Just to emphasize what Rich has already said, it is not appropriate to use the Appendix P flexibility criteria for tank roof nozzles for the reasons he cited. It is not even appropriate to use these criteria for shell nozzles for smaller than 120 ft. diameter tanks or for other than "low-type" nozzles.

Using FEA to calculate the appropriate flexibilities and using a CII flexible anchor is certainly one way to account for the roof flexibility. That should help out a lot since both the tank roof and shell are pretty flexible structures. I assume that you would then use an ASME Division 2 type analysis when you get into evaluating the nozzle and shell stresses that result from the imposed loads.

I assume that the tank has a dome roof since you are designing for a small internal pressure. If this is the case, it might be worthwhile and simpler to try "fooling" the program (sorry Rich) and have CII calculate the flexiblility for you using one of the built-in flexible nozzle options.

The BS 5500 option lets you put in a spherical vessel directly. To use the WRC 297 option, you would probably have to first come up with a cylinder of equivalent stiffness under applied load (perhaps hand calcs. using Roark?), then use that in CII.

I also think that it would be worthwhile to test the pressure drop limit and the "no bends" requirement with the process engineers before resorting to using expansion joints. It's easy for them to "impose" that requirement, but they don't have to worry about bellows design, installation, and possible failure.
The roof flexibility still might not be enough to make things OK with the straight risers and the straight header run between the nozzles. How about offsetting the header from the nozzle centerline so that you can atleast get one elbow into each riser? The process guys might be able to live with that. With a valve in each riser, you also have to worry about any portion of that system being shut off and at ambient temperature while the rest is at 260F.

Hope this, in conjunction with Rich's comments, is helpful to you. Good luck!

Vince

------------------
Vincent A. Carucci
President
Carmagen Engineering
7 Waverly Place
Madison, NJ 07940

Visit our web site at http://www.carmagen.com