## General Mechanical

#### How do I prevent my pressure vessel from flying away?

• michail.stockfelt
Subscriber

If you make a body with an internal pressure, you have to tie it down somehow or you will get rigid body motion. Usually by binding an orifice - tying up the pipe that usually connects to the vessel.

But this can become a problem if you for some reason get any imbalance in the model - accidentally missed some tiny inner surface (of the possibly complex shape) that multiplied with a large pressure results in your vessel tearing at the pipe sideways, or if you actually have no pipe (a balloon for example), or for whatever reason.

I once tried to model a pair of scissors to see what the results were when it cut a wire, and I had some problems balancing the forces. Ended up adding a rod that went out to nowhere just to bind that and get the clipping parts as unaffected by the constraints as possible.

Inertia Relief does not seem to be the answer, I can't get it to work because it seems to assume an acceleration field.

Is there any other trick to nullify resultant forces like that?

• Daniel Shaw
Ansys Employee

Symmetry is commonly used to model pressure vessels.  Symmetry boundary conditions provide numerical stability.

If the issue is numerical stability caused by small unbalanced nodal forces, you can try activating "weak springs" in the Mechanical UI.  Mechanical will then add spring elements with small stiffnesses to the model at DOFs without sufficient restraint.  The weak springs will carry the imbalanced nodal forces and prevent rigid body motion.  You can use a reaction probe to check the amount of force developed in the weak springs.  If it is very small, then they provided stability without affecting the overall model behavior.  If it is large then, the model does not have sufficient restraint and you need to manually restrain additional DOFs.

• michail.stockfelt
Subscriber

Okay, thanks.

These vessels are not symmetic, and the pressures involved are… not low… so the resultant unbalanced nodal forces tend to be quite wild. I need “strong springs”… Will definitely try that reaction probe thing, thanks!

My solution so far is to add features where I can put restraints, I just would prefer not to because they inevitably add quirks.

Another solution would be to use a wossname explicit transient analysis and set the density to “Neutron Star” and let it just fly, I suppose.

• Daniel Shaw
Ansys Employee

You can also try running a Mechanical implicit time transient analysis.  The inertia and damping forces might provide some stability.  It seems like the model is underestrained.  You may need additional restraints.

• michail.stockfelt
Subscriber

👍