Discussion Forum

Hi

COMSOL is not the easiest to do rigid body physics, I would suggest to use a 2D shell for the flexure and a 2D bar for the two "rigid bars", from their relative size you can make them "infinite" rigid compared to the flexure. In this way you have only 3 lines to draw, and you can add "hinges easily, for the shell elements do not forget to couple the displacement AND the rotation

--
Good luck
Ivar

Hi,

If you wish to do things using rigid connectors, you should simply set the two translational degrees of freedom equal; e.g u_rig1 = u_rig2 and v_rig1 = v_rig2. Both rigid connectors should have their center of rotation at the pin center.

But since the outer part is free to rotate around the pin, the model is a mechanism. The way you have sketched it, it could only be used in a transient analysis, since it has one possible rigid body rotation.

With reference to Ivar's answer above: There is nothing such as a 2D shell, so I assume that he means 2D beam. The 2D beam has no geometrically nonlinear formulation, so in that case the rotation of the inner part must be small.

Regards,
Henrik

Hi Henrik

Your right it remain beams in 2D ;)

I do find it's a pity that COMSOL is not better/simpler set up for rigid body mixed models, as you have everything in there. Hope you will add some further BC/Domain conditions for "mechanisms"

Rigid connectors in 3D are nice but require some experience to handle correctly, but OK they are very good in specific cases, particularly for flexure structures with large stiffness differences

--
Good luck
Ivar

Thank you for your responses. I was able to model the mechanism successfully in 3D for low applied loads.

Now I am running into convergence issues when I input a pre-described displacement on the moving rigid connector.
Since I am expecting large deformations, I checked the geometric nonlinearity option and I am using a parametric solver through the study extension option to help the solution converge, however, I have been unsuccessful.

Do you know of any tweaks with regards to the solver settings that may help convergence for a geometric nonlinear problem? i have been following the "large deformation beam" example and I changed the relative tolerance to 1e-6 but still no go. For now, I am going to make my mesh finer in the areas of interest and see what happens.

thanks for all your help :)

Oscar

I got it to converge, one of the global constraint equations for the rigid connector was wrong.

Thanks guys

Oscar