I am having trouble agreeing with our structural dept. about how the value of restraint forces from the pipe to vessel shall be calculated. We have a 100ft tall, 7'-8" dia. vessel with a 24" dia. line coming out of the top and dropping down the length of the vessel. We are designing the pipe to B31.3 which states that all seismic must comply with the Universal Building Code (UBC). In that code there is formulae under section 1632.2 DESIGN FOR TOTAL LATERAL FORCE, which states:

Fp=4.0*Ca*Ip*Wp or alternatively
Fp=(Ap*Ca*Ip/Rp)*(1+3*Hx/Hr)*Wp

These formulae seem to assume that the seismic event is acting on the bottom the attachment (pipe) and the forces get larger for each guide as you go up the vessel. Since the pipe is attached to the top of the vessel, it appears to my intuition that the exact opposite is the case. After modeling the Vessel and the pipe and running a dynamic seismic analysis my intuition was confirmed. The forces that Caesar II produces are similar in magnitude but opposite in the fact that they are smallest at the top of the vessel and get larger as the you look down the vessel.

My question is thus;
"Can I defend the values of Caesar II vs. those of the UBC in respect that the pipe is hanging from the top and not supported at the bottom?"

"Can anyone direct me to example seismic calculations of pipe hanging from pressure vessels?"

The codes that I have reviewed do not appear to have any specific discussion on this particular application. Any input would be greatly appreciated.

In your description system the biggest lateral force will become on the Vessel Anchor.
The biggest lateral force in the Pipe will become on the intersection Vessel/Pipe.
The force distribution will be as follow:
The greatest force are on the down site vessel and will be come smaller in the up site
The greatest force in the pipe will become on the intersection to the vessel and the
forces will be smaller in the down site the pipe.
I think your problem is simple to calculate with Caesar. But you must look for the right
element lenght (lumped mass), the stiffness on the intersection and the mass distribution
on the vessel.
Use the UBC Spectrum.

If we assume that your vessel (tower) is anchored at the base, then the greatest motion will be at the top. Hence your conclusion as to the UBC formulae (greater loads as you go up).

Now if you attach piping to the top of the vessel, and run it down the side, with no other restraints, then you have essentially a whip. The motion at the bottom of the pipe will be larger than at the top, where it attaches to the vessel.

Is this realistic? Perhaps not. The reason is that you must have other restraints down the vessel, and along the pipe once it reaches the ground. These restraints will nullify the "whip" effect and you will move back to (the idea of) increasing loads as you move up.

Why might you obtain different results from those expected? This depends on how you modeled the restraints, specifically your guides down the vessel. If you applied double acting horizontal restraints, then they should be active in the dynamics solution. However, if you applied single directional horizontal restraints, or horizontal restraints with gaps, then they were linearized by the program for the dynamic solution. The linearization process either makes them linear (removes the gap), or it removes them completely. Check your <font color="0000ff">"Active BCs"</font> report (Active Boundary Conditions) to see how restraints were linearized. If this isn't what you want, then you may have to manually adjust the restraints before analyzing the system.


------------------
Regards,
Richard Ay (COADE, Inc.)



[This message has been edited by rich_ay (edited January 14, 2000).]

I agree with Rich and advise you to continue the model past the bottom of your tower, to a logical point of isolation or a boundry condition which can clearly be defined. A so-called anchor (3)fixed linear vectors and incidental rotational stops would fit the bill.

Also the period of the tower and hence the downcomer strapped to it will probably be different than the structure which the "anchor" is attached to hence consider SAM effects (Seismic Anchor Motions)

------------------
Best Regards,

John C. Luf