Horizontal Vessel Support Saddles - Web at edge or web at center? 2

[hull6851]

My company often deals with placing vessels on skids (steel structure) instead of on concrete pads. Since our supporting steel (skid) is constructed of I-beams, my opinion is that we also ought to have vessel support saddles with a corresponding I-beam sections (web of the saddle centered). However, a colleague believes that a L shaped cross section (with web of the saddle at the edge) is fine, and perhaps even preferable (because of apparent cost benefits (less welding to fabricate the saddle). I'll concede that it may be (slightly) less costly, but I feel there is a distinct advantage to aligning the webs of the saddle and the underlying skid structure, both for load path considerations and for bolting. See attached sketch.

Also, although I've *heard* this, I cannot find anywhere where it is written: that the L shape cross section (web at edge) is typically reserved for concrete mounting of a vessel, while I-sections (web at center) is typically reserved for steel (beam) mounting of a vessel (with bolts on either side of the web).

Can anyone share their experience and/or common industry practice, or perhaps even point me to a specification, code, or other resource that would be beneficial?

Thanks!

 

[SnTMan]

hull6851, I have no refs for you, but, all things being equal, web at the edge will give a higher moment of inertia for the support, usually yielding a lighter support (less depth, less ribs, lighter plate, etc., and yes, less welding) than if the web were in the center.

Most fabricators will therefore default to this design without further consideration. As you say, center web may have advantages for mounting on steel, but difficult to say without input from the buyer.

Ideally, support design is part of the drawing review process.

Regards,

Mike

The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[racookpe1978]

It depends a little bit too on the expected stress movement of tank/PV and support: Fab'ed cold, then heats up, but doesn't get colder. Fab'ed at room temperature, stays near room temperature (outdoor weather, mild temperatures). Fab'ed at room temperature, heats up a lot. Fab'ed at room temperature, shipped and installed in very cold climates, etc.

I've seen most use a center flange web assembly, but not very many I-beams or WF shapes.

 

[dhengr]

Hull6851:
I think your instincts are right on the money, I like the symmetrical cross section for the saddle and the more direct load path with the webs lining up. That saddle might be lighter, as long as you account for tank bearing stresses on the saddle and for web crippling. You are missing a top flange on your “L” shaped saddle (top sketch). Can you make that web connection directly to the vessel shell without a top flange? Since it is unsymmetrical it will not take lateral loads, those from tank expansion longitudinally, as well as a more symmetrical section might. I’m not sure I understand how this unsymmetrical arrangement leads to a larger moment of inertia as suggested, rather it leads to some torsional loading which comes with unsymmetrical shapes loaded in bending. Note also, that the “L” as you show it must span alone, with a bit of help from the WF flange tip cantilever bending, whereas your preferred arrangement has two WF type shapes acting in unison to carry the bending load. As for the WF shaped saddle requiring more welding than the “L” shape, neither of them would likely need full pen. welds, using good welding design and practice, and they should both be welded on both sides of the web, at the flanges, so what savings? You should not leave the root of welds like that unprotected and able to bend across the root pass, puting the root in tension. As for stiffeners, you don’t show any, and they may not be needed on both sides of the WF shape if the designer really understands their reason for being.

 

[hull6851]

Thank you, all. Good discussion.

Support design is part of our drawing review process with our vendor's drawings. The disagreement was internally with a colleague as to whether I should be requesting center or edge webs in our supports and what the cost implications were, if any.
Yes, my sketch was done hastily and I didn't appropriately represent the top flange of the section where the saddle connects to the vessel.

Thanks again!

 

[SnTMan]

dhengr, higher MOI for the "L" design results from distance between the centroid of the composite shape and the centroid of the web, the web then contributing to the higher MOI. For a symmetrical design (web on center) the web is on the centroid, thus contributes less to the MOI. Just parallel axis theorem stuff.

My experience is with shell & tube exchangers, for which the governing axial support load is often bundle pull. The axial load is considered to be resisted by a single (fixed) support, as the opposite support usually has slotted bolt holes (floating). The asymmetric supports are widely used in this class of work.

Not shown in the OP's sketch (as noted) are one or more ribs perpendicular to the web and a pad welded between the vessel and the support proper.

Having said all that, if the customer expresses a preference for a symmetrical design, he gets it

Regards,

Mike

The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[r6155]

See PIP VEFV1106 VESSEL, HORIZONTAL SADDLES SUPPORTED ON STRUCTURAL STEEL

Regards
r6155

 

[hull6851]

r6155, thank you for pointing me to the PIP VEFV1106 (which is part of document VEFV1100, if you are trying to find it on IHS or PIP website). VEFV1105 shows details for Horizontal Vessel Saddles Supported on Concrete and VEFV1106 shows details for Horizontal Vessel Saddles Supported on Structural Steel. The summary of the PIP drawings is that horizontal vessels on concrete use a "L" shape saddle design and horizontal vessels on structural steel use a "I" shape design, as I suspected initially.

Sorry for such a delay in responding. We put this issue to rest months ago, only for it to resurface when someone else's opinion was interjected recently. I had to remind myself of the discussion and wanted to summarize the findings and the PIP recommendation.

Thanks again, all.