Hi all,

Can anybody tell me how to model a swivel joint?

Thanks,

Check this newsletter article.

Thanks Richard.

I'll correct my question: How to model a piping system with three swivel joints?

I modelled each swivel joint with 6 restraint with cnode. But the results don't seem right. The displacements of the middle swivel joint don't match my hand calculations. The newsletter suggested to use y-rod for modelling. But it does not work for a well head piping with 3 swivel joints installed.

Any idea how to make it right?

Caesar II may give you results but for one of an infinite number of results. The model is statically indeterminate, i.e., too many unrestrained DOF. The rule of thumb limit is two gimbals and one hinge. You have three gimbals.

As a beam model, CAESAR II represents the stiffness of all elements based on their initial position. The "shortcoming" of this approach is the model is not constrained by geometry; the stiffness matrix is not updated to reflect the deflected position.
Try this - put a very large Y load on a flimsy, horizontal cantilever. Look at the end rotation; it is (or can be) greater than 90 degrees. The deflection may even be greater than the beam length. How is this "possible"? Simply divide the load by the stiffness.
That's what makes the results from that newsletter ball joint model interesting. Our +XROD-style restraint IS constrained by position.
There are engineering tools to evaluate these linkages. CAESAR II is not one of them. Stick to small rotations in CAESAR II.
I recommend you work with the vendor of that joint assembly to figure this stuff out.