Errors in Fluid-Structure Interaction Analysis Using HyperMeshCFD (2024) and Optistruct (2024)

Daiki_Konosu_0622

Hello everyone,

I am working on a fluid-structure interaction analysis of pressure wave propagation using HyperMeshCFD and Optistruct.

When I ran the calculation, The following error statement appeared in the computer console of the fem file. (also described in png file)

"

StdErr output from the solver:

External Code via CCI: *** ASSERTION in Function <iopWriteSock> File <iopSocket.c> Line <612>
External Code via CCI: *** TCP socket write error; No error
Abort(-5) on node 0 (rank 0 in comm 0): application called MPI_Abort(MPI_COMM_WORLD, -5) - process 0

==============================

hwsolver.mesg output from the solver:

**** ABORTING RUN DUE TO UNEXPECTED ERROR CONDITION ****

Timeout ALARM
Timeout ALARM

==============================

Solver run completed.

OptiStruct:: Solver run finished after 4 minutes.

"

picture of error statement.png

This error statement makes it difficult to determine the cause of the problem. We believe that the calculation itself has probably converged and is not an error based on large deformation.

Also, one of my biggest questions is that in the analysis using the 2022.3 version of Acusolve and Optistruct, the calculations completed without error. (exactly the same calculation conditions as the 2024 version)

I also attach my HyperMeshCFD inp, Log files, and Optistruct fem, out files. I would appreciate any suggestions for a solution to this problem.

Thanks in advance. 

0425_str_2-0mm_2024.fem

0425_str_2-0mm_2024.out

0428_flu_2-0mm_2024.1.Log

0428_flu_2-0mm_2024.inp

acupro

The AcuSolve .Log file shows severe mesh distortion errors at the end of the first time step - then stops.

I see you have a multiplier function on the External Code command - with a constant value of 1. Maybe try making that a linear type - where it starts at 0 then goes to 1 over the first 20 time steps - to slowly pass the fluid forces, letting the flow converge better.

Daiki_Konosu_0622

Thank you Acupro for your reply.

As a result, I set the Multiplier Function to linear, but I got the same error statement. I calculated the coefficient by reducing it from 1 to 0.1, etc., but got the same error statement.

I also thought that the time-step might be too large, so I reduced the time increment, but the same error statement appeared.

The error statement

"External Code via CCI: *** ASSERTION in Function <iopWriteSock> File <iopSocket.c> Line <612> External Code via CCI: *** TCP socket write error; No error Abort
(-5) on node 0 (rank 0 in comm 0): application called MPI_Abort(MPI_COMM_WORLD, -5) - process 0"

means that it is due to a severe mesh distortion error. Is this correct?

Thanks again for your comment.

acupro

If that error statement is in the AcuSolve .Log file - that likely means OptiStruct stopped for whatever reason, and closed the connection. If you still have the mesh distortion error and stop on the AcuSolve side - then that TCP socket error would be on the OptiStruct side.

Have you verified the fluid simulation runs properly on its own?

I do find it interesting that the dimensions of the structure vs fluid for the wetted/exchange surfaces are slightly different (larger on the structural side), but the area is about the same:

acuSolve: Min/Max x-coord = -1.525000e-02 1.525000e-02
acuSolve: Min/Max y-coord = -1.525000e-02 1.525000e-02
acuSolve: Min/Max z-coord = -1.450000e-16 2.000000e+00
acuSolve: Total surface area = 1.822637e-01
acuSolve: Total length = 0.000000e+00
acuSolve: No. external code mbds = 0
acuSolve: No. fluid wet surfaces = 1057
acuSolve: Min/Max x-coord = -1.500000e-02 1.500000e-02
acuSolve: Min/Max y-coord = -1.499162e-02 1.499162e-02
acuSolve: Min/Max z-coord = 0.000000e+00 2.000000e+00
acuSolve: Total fluid surface area = 1.822672e-01

What sort of simulation are you running? What is your simulation model?

Daiki_Konosu_0622

Thank you for explaining how the error statement works.

Afterwards, I continued the verification and successfully simulated the calculation. I took your suggestion and changed the mesh size in various ways, thinking it was due to the “shape of the exchange surface”. As a result, the mesh size of the fluid elements was smaller than that of the structural elements, allowing the calculations to run to the end.

We are currently investigating the detailed cause of the problem, but would like to perform simulations in the future, taking the mesh size into account.

Again, thank you for your reply. We look forward to working with you in the future.