[farah.elias]

Hello,

I came across 3 different FSI solvers that can be used inside Foam-Extend:

• Extend-Bazaar FSI toolkit (https://openfoamwiki.net/index.php/E...re_interaction), the main author being Zeljko Tukovic and many co-authors, for instance, Philip Cardiff.
• solids4foam Toolbox (https://bitbucket.org/philip_cardiff...se/src/master/), developed mainly by Philip Cardiff and Zeljko Tukovic.
• FOAM-FSI Solver (https://github.com/davidsblom/FOAM-FSI), developed mainly by David Blom along with other contributors like Thijs Gillebaart and Philip Cardiff.

I understood from another forum I found that solids4Foam is actually based on Extend-Bazaar FSI toolkit. I would like to know the reason(s) behind the switch from the Extend-Bazaar toolkit to solids4Foam toolbox. As well, in the solids4Foam, is the coupling between the fluid and solid solvers done internally in OpenFOAM?

Also, when I came across the FOAM-FSI solver, I understood one thing about it, that a preCICE adapter was written for Foam-Extend-3.2, where this adapter cannot be found on the precice.org for the Foam-Extend variant of OpenFOAM. Besides this additional functionality that I think is not included in solids4Foam, what is the difference(s) between solids4Foam and FOAM-FSI? do they solve the fluid flow and the structural part using similar algorithms (methods/techniques)? and do the same coupling methods exist in both solids4Foam and FOAM-FSI in order to couple both solid and fluid solvers?

Finally, I know there is another fourth option to solve a strongly coupled FSI problem, which is to use a fluid solver (for example using the fluid solvers of OpenFOAM) and an in-house built solid solver and couple them using preCICE. Do you believe that the latter is fairly doable because of the complexity of the coupling process with preCICE? and any comment or advice regarding this approach? or is it preferable to build the (in-house) solid solver inside solids4Foam and run everything inside Foam-Extend?

Thanks a lot in advance.

Regards,
Elias

[bigphil]

Originally Posted by farah.elias

I understood from another forum I found that solids4Foam is actually based on Extend-Bazaar FSI toolkit. I would like to know the reason(s) behind the switch from the Extend-Bazaar toolkit to solids4Foam toolbox. As well, in the solids4Foam, is the coupling between the fluid and solid solvers done internally in OpenFOAM?

solids4Foam is essentially just an evolution of the Extend-Bazaar FSI toolkit, where some differences are:

• There is only one solver (solids4Foam)
• There are many more solid models
• There are many more solid mechanical laws
• There are many more solid boundary conditions
• It works with OpenFOAM.org and OpenFOAM.com as well as foam extend
• solids4foam is under active development

As regards "is the coupling between the fluid and solid solvers done internally in OpenFOAM?", yes, the coupling is the same as in the Extend-Bazaar.

Originally Posted by farah.elias

Also, when I came across the FOAM-FSI solver, I understood one thing about it, that a preCICE adapter was written for Foam-Extend-3.2, where this adapter cannot be found on the precice.org for the Foam-Extend variant of OpenFOAM. Besides this additional functionality that I think is not included in solids4Foam, what is the difference(s) between solids4Foam and FOAM-FSI? do they solve the fluid flow and the structural part using similar algorithms (methods/techniques)? and do the same coupling methods exist in both solids4Foam and FOAM-FSI in order to couple both solid and fluid solvers?

One comment on preCICE with solids4foam, you can find a preCICE adapter for solids4foam on the feature-precice-adapter-openfoam-v1912 branch, see here: https://bitbucket.org/philip_cardiff...ction-preCICE/. I hope to merge this into the master branch in the coming months.
In the example tutorial, it uses the standard built-in OpenFOAM pimpleFoam solver for the fluid, solids4foam for the solid, and preCICE for the coupling.

As regards differences between solids4foam and FOAM-FSI, well FOAM-FSI was developed mostly independently although the solid solvers come from the same foam-extend origin. You will see that the included coupling algorithms are independently developed too and have some different features (I can see there is an interface to a dealii solid solver and also to preCICE). Although FOAM-FSI no longer seems to be under active development.

Originally Posted by farah.elias

Finally, I know there is another fourth option to solve a strongly coupled FSI problem, which is to use a fluid solver (for example using the fluid solvers of OpenFOAM) and an in-house built solid solver and couple them using preCICE. Do you believe that the latter is fairly doable because of the complexity of the coupling process with preCICE? and any comment or advice regarding this approach? or is it preferable to build the (in-house) solid solver inside solids4Foam and run everything inside Foam-Extend?

I guess it depends on what your problem is and what you background experience is. It may be possible to solve your problem in solids4foam, and you could now also try preCICE with an OpenFOAM fluid solver and solids4foam solid solver. Or you can build your workflow around preCICE and use an OpenFOAM fluid solver with one of the available open-source finite element solid solvers.

Feel free to describe your problem here and I (or others) may have suggestions/comments.

