Supported Tank Roof General Queries

[zsa]

Hi all,

I have certain queries that I hope that the good people here will help clear out so that I may better understand

1) API 650 5.10.4.3 states that 'Rafters shall be designed for the dead load of the rafters and roof plates with
the compression flange of the rafter considered as receiving no lateral support from the roof plates and shall be laterally braced if necessary '

Couple of questions here: - What exactly does the term 'compression flange' here refer to? I am assuming that it means that this is the 'flange' of the I-beam that is in contact with the roof plate and is under load/live load from contact with the roof plates and is hence under compression? Am i correct in assuming this or not even close?


The clause goes on to say ' When considering additional dead loads or live loads, the rafters in direct
contact with the roof plates applying the loading to the rafters may be considered as receiving adequate lateral
support from the friction between the roof plates and the compression flanges of the rafters, with the following
exceptions:'

Maybe i dont understand this part because I dont fully comprehend the first part of that clause but what the clause is saying doesn't quite register with me? Like I dont know how to interpret this?


-Lastly, Clause 5.10.2.7 states that 'For all types of roofs, the plates may be stiffened by sections welded to the plates. Refer to 5.10.2.3 for requirements for supported cone roofs'. Is there any criteria through which i can check whether the roof plates and/ or roof structure(rafters) needs to be 'laterally braced' by angles or any other structural member? 5.10.4.3 states that the structure may be braced if necessary but doesn't outline what method to use in order to reach that conclusion. Colleagues at work say that they just follow procedures based on similar past projects that they have worked with- something I am slightly uneasy about because (a) it sounds like a tad bit negligent to not go through with actual analyses and (b) I dont have enough of said experience to eyeball this kinda thing (and my good judgement tells me to never do this regardless of 'experience')

Terribly sorry for the tirade but I just want to further my understanding by asking people who have been involved with the industry because as informative as books are, sometimes you gotta turn to people to understand better.

I actually had a couple of more questions but I dont want to turn this into a post so long that no one would bother reading. And also, I am hoping that the answers to these questions will help clarify the overall picture.

Thanks!

 

[JStephen]

The compression in the first clause is compression due to bending in the rafter. For a simply-supported beam, that's the top flange.

The check for lateral bracing will come from the AISC steel codes or other equivalent design codes referenced in the standard.

 

[zsa]

Thanks for the reply, JStephen.

Building on my previous questions, what design conditions might lead to a requirement of the roof plates themselves requiring stiffening a la 5.10.2.7?

Also, if I have a rafter supported cone roof where the rafters are attached to the top of the shell, how do i perform the buckling analysis of the shell due to the increased load? As far as i can tell, the only 'buckling'we check for is for wind overturning stability (ie the transformed height calculation to check for need of intermediate wind girder) and maybe in seismic analysis. How do i/do i need to check the buckling due to the additional weight of a thickened roof plate along with the roof structure? I was thinking about maybe modelling one shell plate width of area extending all the way up to the roof as a cantilever and use the weights to check for induced bending moments and magnitude of deflection? (That seems like a bit of an oversimplification to me)

 

[dig1]

Depending on the roof configuration, wind may cause suction on the roof and that would place the bottom flange of the rafter in compression. You would need to look if the roof plates are attached to the rafters (if not there is no uplift on the rafter as the wind would lift the plate up), if they are the net load, etc.

 

[JStephen]

For a supported cone roof designed within the maximum rafter spacing, or for self-supporting cone, umbrella, or dome roofs per the standard cone/umbrella design, you don't normally need to stiffen the roof plate itself, or check for required stiffening. However, you can build other roof types that do use stiffeners welded to the plate- do a google image search for "stiffened umbrella tank roof" and you'll find some examples.

Generally, it isn't necessary to check for buckling in the shell due to rafter loads. For compression, API-620 does give an equation for allowable compression in a thin cylinder that could be used. You'd assume the load spreads at 30 or 45 degrees out from the top of the attachment and check compression at the bottom of the attachment, so you'd have several inches of shell available to resist the load. And if necessary, the attachment could be made longer.

On the rafter compression, either reversed loading or putting the rafters on the top of the roof plate or using cantilever supports for the rafters can put an unsupported flange in compression.

 

[zsa]

Thanks for the replies dig 1 and JStephen.

dig1: why would wind pressure place the bottom flange under compression? shouldn't wind loading place the top flange under compression (for rafters under the roof plate)?

JStephen: I understand that the top flange for a simply supported beam is under compression due to its own weight and any perpendicular loads acting on it. However, what do you mean by reversed loading and 'unsupported flange'? Does reversed load imply wind (external) pressure as opposed to internal pressure? And by 'unsupported' flange do you mean laterally unbraced flange or the flange that is not in contact with the roof?

AFAIK, an API compliant roof is only supposed to act as a membrane transferring loads to the shell (and roof structure, if any) and through the shell, on to the foundation, without bearing any load itself. What about the case in which rafters are welded to the roof plate(either under the roof plates or on top)? what kind of stability analysis do i need to do for that case?