Very large sheds to design - design philosophy for stability?

[clarke1973]

Hello all,

I have four large warehouses to design - each 300m long x 90m wide x 15m high to eaves (5 deg roof slope). Construction will be standard lightweight steel clad frames. Large, uninterrupted internal spaces are required so I need to maximize internal column spacing. I am looking at different options for the roof framing, but will most likely be four-span frames (22.5m span) at say 7m centers along the length of the building. Transfer frames (truss or beams) supporting the main frames will span longitudinally between internal columns, giving a column spacing of 21m.

My question is regarding stability. Generally with this type of structure I would obviously portalise the frames, but I am thinking it may be more efficient to adopt a fully braced system, particularly given the height of the columns. The additional horizontal roof bracing would obviously be additional steel, but would it be offset by the saving in reduced perimeter column sizes and also the beam / column connections (pinned connections opposed to rigid)?
Obviously I can do some preliminary numbers for both options and do a tonnage / cost comparison, but does anyone here have any experience with structures of this size and can offer an opinion on the best approach?

Also, a number of areas will have large span cranes so horizontal deflection limits will need to be tighter than normal.

Thanks

 

[Guest262653]

I think that the main issue that will control your choice may well be the longitudinal span of the horizontal bracing system. If you can segment the 300m long structures in 40 or 50 m units and insert the vertical bracing at these location, the use of a fully braced system might be more adequate, especially with the presence of cranes.

Have you also considered using concrete perimetral frames and internal columns and steel trusses? I've seen this system being used quite efficiently in a couple of large industrial buildings.

 

[paddingtongreen]

I'm old fashioned, I like "squaring bracing" on the bottom chord, that squaring bracing then becomes part of a global bottom chord bracing system. Most of these carry the loads to vertically braced bays at the corners but if I can, I vertically brace the middle bays of the walls because there is greater vertical load in the middle columns than the corner columns to counter uplift, and expansion is outward from the center, not squeezed toward the middle. This expansion point is usually minor and not determining.

Michael.
"Science adjusts its views based on what's observed. Faith is the denial of observation so that belief can be preserved." ~ Tim Minchin

 

[ajh1]

The configuration you describe, while long, is pretty much a typical metal building to that industry. The 15m height is tall but not out of the ordinary. It would sound like your biggest problem is going to be the crane systems. The cranes are going to control the permitted lateral sidesway. One option for that condition if the cranes are heavy would be to make the sidewall columns fixed base to limit the deflections. The other area that will be problematic is your desire to eliminate interior columns by using jack beams. Crane beams spanning 21m are considerably more difficult to design than the equivalent 7m beams if the interior columns are left in place. This is particularly true if the other side of the same crane aisle has the normal 7m spans. Differential deflections between the two sides of the aisle will need to be balanced. I would suggest that where you absolutely need the clearance then by all means, use the jack system, but limit it's use to only those areas where you really need it.
As noted by avscorreia, to implement a fully braced system longitudinally would likely require intermediate frame locations that would serve as anchors for the bracing. Those will be very difficult to control deflection on as now instead of one bay of lateral load, they will be carrying 6-8 bays of load. That is unlikely to be doable with cross bracing somewhere to transfer those loads to the ground. Perhaps some sort of outriggers or buttresses outside of the building if the inside must remain clear for the crane aisles. Probably easier to bite the bullet and design each frame to take its own lateral forces. A small set of braces along each eave in the roof plane could be helpful in balancing out any lateral crane loads that are applied to the frames so that instead of one frame taking the entire lateral load of the crane system, a set of 3 or 5 frames could distribute the load out from the single load application point. Unfortunately, the same concept doesn't work for getting rid of wind loads that will likely be significant on a 15m high structure.

 

[clarke1973]

Thanks for all the responses. Mixed opinions it seems on the best approach, but it sounds like either option is equally practical. There will be limited opportunities for internal vertical bracing bays, perhaps 2 or 3 bays within each shed, so the bracing will be working very hard (horizontal and vertical) and uplift will be an issue. I want to avoid piling but isolated pad foundations might not be suitable for the uplift forces... sounds like I will have to do some modeling on both systems..

thanks

 

[TLHS]

If the cranes are reasonably high loads, I'd keep them on separate structure if at all possible. Firstly, from a redundancy standpoint you don't want someone screwing up crane operations to be able to take the whole building down.

Secondly, it means you can keep deflection limits for the building reasonable and foundation conditions reasonable, and then design the crane as a separate system so you can control your spans and supports differently.

If the crane covers the entire square footage of the building that may not make sense, but it's nice to give separate structural systems a try if it's possible.

