Move a solid in a fluid without LSI module (COMSOL 5.2)
 

I want to create a very simple model: A fluid channel with Stokes flow, with a sphere (2D is fine too) in it that gets pushed along by the fluid.

This is easy to do if you use the Fluid-Structure Interaction (FSI) module to connect the solid and liquid physics. But, the FSI module can be limiting (e.g., it does not allow specifying the sphere as a rigid domain -- it must be elastic) so I'd like to set the model up without it.

To do this, I am using Creeping Flow, Structural Mechanics, and Moving Mesh. From Structural Mechanics I added Rigid Domain to the sphere and then tried to couple the forces of the fluid to the sphere with various combinations of things in Moving Mesh (e.g., Prescribed Mesh Displacements, Prescribed Deformations) and Structural Mechanics (e.g., Prescribed Displacement, Applied Force, Applied Moment).

This seems like it should be simple, but no matter what I do my sphere won't move at all :confused:. Any input would be greatly appreciated!

Some progress...
 

I finally got this to work in an ugly manner. I separated the fluid domain and the solid domain into two Prescribed Mesh Displacement nodes under Moving Mesh. By keeping the fluid displacement 0 and displacing the solid over time, I can accomplish XY movement.

But, there is no way to specify angular velocity. I can think of a way to hack it in by splitting the circle in half, but what a kludge. I'd much rather figure out why the Structural Mechanics settings seem to have no effect, since at least some of them have X, Y and theta built in.

See update in freelancer forum -- happy to pay someone to do this! Either way, any help with shear stress, moving walls, and global equations in COMSOL would be appreciated.

Hey MultiPhy. I am working on something similar as well. Here are some tips.

If it is rigid, you don't need Structural mechanics. In 2d, use an ODE sum the forces in the x and y directions and set that equal to the mass times acceleration; also sum the moments about the center of the sphere. You can implement these equations with the "Point ODE" or "Global ODE" physics interfaces. If you have the multibody dynamics module, you may be able use that instead. Also, I think you will need automatic remeshing (check the help if this is knew to you) because the mesh will undergo large deformation and movement.

I don't know the exact problem with Structural mechanics but my guess it has to do with the fact that the deformation is 0 everywhere. Zero deformation is infinitely stiff. For instance, f=K*x and x=0. If we say f=10, what is K? You can't define it. The stiffness is infinite. In FEA, f=K*x is not solved when the displacement is known. Instead, x = 0 is what is used and K is never calculated at that node. I bet the checks in the Structural mechanics module can't deal with the fact that the deformations are 0 everywhere on the sphere.

If you actually want to look at this again, reply and we can work on this. Otherwise, this reply is for anyone who ends up on this thread and needs a general direction.

I am sure i can help you to model the solid motion in comsol ...first you use the fsi module then you are not choose the fix condition ... because it doesnt allow the particle to move throw the channel.... and for youe secend question you can choise whatever material you want so it cant be elastic..

I usualy online on telegram so you can send me any question about comsol I write my ID on telegram below ...
@habibjalili or http://telegram.me/habibjalili