Process structure seismic design 1

[SocklessJ]

I’ll soon be designing an open steel tower to support a heavy reactor and other process equipment. The reactor is supported by the tower, via a support ring at its mid-height.

ASCE 7 states “where the mass of the operating vessel or vessels supported is greater than 25% of the total mass of the combined structures, the structure and vessel designs shall consider the effects of dynamic coupling between each other. Coupling with adjacent, connected structures such as multiple towers shall be considered if the structures are interconnected with elements that will transfer loads from one structure to the other.”

Some more questions I have are:
1. Do you know of any good references for design of process structures like this? I already have ASCE’s guides for seismic and wind design of petrochemical facilities.

2. Is a Modal Response Spectrum Analysis overkill? The location is SDC C, however due to heavy masses potential irregularities, ELF might not capture the response. Also, page 195 (item 6) of the ASCE petrochem guide recommends it. The biggest problem I see is combining the Direct Analysis Method with a MRSA. Some options might be:

a. If a MRSA is not required by code, maybe I could do one separately without the P-D analysis, just as a gut check.
b. STAAD has an “Individual Modal Response” option, which will form a primary load case with scaled forces corresponding to each mode shape. I could try running a Direct Analysis for each mode (combined with dead load), SRSS the results, and code check that.​

3. How should I model the reactor to account for the coupling? I’m thinking a series of lumped masses connected by members with section properties matching that of the vessel shell. My guess is that it will want to rotate about its mid-height support level. This wouldn’t be captured if it was treated as a single lumped mass.

4. In addition to the reactor there will be other pieces of (lighter) connected equipment supported by the structure. Since this equipment will probably be <25% of the total mass, I assume it can be decoupled from the structural model and the mechanical engineer will simply provide reactions for the equipment, pipe supports, etc. Is this typically how it works?

Any thoughts would be greatly appreciated.

 

[WARose]

1. Do you know of any good references for design of process structures like this? I already have ASCE’s guides for seismic and wind design of petrochemical facilities.

You've got the only ones I can think of.

2. Is a Modal Response Spectrum Analysis overkill? The location is SDC C, however due to heavy masses potential irregularities, ELF might not capture the response. Also, page 195 (item 6) of the ASCE petrochem guide recommends it.

You go where the code takes you. If it mandates it.....then you have to do it. However you can use one to back check the other. I've put a lot of equipment on (flexible) steel frames where I did a MRSA to be sure I was capturing the right amount of seismic load. (Even if the code allowed a ELF.) Most times, you wind up with about the same number you'd get using the ELF.

3. How should I model the reactor to account for the coupling? I’m thinking a series of lumped masses connected by members with section properties matching that of the vessel shell. My guess is that it will want to rotate about its mid-height support level. This wouldn’t be captured if it was treated as a single lumped mass.

Unless the vessel is somehow altering the lateral load resistance of the structure (in terms of stiffness).....the lumped mass approach is probably best. Most equipment is along for the ride in a lateral event.

The thing you have to watch for (as much as anything) for equipment supports in such a structure for seismic load is adequate anchorage. (See chap. 15 in ASCE 7-10.) You won't come out of the ELF (for the whole thing) with the right number there.

4. In addition to the reactor there will be other pieces of (lighter) connected equipment supported by the structure. Since this equipment will probably be <25% of the total mass, I assume it can be decoupled from the structural model and the mechanical engineer will simply provide reactions for the equipment, pipe supports, etc. Is this typically how it works?

Yep.