Philip

[pi06jl6]

I can say only the following about the packages: many good research articles have come from these packages, I would recommend preCICE and pick your poison from https://precice.org/adapters-overview.html.

OF is indeed a valuable toolbox.

But it requires significant insight into the algorithms, a steep learning curve.

I wish the best luck in your endeavors

[Tianyz]

Hi, I'm using solids4Foam to simulate some FSI cases, but there is an error in my case.

--> FOAM FATAL ERROR: 
Inlet patch name inlet not found.

    From function setInletVelocity::setVelocity()
    in file functionObjects/setInletVelocity/setInletVelocity.C at line 78.

FOAM aborting

Maybe this error is related to the controDict file, but it doesn't work even I comment the functionobject in the controDict file. Can anyone please help me with it? Thank you in advance.

[Tianyz]

Originally Posted by Tianyz

Hi, I'm using solids4Foam to simulate some FSI cases, but there is an error in my case.

--> FOAM FATAL ERROR: 
Inlet patch name inlet not found.

    From function setInletVelocity::setVelocity()
    in file functionObjects/setInletVelocity/setInletVelocity.C at line 78.

FOAM aborting

Maybe this error is related to the controDict file, but it doesn't work even I comment the functionobject in the controDict file. Can anyone please help me with it? Thank you in advance.

I fixed it, just comment this function, tried again, and then it works.

[Tianyz]

Hi there,

I'm new to solids4Foam. And I'm doing deformable overset mesh and FSI based on OpenFOAM. I went through OpenFOAM-2106 initially and then switch to solids4Foam because the latter one supported deformable overset mesh. However, I noticed that there are some differences between these two versions. In OpenFOAM-2106, all of the configurations can be set within a single file - dynamicMeshDict, there are a lot of entries, for instance, mass, restraints, and so on. For solids4Foam, apart from the dynamicMeshDict file, another file oversetMeshDict is necessary, so do they have the same entries ? And what's the function of the motionSubset folder? Can anyone please explain these to me?
I'll appreciate it!


Best,
Tian

[Tianyz]

Hi there,

I created a case based on the overset folder from the tutorial case in solids4Foam (master brunch).
The case is flow around the long slender cylinder, which can deform according to the force exerted by the fluid domain on the surface of the structure.

Presumably, the two ends of the cylinder should be fixed on the wall, while the middle of the cylinder should deform. But the truth is the case diverged, and one end of the cylinder moved.


I uploaded the details about this case and the log file. Any comments about this case will be appreciate!

[Tianyz]

Originally Posted by tianyz

hi there,

i created a case based on the overset folder from the tutorial case in solids4foam (master brunch).
The case is flow around the long slender cylinder, which can deform according to the force exerted by the fluid domain on the surface of the structure.

Presumably, the two ends of the cylinder should be fixed on the wall, while the middle of the cylinder should deform. But the truth is the case diverged, and one end of the cylinder moved.


I uploaded the details about this case and the log file. Any comments about this case will be appreciate!

D.png

DD.png

p.png

U.png

log.txt

[quarkz]

Originally Posted by Tianyz

Hi there,

I created a case based on the overset folder from the tutorial case in solids4Foam (master brunch).
The case is flow around the long slender cylinder, which can deform according to the force exerted by the fluid domain on the surface of the structure.

Presumably, the two ends of the cylinder should be fixed on the wall, while the middle of the cylinder should deform. But the truth is the case diverged, and one end of the cylinder moved.


I uploaded the details about this case and the log file. Any comments about this case will be appreciate!

Hi Tianyz,

You have a very interesting case. Btw, you mentioned that the foreground overset mesh in OF can deform. Is that possible in the original OF?

If it is possible with solidFoam, is there a tutorial which shows us how it can be done? Hope you can give me the link.

Btw, my problem involves the flow past a wing. The wing deforms in a prescribed manner. I am interested to see what happens to the wing and flow field when the wing deforms in a 6DOF scenario (mainly 2DOF in pitch and y translation).

As mentioned earlier, it is not possible to deform my foreground overset mesh in the standard OF. Hence, I'm looking for alternatives.

Thanks!

I have

[bigphil]

FYI, I added a tutorial showing how to couple solids4foam and overPimpleDyMFoam via preCICE to the new solids4foam website: https://solids4foam.github.io/tutori...tCylinder.html

[Tianyz]

Originally Posted by quarkz

Hi Tianyz,

You have a very interesting case. Btw, you mentioned that the foreground overset mesh in OF can deform. Is that possible in the original OF?

If it is possible with solidFoam, is there a tutorial which shows us how it can be done? Hope you can give me the link.

Btw, my problem involves the flow past a wing. The wing deforms in a prescribed manner. I am interested to see what happens to the wing and flow field when the wing deforms in a 6DOF scenario (mainly 2DOF in pitch and y translation).

As mentioned earlier, it is not possible to deform my foreground overset mesh in the standard OF. Hence, I'm looking for alternatives.

Thanks!

