Problems running FSI heat transfer analysis with AcuSolve and OptiStruct

Lucia R

Hi,

I am doing a heat transfer FSI analysis with Acusolve and OptiStruct and it is running but the results are wrong because it obtains temperatures about 3e135 K. 

I did the transient heat transfer analysis in Acusolve alone and it works correctly, and in Optistruct too, but when I conect both of them the results seem crazy. 

Apart from this, the FSI is running from 3 days ago and it is in the time step 23 (the CFD analysis alone was completed in 6 hours).

Can anyone help me with this problem? I put below the files .log and .out of this problem.

Thanks in advance!

 

 

acupro

Lucia R said:

I don't have the 2019.0 version.

 

I'm going to share with you the model files, can you try to run them in the 2019.0 version to check that it works?

 

Thanks!

 

The files are included in the following Drive link:

https://drive.google.com/file/d/1hPyhdTMZRfYPj-yXuQBo_IjMedaOiOvn/view?usp=sharing

 

The coupling for this does, indeed, work with 2019.0 of both solvers.  I did make a few changes to the AcuSolve .inp file.  I reduced min stagger iterations to 2 - as there's really no need to force a minimum of 4 - it will do as many as needed.  You don't need both Simple BC and External CS for the same surface - so I commented Simple BC.  I modified the pressure initial condition.

I've attached the updated .inp file.

The 2019.0 release should still be available through Altair Connect.  Make sure you check the box for 'Include Older Packages'.

I'm still hopeful this will all be corrected/fixed for the 2021.1 release packages.

acupro

Starting with line 532 of the .Log file, it describes the External (optiStruct) surface then the fluid (AcuSolve) surface:

acuSolve: Min/Max x-coord = 3.753500e+001 3.857000e+001
acuSolve: Min/Max y-coord = -3.319230e+001 3.317798e+001
acuSolve: Min/Max z-coord = -5.000000e+001 5.000000e+001
acuSolve: Total surface area = 5.280686e+003
acuSolve: Total length = 0.000000e+000
acuSolve: No. external code mbds = 0
acuSolve: No. fluid wet surfaces = 229566
acuSolve: Min/Max x-coord = 3.673500e+001 3.937000e+001
acuSolve: Min/Max y-coord = -3.299999e+001 3.299999e+001
acuSolve: Min/Max z-coord = -5.000000e+001 5.000000e+001
acuSolve: Total fluid surface area = 1.249814e+004

The min/max dimensions are slightly different, and the overall fluid surface area is more than twice that of the structural area.  For the coupled analysis to work properly, the extent (size, shape, location, units) of the exchange surface need to be the same in each solver.  (The mesh doesn't need to be the same.)

At the first time step, also from the .Log file:

acuSolve: External-Code stagger "external_code":
acuSolve: Ave distance = 3.836261e-001
acuSolve: Max distance = 8.000000e-001

This again indicates there is a mismatch between the surfaces in the two solvers - the average distance from the AcuSolve integration point to the OptiStruct integration point is 0.38, while the average is 0.8.  (Do we assume this is meters?)

Please double-check to ensure the size, shape, location, units are represented the same in both solvers.  You should also make sure the temperature 'base' is the same, for example, that a value of 273 means the same thing in both solvers.

 

Lucia R

Hi Acupro,

I have modified the exchange surface and now it seems to be good (units are mm):

 

acuSolve: Min/Max x-coord = 3.857000e+001 3.937000e+001
acuSolve: Min/Max y-coord = -3.000000e+001 3.050000e+001
acuSolve: Min/Max z-coord = -5.000000e+001 5.000000e+001
acuSolve: Total surface area = 6.077552e+003
acuSolve: Total length = 0.000000e+000
acuSolve: No. external code mbds = 0
acuSolve: No. fluid wet surfaces = 124896
acuSolve: Min/Max x-coord = 3.857000e+001 3.937000e+001
acuSolve: Min/Max y-coord = -3.000000e+001 3.050000e+001
acuSolve: Min/Max z-coord = -5.000000e+001 5.000000e+001
acuSolve: Total fluid surface area = 6.163713e+003

 

 

But in the .log the time steps still wrong, I think that the units would be different in both models, but I don't know how to check this and change it.

