as an example to setup my case. the velocity is 1.5m/s with Re 42000 the channel has a height of 5cm and a square obstruction(fixed at the bottom of the wall) 2.5cm is located 37.5cm downstream of the inlet and 125cm u/s of the outlet. Fully developed flow at the inlet.

My first issue is How can i setup the mechanical property of the interface I was using rubber type linearElastic; so that I can visualize a deflection because when ever I use a smaller E the simulation crashes with floatingpointexception I see that its because the deflection is very big and the K values become very big.

I need to know how I can set up the mechanical property properly.

Thank you

[bigphil]

What do you mean by "how I can set up the mechanical property properly"?

This depends on what solid material you want to simulate. For example, if you are simulating aluminium and if it is expected to stay in the elastic regime, then you would look up the Young's modulus (E) and Poisson's ratio (nu) for your particular aluminium alloy.

[habaw]

Hello
Thank you for the replay

What I mean is I set up my case and my simulation works when I use a rigid material with higher modulus of elasticity there is no visible deflection. Whenever I reduce the elasticity the simulation fails with following error and the simulation stops after running for some time . All the values diverge (logsolids4Foam)
What could be the reason for this?


ERROR:
[0] #0 Foam::error:rintStack(Foam::Ostream&) at ??:?
[0] #1 Foam::sigFpe::sigHandler(int) at ??:?
[0] #2 ? in "/cvmfs/soft.computecanada.ca/gentoo/2020/lib64/libc.so.6"
[0] #3 Foam::multiply(Foam::Field<double>&, Foam::UList<double> const&, Foam::UList<double> const&) at ??:?
[0] #4 Foam::tmp<Foam:imensionedField<double, Foam::volMesh> > Foam:perator*<Foam::volMesh>(Foam::tmp<Foam:im ensionedField<double, Foam::volMesh> > const&, Foam:imensionedField<double, Foam::volMesh> const&) at ??:?
[0] #5 Foam::RASModels::kEpsilon<Foam::IncompressibleMome ntumTransportModel<Foam::kinematicTransportModel> >::correct() at ??:?
[0] #6 Foam::fluidModels:impleFluid::evolve() at ??:?
[0] #7 Foam::fluidSolidInterfaces::AitkenCouplingInterfac e::evolve() at ??:?
[0] #8 ? at ??:?
[0] #9 __libc_start_main in "/cvmfs/soft.computecanada.ca/gentoo/2020/lib64/libc.so.6"
[0] #10 ? at ??:?
[gra136:10338:0:10338] Caught signal 8 (Floating point exception: tkill(2) or tgkill(2))
==== backtrace (tid: 10338) ====
0 0x0000000000020753 ucs_debug_print_backtrace() /tmp/ebuser/avx2/UCX/1.8.0/GCCcore-9.3.0/ucx-1.8.0/src/ucs/debug/debug.c:653
1 0x0000000000038980 killpg() ???:0
2 0x0000000000038901 __libc_signal_restore_set() /cvmfs/soft.computecanada.ca/gentoo/2020/usr/src/debug/sys-libs/glibc-2.30-r8/glibc-2.30/signal/../sysdeps/unix/sysv/linux/internal-signals.h:84
3 0x0000000000038980 killpg() ???:0
4 0x000000000061acce Foam::multiply() ???:0
5 0x00000000003990b0 Foam:perator*<Foam::volMesh>() ???:0
6 0x00000000001b6efe Foam::RASModels::kEpsilon<Foam::IncompressibleMome ntumTransportModel<Foam::kinematicTransportModel> >::correct() ???:0
7 0x00000000003d04cc Foam::fluidModels:impleFluid::evolve() ???:0
8 0x00000000003f442d Foam::fluidSolidInterfaces::AitkenCouplingInterfac e::evolve() ???:0
9 0x00000000004025db main() ???:0
10 0x0000000000023e1b __libc_start_main() /cvmfs/soft.computecanada.ca/gentoo/2020/usr/src/debug/sys-libs/glibc-2.30-r8/glibc-2.30/csu/../csu/libc-start.c:308
11 0x0000000000402b1a _start() ???:0

[bigphil]

For this problem, if the stiffness of the solid is too low, the solver may fail for multiple reasons, such as, the fluid mesh becomes excessively distorted, the solid solver may struggle to converge, the fluid-solid coupling procedure may struggle to converge.

