Applying Moments to line 3 beam using ccx and nonlinear 3d dynamic

barrti

Greetings
Its probably too late here but I cant get this model to run. What am I missing? Every time I add a moment to a node on the shaft on the x-axis to rotate the shaft, the software fails to find a solution.

Victor

I'll look again later but you might need to use a force couple on an additional element instead. Force on the wheel works.

Victor

I'm not sure why this happened but adding a dummy element with zero density to the node with the moment makes it solve.

I also turned on automatic time stepping which helps it get further into the solution but it only rotates about 45 degrees.

barrti

Thanks Victor

That is very wierd? Do u think that this is a ccx bug/peculiarity?

Maybe the arm you added gives the moment something to push on:)

Some searching of the Calculix ccx info on beam elements:

http://web.mit.edu/calculix_v2.7/CalculiX/ccx_2.7/doc/ccx/node53.html

About a page into this link they talk about knots:
"In addition, a knot is also generated if

-a rotational degree of freedom in the node is constrained through a SPC or MPC. In this case, the knot introduces the necessary rotational degrees of freedom.

-a bending moment or torque is defined in the nodes. Here too, the rigid body introduces the necessary rotational degrees of freedom."

Just a guess - is it possible that this has something to do with it?

I did get it to run to completion by using auto timestep with 100,000 steps. Took about 5mins. I also used a gentler torque ramp

Thanks again.

Regards

Victor

I think you're probably onto it with the knots. I don't understand them well enough to say though. Generally, all the 6-dof elements in CCX are a bit flimsy and it sometimes fails in surprising ways. I've seen the geometry of connections between elements determine success or failure of the solver on shells too.

barrti

Victor,
Thanks for that. FEA is an art often it seems!

Cheers