 

acuSolve: Time-Step= 2 ; timeInc= 1.000000e-002 ; time= 1.000000e-002
acuSolve: External-Code stagger "external_code":
acuSolve: Ave distance = 6.796082e-004
acuSolve: Max distance = 1.111094e-001
acuSolve: Min allowed gap = 6.130143e+074
acuSolve: Max allowed gap = 4.516963e+075
acuSolve: struct-heat res ratio = 3.732756e-003
acuSolve: Min struct. heat flux = -1.944140e+000
acuSolve: Ave struct. heat flux = -6.062449e-001
acuSolve: Max struct. heat flux = 6.415443e-001
acuSolve: Std struct. heat flux = 2.584384e-001
acuSolve: Min temperature = -1.290625e+004
acuSolve: Ave temperature = -3.878769e+003
acuSolve: Max temperature = 1.985994e+003
acuSolve: Std temperature = 2.307468e+003
acuSolve: Converged

 

How can I check the heat flux units in both models?

 

Thanks for your help.

 

acupro

Lucia R said:

Hi Acupro,

I have modified the exchange surface and now it seems to be good (units are mm):

 

acuSolve: Min/Max x-coord = 3.857000e+001 3.937000e+001
acuSolve: Min/Max y-coord = -3.000000e+001 3.050000e+001
acuSolve: Min/Max z-coord = -5.000000e+001 5.000000e+001
acuSolve: Total surface area = 6.077552e+003
acuSolve: Total length = 0.000000e+000
acuSolve: No. external code mbds = 0
acuSolve: No. fluid wet surfaces = 124896
acuSolve: Min/Max x-coord = 3.857000e+001 3.937000e+001
acuSolve: Min/Max y-coord = -3.000000e+001 3.050000e+001
acuSolve: Min/Max z-coord = -5.000000e+001 5.000000e+001
acuSolve: Total fluid surface area = 6.163713e+003

 

 

But in the .log the time steps still wrong, I think that the units would be different in both models, but I don't know how to check this and change it.

 

acuSolve: Time-Step= 2 ; timeInc= 1.000000e-002 ; time= 1.000000e-002
acuSolve: External-Code stagger "external_code":
acuSolve: Ave distance = 6.796082e-004
acuSolve: Max distance = 1.111094e-001
acuSolve: Min allowed gap = 6.130143e+074
acuSolve: Max allowed gap = 4.516963e+075
acuSolve: struct-heat res ratio = 3.732756e-003
acuSolve: Min struct. heat flux = -1.944140e+000
acuSolve: Ave struct. heat flux = -6.062449e-001
acuSolve: Max struct. heat flux = 6.415443e-001
acuSolve: Std struct. heat flux = 2.584384e-001
acuSolve: Min temperature = -1.290625e+004
acuSolve: Ave temperature = -3.878769e+003
acuSolve: Max temperature = 1.985994e+003
acuSolve: Std temperature = 2.307468e+003
acuSolve: Converged

 

How can I check the heat flux units in both models?

 

Thanks for your help.

 

I'm not sure about OptiStruct - but AcuSolve (the solver) is unit-less.  The solver has no knowledge of units - it only has values/numbers.  Thus, it is up to the user to ensure those values represent a consistent set of units.  If you dimensions/lengths are in mm, then all units BCs, etc need to have values consistent with the length dimension being mm.  So density has to be be in kg/mm^3 rather than kg/m^3, etc.  Same for thermal properties, etc.  Most users find it simpler to represent all BCs, properties, lengths, etc. in MKS units across both solvers, in order to avoid confusion.  So that would be lengths in m, density in kg/m^3, etc.

My guess would be that this is a case of values for properties, BCs, etc. being inconsistent with lengths represented in mm.

Lucia R

acupro_21778 said:

I'm not sure about OptiStruct - but AcuSolve (the solver) is unit-less.  The solver has no knowledge of units - it only has values/numbers.  Thus, it is up to the user to ensure those values represent a consistent set of units.  If you dimensions/lengths are in mm, then all units BCs, etc need to have values consistent with the length dimension being mm.  So density has to be be in kg/mm^3 rather than kg/m^3, etc.  Same for thermal properties, etc.  Most users find it simpler to represent all BCs, properties, lengths, etc. in MKS units across both solvers, in order to avoid confusion.  So that would be lengths in m, density in kg/m^3, etc.

