Supported Tank Roof - Laterally Braced Rafters

[Nick7805]

I'm working on a design for an internal raftering support system for a 120ft diameter API-650 tank. I'm looking at a lateral bracing design for the rafters - I've reviewed some of our old projects, and we've always provided lateral bracing to internal raftering systems by welding 1/4" flat bar to join the tops of two rafters together to prevent buckling of the top flanges. This is done in a full ring, joining every rafter in the tank to an adjacent rafter. My question is, is this method of lateral bracing adequate?

How would seismic loading affect a raftering system braced this way? Could strong wind loading potentially lead to buckling of the unbraced bottom flanges? This tank is being designed for an area where hurricanes and seismic activity are of concern.

I can provide a sketch if my descriptions are unclear.

Thank you.
Nick

 

[IFRs]

These lateral braces have two purposes that I know of.
1) stability during erection, for this case use one or more diagonal sets of bracing creating triangles back to the tank shell, especially for large tanks and multiple rings of columns.
2) weak axis bracing of the rafters when under dead load only and the roof plates do not offer weak axis bracing (API 650 5.10.4.2 and 5.10.4.3), for this case you can calculate the critical length using AISC methods for the spacing of the braces along the rafters. If you find that concentric rings of bracing alone is not enough for stability, it is easy and inexpensive to cross back and forth between a pair (or pairs) of rafters all the way to the tank shell to satisfy your concerns.

I don't normally consider for new tanks wind uplift on the roof plate inducing compression of the rafters, also it is a transitory load. On the other hand I have had to replace cone roof structures damaged by hurricane force winds acting horizontally on the tank shell making the rafters bent and the girders rolled over. If you are designing for extreme conditions, you may need to review unusual load cases.

 

[HTURKAK]

Yes ..please..

Cross bracing is provided in the plane of the roof in general in four bays for the roofs more than 15 metres in diameter. Is this internally column supported roof ? 1/4" flat bar could be satisfactory for strut ring but bracing would need greater size and in general angles used.






Use it up, wear it out;
Make it do, or do without.

NEW ENGLAND MAXIM

 

[JStephen]

The current API-650 has a requirement that rafters be adequately stable without considering support from friction with roof plates when considering the roof plate dead load only. This requirement wasn't always in there, so older designs had less inherent stability than under the current standard. But in either case, you may have situations where the calculations show all is good but experience shows additional support is prudent.
Bracing the top of one rafter to the top of the next doesn't theoretically accomplish anything since both (or all) rafters can go the same direction at the same time. But in practice, it will help. I believe the AISC requirements call for bracing against rotation, not just lateral movement of the compression flange. However, although AISC is referenced in the tank standards, bracing and detailing of roof rafters are not normally as rigorous as would be done with structural steel in other applications.

 

[Nick7805]

HTURKAK, yes it an internally column supported roof design. Some images of the bracing system are attached for clarity. In these images, MK 12 is 1/4"x2" flatbar.

JStephen, could you please elaborate on the case where all rafters move the same way? Is this a lateral buckling case you're referring to? This seems like an unlikely situation to me since every single rafter would have to be overloaded to the point of buckling, correct? Could this problem be alleviated by cross-bracing between members instead of using a flatbar ring? This should prevent rotation.

 

[Geoff13]

Imagine the centre column rotating about its own axis. The current ring of flat bars will not prevent this.

You must add some diagonal flat bar braces from the shell to the current ring. Probably a single X would be sufficient, spanning 4 or 5 rafters.

 

[HTURKAK]

Err. Nick7805 (Mechanical)(OP),

I looked to your sketch and apparently needs some improvement . My points ,

- The roof internally center one column supported. In order to get more economical design, you may consider circumferential columns arranged around center column. The ring of columns would circumferentially be linked by girders which will support radial rafters.

- The maximum allowed spacing between rafters is around 1.7m ( 5.5 ft). In your case , The spacing near the perimeter of shell =120*Π/44=8.6 ft. You need to increase the no. of rafters for the first ring or increase the roof plate thk.

- You may consider truss supported roof without columns for 120ft dia.

- The preliminary proposed roof bracing could be as shown below,

Use it up, wear it out;
Make it do, or do without.

NEW ENGLAND MAXIM