Flexible Ball Impacting Water Surface

[Trinidad Mcdavids]

Hello, 

I am trying to simulate a case of flexible ball impacting water surface (similar to the case cylinderDrop, the difference is the cylinder is elastic install regid). The ball mesh was imported from obj file instead of generated manually.

1. I made an example case out of "fsi_felxbile element" & "fsi_cylinderdrop", but is a bit lost regarding the BCE setting? how can the BCE be applied for the imported fem mesh? 

2. The examples case i made is attached. It seems the case can not be compiled with many errors. Could you please kindly help to quick check what is wrong in my codes?

3. Regarding the post processing, i would like to show the deformation & stress of the elastic body as well as the pressure/velocity contour map of the fluid to be exported to vtk file. how can this be done?

[Wei Hu]

Hi Trinidad,

In the current implementation of Chrono::FSI, we have very limited support for interaction with flexible bodies. We only support flexible bodies described with the ANCF cable element or ANCF shell element. Unfortunately, we can not read directly from a mesh file. You should manually input those elements and nodes to a mesh. The ANCF shell elements should also be rectangular elements which I believe it's not the case with the mesh you import from an OBJ file. Another important thing you need to take care of is related to  *NodeNeighborElement_mesh* and *_2D_elementsNodes_mesh*. The two vectors contain important information that is needed for the simulation and should be initialized with the correct value. You can take a look at that flexible element demo and see how these two vectors are initialized therein. 

We haven't maintained the flexible bodies support in FSI for about three years, so it probably takes a while to modify the code and make it works for your case. However, the good news is, we restarted the maintenance of this feature two weeks ago. The ongoing work in the next couple of months is to make it available to directly read a mesh from an OBJ file and generate BCE automatically, which I believe is exactly the feature you want to use in your demo, so the user can use this code much easier. We'll make it as fast as possible, please stay tuned with the *main* branch of Chrono.

Thank you,

Wei

--
You received this message because you are subscribed to the Google Groups "ProjectChrono" group.
To unsubscribe from this group and stop receiving emails from it, send an email to projectchron...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/projectchrono/d7917585-9dd6-4989-9878-dd86fb95d425n%40googlegroups.com.

[Trinidad Mcdavids]

Dear Wei,

Since my case is a bit urgent, I would like to try the manual method of generating the mesh. Do you think it is feasible?

I can generate quad shell mesh using gmsh, do you think it is possible to export to the manual mesh vector? I can write a small function to do this, e.g read gmsh .msh file and export the vectors.

Finally, how about the post-processing? do you think it is possible to extract the flexible body stress & deformation to vtk file? how about fluid contour map?

[Wei Hu]

Hi Trinidad,

Yes, if you can generate a mesh file based on the FSI requirement, I think it's doable. You can drop your updated mesh file and new code here, so we can take a look at it, and probably can do a test for you at our end.

For post-processing, we can extract the stress and deformation of the flexible body and also the pressure/velocity field of the fluid.

Thank you,

Wei

To view this discussion on the web visit https://groups.google.com/d/msgid/projectchrono/aa66fa4f-5911-4e2b-8ccb-a16200735e89n%40googlegroups.com.

[Trinidad Mcdavids]

Hi Wei,

I tried to get myself familiar on the vector structure, but it seems a bit challenging.

Do you have existing functions that can extract the required elements/nodes vector from existing mesh files? I attached a nastran BDF file, if needed, i can also convert it to gmsh .msh file or abaqus .inp file. 

Thanks & Regards,

Trinidad

[Trinidad Mcdavids]

file now attached. three format provided.

[Wei Hu]

Can you please convert the mesh file to OBJ format so we can take a look. Keep in mind that we only support mesh with rectangle elements now. Otherwise, it probably will take a while to get this case built correctly.

Thank you,

Wei

To view this discussion on the web visit https://groups.google.com/d/msgid/projectchrono/f4574bc1-e0bd-4490-9c3e-a6b10fe88274n%40googlegroups.com.

[Trinidad Mcdavids]

Hi Wei,

The updated obj mesh is attached. The mesh is of quad elements, but not rectangle elements.

I am not sure whether this would be OK. 

[Wei Hu]

Hi Trinidad,

Thanks for the mesh file, but I'm not sure if the quad element is Ok or not, I'll need to talk to our lab member who is working on the FEA to see if we can model this problem using the ANCF shell element.

Radu,

Do you think this is doable using the ANCF 3423 shell element?

Thank you,

Wei

To view this discussion on the web visit https://groups.google.com/d/msgid/projectchrono/e481ea96-26d8-4252-a196-abde4a4c6da7n%40googlegroups.com.

[Wei Hu]

Hi James,

I took a look at your mesh file, since it's not discretized with rectangle elements. We currently cannot run it directly in FSI, we still have issues generating BCE particles from an OBJ file like that. However, we are working on a similar problem which might be helpful with your project. We can run a simulation of a flexible toroidal tire moving on deformable terrain (which should be easy to switch to fluid since both of them are modeled using the SPH method in the Chrono::FSI module). The reason I think it might be helpful is that your mesh just looked like a tire. If you can use the toroidal tire mesh instead of your own mesh. I think things should be much easier to move forward.

Here is an animation of the flexible tire moving on deformable terrain. We are going to commit the code to GitHub soon, probably in a couple of days.

Let me know what you think.

Thank you,

Wei

On Wed, Aug 17, 2022 at 1:36 AM James <kevin...@gmail.com> wrote:

Hi Wei,

Do you have any update on this?

To view this discussion on the web visit https://groups.google.com/d/msgid/projectchrono/36808434-d63b-4dca-91af-a93d83895e0cn%40googlegroups.com.