My guess would be that this is a case of values for properties, BCs, etc. being inconsistent with lengths represented in mm.

I've represented all BCs, properties, lengths, etc. in MKS units across both solvers. AcuSolve seems to be running but an error occurs after some iterations (line 778 of the .Log file): 

 

acuSolve: External-Code stagger "external_code":
acuSolve: Skipped
acuSolve: CFL timeInc = 3.516369e-007
acuSolve: Error receiving data from external code to message <stop false>; expected <stop> received <region_id 1>

 

Do you know what is the problem and how to solve it?

 

Thanks!

acupro

Lucia R said:

I've represented all BCs, properties, lengths, etc. in MKS units across both solvers. AcuSolve seems to be running but an error occurs after some iterations (line 778 of the .Log file): 

 

acuSolve: External-Code stagger "external_code":
acuSolve: Skipped
acuSolve: CFL timeInc = 3.516369e-007
acuSolve: Error receiving data from external code to message <stop false>; expected <stop> received <region_id 1>

 

Do you know what is the problem and how to solve it?

 

Thanks!

I notice you're using version 2017.  It is possible there were problems with the exchange in that version.  Can you update to the latest release - 2021 - and see if that corrects the problem?  Based on the error, and the fact that the exchange information seemed reasonable, I would guess it's a problem with that version of the software.

Lucia R

acupro_21778 said:

I notice you're using version 2017.  It is possible there were problems with the exchange in that version.  Can you update to the latest release - 2021 - and see if that corrects the problem?  Based on the error, and the fact that the exchange information seemed reasonable, I would guess it's a problem with that version of the software.

I updated to the release 2020 and now AcuSolve run all time steps but it seems that the exchange with OptiStruct is not working.

 

I put below the .log and .out files, and also the cci file.

 

In the .out file:

 

Opening E:\MTR_Pila\FSI\unidades_m\3-FSI/cci.txt
External Code via CCI: *** ERROR: Expected <send> but received <stop false> from AcuSolve
External Code via CCI: *** ERROR: Received fatal error from AcuSolve in response to <temperature 180540>: fatal_error expected <stop> received <region_id 1>
**** ABORTING RUN DUE TO UNEXPECTED ERROR CONDITION ****

 

I hope you can help me.

 

Thank you very much.

acupro

Lucia R said:

I updated to the release 2020 and now AcuSolve run all time steps but it seems that the exchange with OptiStruct is not working.

 

I put below the .log and .out files, and also the cci file.

 

In the .out file:

 

Opening E:\MTR_Pila\FSI\unidades_m\3-FSI/cci.txt
External Code via CCI: *** ERROR: Expected <send> but received <stop false> from AcuSolve
External Code via CCI: *** ERROR: Received fatal error from AcuSolve in response to <temperature 180540>: fatal_error expected <stop> received <region_id 1>
**** ABORTING RUN DUE TO UNEXPECTED ERROR CONDITION ****

 

I hope you can help me.

 

Thank you very much.

Apologies...  In searching through another similar thread, it appears the thermal coupling for AcuSolve / OptiStruct has some ongoing issues.  The previous suggestion was to try the simulation using AcuSolve and OptiStruct version 2019.0.  Are you using full version - or EDU version?  Are you able to download 2019.0 for both solvers and try again?

Junta_20361

acupro_21778 said:

Apologies...  In searching through another similar thread, it appears the thermal coupling for AcuSolve / OptiStruct has some ongoing issues.  The previous suggestion was to try the simulation using AcuSolve and OptiStruct version 2019.0.  Are you using full version - or EDU version?  Are you able to download 2019.0 for both solvers and try again?

2019 doesn't work neither.

Actually, with the same input file, 2019 shows another error:

like "external node not defined..."

In 2020, it shows "the external code stagger skipped"

 

 

acupro

Junta_20361 said:

2019 doesn't work neither.

Actually, with the same input file, 2019 shows another error:

like "external node not defined..."

In 2020, it shows "the external code stagger skipped"

 

 

Is that latest with both OptiStruct and AcuSolve 2019.0?  Are you able to include/attach the .fem, .out, .inp, and .Log files?