 

[dik]

I'd break the 300m into three 100m sections for joints and use x-bracing with double columns at the joints. break the spans into something reasonable for trusses/joists @ 2m approx. and use plastic design for the roof beam design. I'd use elastic beam-column design for the columns. I'd use horiz girts @ 2.1m approx or whatever works well with the first one at 7'+ to clear mandoors and insulated? prefinished metal siding. OH doors are another issue. I'd likely use the roof deck as a diaphragm to transfer lateral loads to x-bracing in both directions and coord x-bracing with openings. You have enough building to mock up a couple of layouts and get pricing from a local steel contractor.

Dik

 

[Agent666]

I would definitely portalise in the 90m direction, I think if you have limited options for vertical bracing along the 300m length then you will struggle with the roof diaphragm. If you are bracing the roof then you need to consider that all rafters need to be designed for considerable axial loads in addition to bending forces. This becomes problematic as effective lengths can be long if your column spacing is 20+m. Investigate tapered sections to get the most bang for your buck, you can preset out a lot of the dead load deflections so frame can ignore this to a degree by altering the fabricated geometry to give an initial deflected shape close to the undeflected original geometry.

In the 300m direction I would brace the external walls and provide a band of bracing in the roof at these locations. Add enough bracing so that bracing forces in the roof and walls and resulting foundation loads are reasonable without heavily influencing the strength sizes of the portals in the other direction or needing ridiculous foundations.

Pay attention to the crane pointers by others, the deflections to find the frames will be governed by these depending on the size of the crane.

I have seen one building years ago that had large external props from the top of the columns to the ground at each frame, so Bracing doesn't have to be internally in the building is for it impacts on layout.

 

[dik]

Because of the height and the width, I would use x-bracing if possible and coordinate it with any openings. I think a portal would be too flexible, and it's easy to design the deck as a diaphragm. I'd also use HSS columns for the interior and W for the exterior columns. Depending on the location of the crane, W columns may be in order for some/all of them.

Dik

 

[TLHS]

I know designing the roof deck as a diaphragm is pretty standard, but I'd at least recommend sitting down and considering what might make them cut the roof open at some point.

I've seen people add giant openings or decide they're going to rip the roof apart to crane something into industrial buildings. With bracing you know that a halfway competent tradesman will know it's important (doesn't mean I haven't seen bracing cut). With diaphragms, it's a lot less obvious. It's the same reason you may want to avoid using the floor slab to tie the foundations together in an industrial building. Someone can come along later to install an equipment foundation or a trench or something and cut critical structure because it's not obvious that it's important.

With a warehouse I expect the roof situation is probably okay, but you never know.

 

[dik]

and cut through your x-bracing... OK as a diaphragm and if they want a skylight, then deal with it at the time...

Dik

 

[hokie66]

I would use the more traditional approach of portal frames in the 90 metre direction and bracing in the other direction. But 300 metres is too long without movement jointing. 3 at 100 metres would be my choice. Not being in a country where roof decking is used for diaphragms at all, I would be prejudiced against that. I would think concealed fastener metal roofing would be the best for watertightness, so would use roof bracing trusses.

 

[dhengr]

Clarke1973:
Regarding bracing, and some column lines with bracing on the interior of the bldg., etc., I think you have to have a serious talk with your client. Hell they all want them to be 300m by 90m, with no interior columns, so they can do anything anyplace, certainly no x-bracing every third column line, they just couldn’t work around that. This really requires some interior usage study and traffic flow study as part of the design. After all, they do want their bldg. to stay standing in a light wind, or EQ or snow loading. The starting point... We all have no problem designing a 20m by 20m shed to withstand reasonable lateral loads. But then, since the 20x20 works, most clients don’t understand that you can’t just stack these 20x20 bays side by side, 3 wide and 20 long, and eliminating all the end walls, all but two, when these end walls provide all the lateral resistance. Nor do they understand that there is just no way to make that 300m long roof diaphragm work. You have to explain that to them, and convince them that you can provide a workable design, with some fixed openings, and with some bracing at every third or forth column line. Additionally, in that large a facility, their insurance company or the local BO and bldg. code may require fire walls at some reasonable intervals in the length of the bldg. A way you might save them some construction time and money is to consider panelizing the roof system on that large a group of bldgs. You build (mass produce) roof panels, 4-5-6 stl. jsts. wide with decking installed on the ground, and lift these panelized roof sections into place, as you move down the bldg. The crane just backs down the length of the bldg. and the stl. jsts. and decking flow in from either side where smaller equipment helps with the panel assembly.

 

[dik]

I'd be looking at an EPDM single ply membrane roof to accommodate roof slopes and control/expansion joints... Prefinished metal, in these environs, is a tad more expensive.

Dik