[habaw]

Thank you for your prompt replay


I have noticed that my turbulent case with RANS model runs perfectly with openFoam only and also it runs in solids4Foam whenever the interface/flap/ is rigid. Even the slightest flexibility on the flap crushes the case file.
Also I tried with the tutorials for perpendicular flap by changing the turbulence property into RANS and simulation diverges after some iteration.

Could I get any insight on this issue please.
Thank you

[be_inspired]

Hi,

Could it be possible to impose the deformation of the surface giving a displacement field? That is, use the same inputs as for the laplacian solver.

I want to impose the deformation of the flap along the flap length.

Best Regards
Marcelino

[bigphil]

If the displacement field is known, then I guess it would not be an FSI problem. In that case, you can take a look at the various mesh motion solvers and their boundary conditions.

[be_inspired]

I want to use RBF solver because I am managing large deformations. Laplacian approach is not working properly in those cases after several movements and slerp is only available for solid body movements.

I have compiled RBFMeshMotionSolver ( available in solids4foam) into Openfoam v2210 and v2012. For a mesh of 2 M cells, in serial it does nothing after 1 hour ( no errors, nothing, only taking a lot of memory). In parallel, it fails because of an error.

terminate called after throwing an instance of 'std::bad_alloc'
terminate called after throwing an instance of 'std::bad_alloc'
what(): std::bad_alloc
[node021:15401] *** Process received signal ***
terminate called after throwing an instance of 'std::bad_alloc'
what(): std::bad_alloc
[node021:15402] *** Process received signal ***
what(): std::bad_alloc
[node021:15404] *** Process received signal ***
[node021:15404] Signal: Aborted (6)
[node021:15404] Signal code: (-6)
[node021:15401] Signal: Aborted (6)
[node021:15401] Signal code: (-6)
[node021:15402] Signal: Aborted (6)
[node021:15402] Signal code: (-6)

[habaw]

Is it possible to conducted a 3D Solids4Foam case in LES in parallel. But It works perfectly fine in series but not parallel. A mesh tolerance error occurs.
Please need you help and Thank you

[bigphil]

Are you using OpenFOAM-9? If so, you could try OpenFOAM-2012 (or newer) and check if that works.

[habaw]

Thank you for your reply
Yes I am using OF9 which I compiled on graham cluster computed Canada to run my case. I tried compiling using v2012 but was unable.
What change make parallel run work in 2012 but not OF9 if I may ask?

[habaw]

I tried the 2012 version but still was not able to run the case in parallel same as OF9. So I run my case in series first and used the latest time to run in parallel but the following error again occurs for every processor. Please help? Thank You
/Below is portion of the error message/



--> FOAM FATAL ERROR:
[2] face 36 area does not match neighbour by 190.244% -- possible face ordering problem.
patchrocBoundary2to1 my area:0.466667 neighbour area:0.0116666 matching tolerance:0.000308029
Mesh face:100349 vertices:4((3.67606e-11 0 0) (3.66314e-11 0.133333 0) (-3.5 0.266667 0) (-3.5 0.133333 0))
If you are certain your matching is correct you can increase the 'matchTolerance' setting in the patch dictionary in the boundary file.
Rerun with processor debug flag set for more information.
[2]
[2] From function virtual void Foam:rocessorPolyPatch::calcGeometry(Foam::Pstre amBuffers&)
[2] in file meshes/polyMesh/polyPatches/constraint/processor/processorPolyPatch.C at line 269.
[2]
FOAM parallel run exiting