Junta_20361

acupro_21778 said:

Is that latest with both OptiStruct and AcuSolve 2019.0?  Are you able to include/attach the .fem, .out, .inp, and .Log files?

hi Acupro,

you can find the hm files for both acu and OS with the log file from below link:

https://drive.google.com/drive/folders/1dh707P1Ykdy2eTMIlaUvTb9Gzp_284_9?usp=sharing 

thanks!

acupro

Junta_20361 said:

hi Acupro,

you can find the hm files for both acu and OS with the log file from below link:

https://drive.google.com/drive/folders/1dh707P1Ykdy2eTMIlaUvTb9Gzp_284_9?usp=sharing 

thanks!

It looks like you tried AcuSolve 2019.1.  Can you get 2019.0?

It would be most helpful if you can include the actual model files from your latest attempt.  AcuSolve (.inp and MESH.DIR or HYPERMESH.DIR) and OptiStruct (.fem).  I want to test exactly what you tried, but it seems the files I exported from HM are not complete for the coupled solution.

Junta_20361

acupro_21778 said:

It looks like you tried AcuSolve 2019.1.  Can you get 2019.0?

It would be most helpful if you can include the actual model files from your latest attempt.  AcuSolve (.inp and MESH.DIR or HYPERMESH.DIR) and OptiStruct (.fem).  I want to test exactly what you tried, but it seems the files I exported from HM are not complete for the coupled solution.

hi Acupro,

please find those solver decks in the attachment

acupro

Junta_20361 said:

hi Acupro,

please find those solver decks in the attachment

I ran these files without modification in 2019.0 - OptiStruct and AcuSolve.  There's an odd error message on the OS side at the very end, but other than that it all worked fine.  I believe the problem is on the AcuSolve side, so you need to use 2019/2019.0 AcuSolve (you may need to Include Older Packages), but I think you could use a newer OptiStruct version.  Hopefully this connection issue will be resolved in the next AcuSolve release.

I've attached the Log files from the 2019.0 run.  

Junta_20361

acupro_21778 said:

I ran these files without modification in 2019.0 - OptiStruct and AcuSolve.  There's an odd error message on the OS side at the very end, but other than that it all worked fine.  I believe the problem is on the AcuSolve side, so you need to use 2019/2019.0 AcuSolve (you may need to Include Older Packages), but I think you could use a newer OptiStruct version.  Hopefully this connection issue will be resolved in the next AcuSolve release.

I've attached the Log files from the 2019.0 run.  

hi Acupro,

Thank you for you help on this,

I will try with acusolve 2019.

I also tried with version 2021. The error still exist with "external stagger skipped".

Anyway, thank you very much!

Junta_20361

acupro_21778 said:

I ran these files without modification in 2019.0 - OptiStruct and AcuSolve.  There's an odd error message on the OS side at the very end, but other than that it all worked fine.  I believe the problem is on the AcuSolve side, so you need to use 2019/2019.0 AcuSolve (you may need to Include Older Packages), but I think you could use a newer OptiStruct version.  Hopefully this connection issue will be resolved in the next AcuSolve release.

I've attached the Log files from the 2019.0 run.  

hi, 
i would like to update about this. Using Solver 2019 for both Acu & OS
If i run from Hypermesh, the error still occurs.
But if i run from OptiStruct Solver Deck as below, the run can complete without error.
 

Lucia R

Junta_20361 said:

hi, 
i would like to update about this. Using Solver 2019 for both Acu & OS
If i run from Hypermesh, the error still occurs.
But if i run from OptiStruct Solver Deck as below, the run can complete without error.
 

Hi,

I have tried what you said in the latter reply but I obtain the same error: "Undefined external node..."

 

Can anybody help me?

 

I have tried with versions 2017.2, 2019.1 and 2020 and it doesn't work in none of them. Even I have tried to launch AcuSolve in version 2017 and OptiStruct in a newer version but it neither works. 

 

I don't have access to another versions, so I hope that anybody knows how to solve this problem.

 

Thanks in advance.

 

 

acupro

Lucia R said:

Hi,

I have tried what you said in the latter reply but I obtain the same error: "Undefined external node..."

 

Can anybody help me?

 

I have tried with versions 2017.2, 2019.1 and 2020 and it doesn't work in none of them. Even I have tried to launch AcuSolve in version 2017 and OptiStruct in a newer version but it neither works. 

