Columns supporting vertical cylindrical tank?

[agfroese]

Does anyone know of any reference material for the design of columns supporting vertical cylindrical tanks (or hopper bottom tanks). I have the reference "Design of steel bins for storage of bulk solids" by Gaylord & Gaylord, but they only discuss the columns briefly and do not deal with any unbraced columns.

I'm looking at at tank/bin with a hopper bottom, and I want to analyze the bin with only using radial bracing from the discharge (ie. no x-bracing or anything along the perimeter of the column circle)...I can do this in RISA but I can't find any way to back up the work with hand caclulation checks...

Anyone have any ideas?

Picture attached:

http://files.engineering.com/getfil...de3-95f0-7f87d1db1983&file=risa_hopperbin.png

 

[weab]

All bin references that I have come across only deal with the bin design portion. I would think that perhaps references on water tower design may address this. Water towers columns may be designed similarly. I've never designed one so I can't help you on this.

I will say that i'm not sure that, in your sketch, you have accomplished anything. I don't think that you have created an effective diaphragm at the bracing level. Even if you did, I don't think that you decreased the column cantilever length. You may have only assured that all columns deflect equally.

 

[JStephen]

Water tower design tends to be proprietary and not published. However, there is some information in the various pressure vessel handbooks, typically more applicable to shorter columns.

 

[agfroese]

weab I know what you're saying, this radial bracing will not provide much support for the columns. This system of bracing we usually only use in areas of low wind pressure, and very low ground motion. Generally we introduce more bracing along the perimeter of the columns in horizontal and diagonal fashions. I've just been trying to develop a way of checking the computer analysis results with some hand calculations and I can't seem to develop anything on my own since its indeterminate, so I've been trying to see if there is any research or reference material on this...

 

[BAretired]

The type of failure which you must guard against is torsional buckling of the structure. There is nothing in your sketch which prevents the structure from collapsing in a torsional mode. Radial bracing is useless in this regard. This has been observed many times in similar structures which have failed in torsional buckling.

BA

 

[saplanti]

You had better provide (compression-tension or cross) braces between columns to provide stability under lateral and/or torsional loads on the bin. Radial bracing alone, as BAretired indicated, does not provide stability to the bin, it will only shift the colleps mechanism slightly.

Depending on the code you need to comply, you may have wind, seismic or lateral stability loads (I would add some kind of accidental torsion) to apply onto the bin. I guess you will come to the same conclusion after the analysis.

Regards,

Ibrahim Demir

 

[ash060]

I agree with BA if you get any differential ground motion or wind pressures the bin will want to spin like a top.

You could fix the bases and make the columns larger and try doing a cantilever column system. I would probably fix the bases regardless with your condition.

 

[JStephen]

Unbraced legs can in fact be used. They would need to extend up the side of the silo far enough to distribute stresses in the silo (including bending loads) or have additional structure built up at the cone-shell junction for that stiffness. If the legs are very long, it would generally be worthwhile to cross-brace them.

 

[SAIL3]

It is an interesting concept. I would not feel confident
designing it this way, though, for the following reasons:
Column Bracing:
1. Using a common brace point for all the columns would
eliminate any redundency in the design.
2. The cone hopper would certainly be stiff enough to
provide bracing for the columns but I feel that to
activate that stiffeness one may require too much lateral
displacement of the cone and cast doubt on it's
effectiveness as a brace point.(It's not a hard point)
3. Any deflection of the cone from it's own loading would
be transferred to the columns with unintended
consequences.
Lateral Loading:
1. As pointed out by the other comments,it does not provide
a mechanism for transferring lateral loads to grade and
therefore would be unstable.
I have used a ring girder, built into the bottom of the
tank and made a moment conn. of the columns to the girder.
Better still, as was pointed out, run the columns up the side of the tank a certain distance and accomplish the same
thing.
Probably, the simplest and most effective is to use x-bracing.
Still, it is an interesting concept and brings up an
interesting question. Is using a common brace point for all
the columns a wise idea.

 

[BAretired]

The structure, as depicted, is unstable under gravity load alone if each column is pinned top and bottom. There is no need for lateral or torsional loading to cause collapse. Each column is constrained to move in a direction tangential to the reservoir which gives the appearance of a torsional failure.

If the columns are restrained at the top, bottom or both, the structure can be made stable.

BA

 

[agfroese]

Yes, the columns are restrained at the top. Welded to Ring girder, with gussets as well. I have found similar results, and agree that a pin connection at the top of the column would make the model unstable, yes, but I do have restraints at the top of the columns.

The base plate is anchored to concrete with expansion bolts, but I still have always assumed a pin connection.

 

[gobsmacked]

The choice of whether to use radial and/or circumferential braces would depend on the orientation of the column section, as well as length and radius of gyration etc.

 

[saplanti]

Have you considered the differantial vertical displacement at/between the column bases?

What type of foundation do you use? Does this foundation cause large differantial displacement under the base plates? If so how do you overcome that? If you did not consider any of these you must have very thick slab that connection the columns, and under the slab is well prepared for the uniform load and displacement under the column bases by geothechnical and civil engineers.
I had an experience that the tank was leaning one side during the filling for hydrotest due to the lack of geotechnical investigation.

I would still not connect the cone hopper lower end to the columns due to the expected load in the hopper, thermal expansion/stresses between the hopper and columns. Depending on the connectivity of the braces onto columns and the hopper, columns may be under lateral loads as well. Does your model considers all these actions may be required during the operation with environmental loads and at rest?

Hope it helps.

Regards,

Ibrahim Demir

 

[BAretired]

If the columns are to be considered hinged at the base and connected to a ring beam at the top, then it is not difficult to determine the buckling capacity of the system under purely gravity loading. Under wind loading or seismic loading, it is a little more difficult but not out of reach.

I am sorry, but I did not follow the argument of saplanti. Perhaps he could elucidate. Why would anyone consider large differential vertical displacements between columns? It is normally assumed that the foundation is unyielding. If not, then I would agree that the problem becomes much more complicated.

BA

 

[gobsmacked]

RE: Saplanti's point. If you have 4 columns and the load is distributed 25% to each one, then, in the case of differential settlement of one of the column bases, the distribution of loads becomes 0, 50, 0, 50%. Even if you were to assume an "unyielding" foundation, there must be some consideration given to the waywardness of this assumption. I don't think that the load and material strength factors of safety necessarily cover this point.

 

[saplanti]

I have just wanted to say about the possible differantial settlement under the column bases in case they are on seperate bases or on thin slab with seperate bases which may be additional problem to cause differantial displacements under the column bases.

Although the columns are on the ring foundation with suffient thickness, if the geotechical work is not done properly and the soil sattles unexpectedly you still may have problem by loading the columns unevenly and with the second order effect due to uneven settlement. This can be seen clearly during the test or during the loading the tank. If the settlement continues and does not correct itself under the loading you may really have problem.

You had better get foundation data from the civil/geotechnical department just for the confirmation that you have no problem with the foundation settlement/uneven settlement/differantial settlement.

Hope it clarifies.

Ibrahim Demir

 

[BAretired]

saplanti is correct. Uneven settlement of the foundations could give rise to variable loading on the columns, so foundation data from the geotechnical consultant is important.

The OP stated:

I want to analyze the bin with only using radial bracing from the discharge (ie. no x-bracing or anything along the perimeter of the column circle).

You cannot do that. With only radial bracing, the structure is a mechanism. A ring beam however, moment connected to the top of each column constitutes circumferential bracing and prevents the structure from being a mechanism.

Whether or not it is advisable is another matter.

BA