metal storage tank design

[cmbyrd77]

I am currently working on a metal storage tank design for Glycol (density a little less than water), and I've seen some previous designs with tanks stiffenend with single angles. Is there any literature out there to figure out the "ring" tension and distribution coeffecients. I'm assuming that the coeffecients in the PCA book on Rectangular Concrete Tanks will not suffice.

 

[JedClampett]

What's the size, height and shape of your tank?
If it's round, you could use AWWA D100 or API design specifications. If it's rectangular, you can model it using finite elements or design it by assuming the plates span vertically to the stiffeners and following the loads to the corners. But be prepared for a lot of stiffeners and very thick plates.

 

[cmbyrd77]

Thank you JedClampett. It is a rectangular tank, and I did think about modeling it in a Finite Element Analysis program. So would you treat the plate and stiffener as a composite section and creat the shape in like RISASection and then model it based off of those section properties? Would you model the full tank or perhaps just a 1 ft. strip?

 

[theonlynamenottaken]

The literature I'd recommend for tank/vessel design beyond the codes (ASME, AWWA, API) are:
"Structural Analysis and Design of Process Equipment," Jawad & Farr
"Process Equipment Design," Brownell & Young
"Tubular Steel Structures," Troitsky
"Design of Welded Steel Structures," Blodgett

 

[theonlynamenottaken]

The common approach for rectangular tank panels is to analyze them as strips, and then designs the "rings" using the strip reactions. You do use a composite section for your stiffeners, using an effective width of shell plate. Going completely from memory I believe you can use ~6t of the shell plate on either side of a stiffener. The books I recommended and API all give guidance on stiffeners and their composite action.

 

[cmbyrd77]

I didn't even think to look in my Blodgett. Thank you.

 

[JStephen]

If the tank is not round, give some serious thought to making it round. You can considerably simplify the design, the construction and make it cheaper, and meet the various codes that way.

Also, take a good look at what, if any, codes this needs to follow, particularly fire codes; whether it needs to be double-walled, etc.

If the tank is small (say, 5,000 gallons or less), it likely isn't worthwhile to do a finite element analysis. Run stiffeners vertically or horizontally, depending on the tank dimensions, treat the plate in between as little beams (or rectangular plates per Roark), transfer that load to the stiffeners, treat part of the wall as the flange of the stiffener. It's very easy to spend $5 of your time saving $2 worth of material on something like this.

Details of rectangular tanks aren't covered by the various codes, so they aren't standardized, and you could have a lot of variation in allowable stress, design assumptions, and detailing.

If you'll be having a fabricator construct the tank for you, they could quite likely design it as well- check and see- but they'd have the same issues to deal with.

 

[cmbyrd77]

Thank you all for your help. It appears that Blodgett pulls through again. What a book! I do have a question regarding the stress they use in their example of 20,000 psi. I know the book was copyrighted in 1966, and I'm not sure than if that is a reduced stress from 36 ksi steel or if they only had 20 ksi steel back then and it was not reduced. Any help would be greatly appreciated.

 

[JStephen]

I would assume 20,000 psi was taken as the allowable stress for material with a higher yield strength. Depending on the tank standard used, A36 steel will be designed for 12,750-23,200 psi in current tank standards.

 

[SAIL3]

I usually run the stiffeners horizontally with good continuity at the connection of stiffener to stiffener at the corners...a band of stiffeners all the way round increases the integrity of the tank...if feasible a round tank is always more economical....if glycol is a hazardous material, it would be another reason to go round...20000psi is the allowable stress, at least for non-hazadardous...yes, Blodget is a treasure trove of usefull and practical info on design...another ref;

AISI
"Useful information on the design of Bins and Silos" by John R. Buzek

 

[cmbyrd77]

So I'm guessing then that they came up with 20,000 psi from 0.667x30,000 psi?

 

[JStephen]

In my 5th Edition of the AISC Manual from 1947, the allowable stress for tension in structural steel is 20,000 psi, same for tensile bending in structural steel. That's for A7, but is just given as a stress, not a function of yield stress. Elsewhere, they show the Elastic Limit of structural steel for bridges, buildings, and ships as 33,000 psi, not sure if that matches up to A7. Anyway, Blodgett probably used it as it was a nice round number that matched the typical steel stresses used. Looks like that was bumped up to 22,000 psi for A36 in 1960. So maybe it was 0.6Fy in either case.

 

[cmbyrd77]

As far as the stiffeners go, there isn't much in Blodgett about how to check those. I realize there is a type of "ring tension" happening, but how would one go about calculating that tension. Concrete tanks are much easier.