[MultiPhy]

[Yes, I understand this is a for-pay freelance section!]

I want to simulate a rigid body in a creeping flow liquid using COMSOL (5.2). I have already essentially solved this problem using COMSOL's FSI module. However, the solid was not truly rigid as the FSI module does not offer the Structural Mechanics choice of making the solid a rigid domain. The solid body deformations, even though irrelevant to the problem, increase computational overhead, and I think cause convergence issues.

Given this, I would like to use the Creeping Flow physics for the liquid, and the Structural Mechanics physics for the solid. I would add "Rigid Domain" to the solid. I believe "Moving Mesh" needs to be applied to the solid also -- I assume the liquid domain takes care of mesh deformations by itself.

Here's where it gets hazy: The fluid needs to exert force on the solid body and cause it to move appropriately. For some reason, no matter how I try assigning positions or forces to the solid (there seem to be many possibilities, including "Prescribed Deformation" and "Prescribed Mesh Displacement" in the "Moving Mesh" node, and "Prescribed Displacement" and "Prescribed Velocity" in the Structural Mechanics node, among other possibilities) the solid doesn't move.

Depending on exactly how I set it up, sometimes nothing moves, and sometimes the solid's boundary moves without the actual solid (the domain?) moving. I don't see how that's even a possible outcome as I would think that boundaries would be tied to their domains. Nonetheless, I can use probes to confirm that there are forces acting upon the solid, and if I probe or plot boundary displacements, in some setups I see movement, but if I check the XY coordinates of the solid domain, it never moves.

Obviously this should be a very simple problem. All I want initially is a 2D rectangle, representing a pipe, with an inlet at one end, an outlet at the other, and a circle inside the rectangle, representing the solid. Assign "water" to the rectangle's domain, any solid to the circle's domain, make the circle a rigid body, and allow it to move with the fluid flow. In fact, for an initial model, I would settle for the solid just moving at all (e.g., by specifying an arbitrary velocity), but moving in a physically-correct manner given the fluid flow would be preferable.

While I think that inspecting the model will show me what I did wrong, a brief write-up (pointing out not only why you set the model up as you did, but what I did wrong) would be useful.

Note that this is really just a proof-of-concept for what will end up being a somewhat more complicated model that needs to be implemented in 3D and allow both Time Dependent and Stationary analysis, so if you can help with the next phase too, that would be great, but right now I just want to get the solid to move!

[MultiPhy]

I've got translation of the circle working in one model, and rotation in another model. Having problems merging them and getting realistic results. Anyone familiar with shear stress, moving walls, and global equations in COMSOL?