 

I don't have access to another versions, so I hope that anybody knows how to solve this problem.

 

Thanks in advance.

 

 

This will most likely only work with 2019.0 of AcuSolve and OptiStruct.  Are you able to obtain those base 2019.0 versions?

Are you able to share the model files (consistent MKS for both solvers) here?  If not, it's probably best to contact your local support team - assuming you can share the files directly with them.

Junta_20361

Lucia R said:

Hi,

I have tried what you said in the latter reply but I obtain the same error: "Undefined external node..."

 

Can anybody help me?

 

I have tried with versions 2017.2, 2019.1 and 2020 and it doesn't work in none of them. Even I have tried to launch AcuSolve in version 2017 and OptiStruct in a newer version but it neither works. 

 

I don't have access to another versions, so I hope that anybody knows how to solve this problem.

 

Thanks in advance.

 

 

hi Lucia,
i tried 17,18, 19.1, 20 & 21. 
none of them work.

But AcuSolve & OS version 19.0 can run very well for this simulation as acupro said.

you should try with that version 

Lucia R

acupro_21778 said:

This will most likely only work with 2019.0 of AcuSolve and OptiStruct.  Are you able to obtain those base 2019.0 versions?

Are you able to share the model files (consistent MKS for both solvers) here?  If not, it's probably best to contact your local support team - assuming you can share the files directly with them.

I don't have the 2019.0 version.

 

I'm going to share with you the model files, can you try to run them in the 2019.0 version to check that it works?

 

Thanks!

 

The files are included in the following Drive link:

https://drive.google.com/file/d/1hPyhdTMZRfYPj-yXuQBo_IjMedaOiOvn/view?usp=sharing

 

acupro

Lucia R said:

I don't have the 2019.0 version.

 

I'm going to share with you the model files, can you try to run them in the 2019.0 version to check that it works?

 

Thanks!

 

The files are included in the following Drive link:

https://drive.google.com/file/d/1hPyhdTMZRfYPj-yXuQBo_IjMedaOiOvn/view?usp=sharing

 

The coupling for this does, indeed, work with 2019.0 of both solvers.  I did make a few changes to the AcuSolve .inp file.  I reduced min stagger iterations to 2 - as there's really no need to force a minimum of 4 - it will do as many as needed.  You don't need both Simple BC and External CS for the same surface - so I commented Simple BC.  I modified the pressure initial condition.

I've attached the updated .inp file.

The 2019.0 release should still be available through Altair Connect.  Make sure you check the box for 'Include Older Packages'.

I'm still hopeful this will all be corrected/fixed for the 2021.1 release packages.

Lucia R

acupro_21778 said:

The coupling for this does, indeed, work with 2019.0 of both solvers.  I did make a few changes to the AcuSolve .inp file.  I reduced min stagger iterations to 2 - as there's really no need to force a minimum of 4 - it will do as many as needed.  You don't need both Simple BC and External CS for the same surface - so I commented Simple BC.  I modified the pressure initial condition.

I've attached the updated .inp file.

The 2019.0 release should still be available through Altair Connect.  Make sure you check the box for 'Include Older Packages'.

I'm still hopeful this will all be corrected/fixed for the 2021.1 release packages.

Could you attach the solution files .h3d and .log or any screenshot of the results?

 

Thanks.

acupro

Lucia R said:

Could you attach the solution files .h3d and .log or any screenshot of the results?

 

Thanks.

If you point me to a location where I can drop the files, I will do so.  I've attached the AcuSolve .Log and OptiStruct .out files.

acupro

Lucia R said:

I updated to the release 2020 and now AcuSolve run all time steps but it seems that the exchange with OptiStruct is not working.

 

I put below the .log and .out files, and also the cci file.

 

In the .out file:

 

Opening E:\MTR_Pila\FSI\unidades_m\3-FSI/cci.txt
External Code via CCI: *** ERROR: Expected <send> but received <stop false> from AcuSolve
External Code via CCI: *** ERROR: Received fatal error from AcuSolve in response to <temperature 180540>: fatal_error expected <stop> received <region_id 1>
**** ABORTING RUN DUE TO UNEXPECTED ERROR CONDITION ****

 

I hope you can help me.

 

Thank you very much.

This appears to be working now in the 2021.1 release version.