I have

Hi quarkz,

Sorry for to reply delay, but to my knowledge, I did this deform overset mesh based on Foam-extend 4.1 and solid4Foam, everything works well, but can not run in parallel. There is a tutorial come with soild4Foam where you can check it. Hope this information is helpful

Best,
Tian

[Tianyz]

Originally Posted by bigphil

FYI, I added a tutorial showing how to couple solids4foam and overPimpleDyMFoam via preCICE to the new solids4foam website: https://solids4foam.github.io/tutori...tCylinder.html

Many thanks for your information, Cardiff! It is useful!

[quarkz]

Originally Posted by Tianyz

Hi quarkz,

Sorry for to reply delay, but to my knowledge, I did this deform overset mesh based on Foam-extend 4.1 and solid4Foam, everything works well, but can not run in parallel. There is a tutorial come with soild4Foam where you can check it. Hope this information is helpful

Best,
Tian

Hi Tianyz, no worries. Do you remember what's the name of the tutorial? Or what's the simulation about? cylinder, airfoil? This will make it easier to search.

Do you know why it can't work in parallel? Does the problem lie in OF or Solid4Foam? Maybe I can give it a try to see if it can work in parallel. If it can't, then its use will be very limited since most problems need to run in parallel. Maybe we can check with the developers to see how this problem can be solved.

[Tianyz]

Originally Posted by quarkz

Hi Tianyz, no worries. Do you remember what's the name of the tutorial? Or what's the simulation about? cylinder, airfoil? This will make it easier to search.

Do you know why it can't work in parallel? Does the problem lie in OF or Solid4Foam? Maybe I can give it a try to see if it can work in parallel. If it can't, then its use will be very limited since most problems need to run in parallel. Maybe we can check with the developers to see how this problem can be solved.

Hi quarkz,

There is a case in fsi folder in soilds4foam named beamIncrossFlow, where a subfolder overset is there. As for the reason why it can't run in parallel, sorry, I don't know too much, I think it should relate to OpenMP. As Cardiff replied, maybe OpenFOAM coupled with preCICE is another way to do it.

Best,
Tian

[Youngxl]

Originally Posted by bigphil

FYI, I added a tutorial showing how to couple solids4foam and overPimpleDyMFoam via preCICE to the new solids4foam website: https://solids4foam.github.io/tutori...tCylinder.html

Hi Philip,
As far I know, this example is already in solids4faom-v2.0. I want to ask if this example can run on foam-extend 4.1, as I see on the preCICE official website that it does not apply to this version of fe.

With regards,
Young

[bigphil]

Originally Posted by Youngxl

Hi Philip,
As far I know, this example is already in solids4faom-v2.0. I want to ask if this example can run on foam-extend 4.1, as I see on the preCICE official website that it does not apply to this version of fe.

With regards,
Young

Hi Young,

The preCICE adapter does not currently support foam-extend. I am sure it should not be much effort to port it to foam-extend, for example, using compiler directives like in solids4foam, but it is on my long to-do list.

If someone works the adapter and starts this work, I am happy to advise and make suggestions.

Philip

[Youngxl]

Originally Posted by bigphil

Hi Young,

The preCICE adapter does not currently support foam-extend. I am sure it should not be much effort to port it to foam-extend, for example, using compiler directives like in solids4foam, but it is on my long to-do list.

If someone works the adapter and starts this work, I am happy to advise and make suggestions.

Philip

Okay, thank you Philip. I am new to preCICE and maybe it will take me much time to do it. However, I will try it.

Young

[bigphil]

You can find the preCICE adapter here: https://github.com/precice/openfoam-adapter.

I suggest you fork the repository and then try compile it with foam-extend-4.1.

You may have to remove some checks saying foam-extend is not supported, and then go through the compiler errors one-by-one. It may be easiest to temporarily disable the compilation of many classes in the Make/files and add them back as you go. You can commit to your fork as you make progress and feel free to include me (@philipcardiff) in comments on issues you face.

[Youngxl]

Originally Posted by bigphil

You can find the preCICE adapter here: https://github.com/precice/openfoam-adapter.

I suggest you fork the repository and then try compile it with foam-extend-4.1.

You may have to remove some checks saying foam-extend is not supported, and then go through the compiler errors one-by-one. It may be easiest to temporarily disable the compilation of many classes in the Make/files and add them back as you go. You can commit to your fork as you make progress and feel free to include me (@philipcardiff) in comments on issues you face.

Okay, thanks for your suggestion.

[Tianyz]

Originally Posted by bigphil

FYI, I added a tutorial showing how to couple solids4foam and overPimpleDyMFoam via preCICE to the new solids4foam website: https://solids4foam.github.io/tutori...tCylinder.html

Hi, Prof. Cardiff,

Just want to let you know that the newest solids4Foam website doesn't work now, I will try it later.

Another question I'm wondering is have you ever considered the possibility of adding this solids4Foam solver in openfoam esi? Or is it a big project to port a limited library of solids4foam into openfoam esi? Since I should use some modified turbulence model compiled based on OpenFOAM esi, it'll be much more compatible to run the case within one package to some extent.

Best,
Tian