Staad - Water tank

[ak.t]

Dear All,

What does it mean "need to check the design of tank manually as per the first principles as in staad it concentrates the moments according to nodal deflections.)
Kindly explain.

Thank you,

 

[JoshPlumSE]

I think you need to explain your question better. Where does that quote come from? Context is important.

I suspect what the quote is trying to say is that "allowable stress" for a tank is not something that's calculated by STAAD. That the analysis may give you stresses, but that you need to determine the allowable stress or capacity based on yield or buckling and that the program will not do this for you.

 

[ak.t]

I did analysis of UG rectangular water tank using STAAD. Thus i got comment from reviewer. I am confused. Can't we rely on STAAD then?

 

[JoshPlumSE]

There are limitations to the type of analysis that STAAD does. I'm not a STAAD expert, but I believe the 2nd order effects are NOT considered for plate elements. So, any buckling or P-Delta of the walls of the tank are not accounted for.

Note: Even if the plate elements handle 2nd order effects, chances are this is only related to perfectly elastic materials and would not be sufficient to give you an adequate guarantee of safety.

 

[JoshPlumSE]

Caveat -

I mentioned that I am not an expert in STAAD. But, I failed to mention that I current work for one of their competitors (CSi) and used to work for another competitor (RISA). So, there is potential for bias in my opinions or statements of STAAD. While I have NOT intentionally stated anything incorrect, I want you to understand that I'm not an unbiased person in matters of structural analysis.

 

[ak.t]

It is okay. Ur response itself highly appreciable. I use CSI-ETABS & SAFE too. Can i design the water tank using ETABS then?
Do u have any guidelines for the same? Thank you.

 

[WARose]

There are limitations to the type of analysis that STAAD does. I'm not a STAAD expert, but I believe the 2nd order effects are NOT considered for plate elements. So, any buckling or P-Delta of the walls of the tank are not accounted for.

That's not accurate. I've used STAAD for P-Delta effects and it does consider those forces for plate elements. You do haver to enter the command however. (I.e. for P-Delta.)

 

[JoshPlumSE]

WARose -

That's good to know. That actually stirs up a memory from back in my RISA days. I remember that some competitor did P-Delta for plates and I wondered how it was done. Very easy for linear elements. Whether you're doing secondary shears, or modifying the stiffness matrix, this is pretty simple.... for column elements. That's because your axial force is in a single direction and your shape functions are simple / predictable.

For plate elements, you've got a 2-way axial force problem in addition to a more complex shape function.

My thoughts on this (for STAAD):
1) I'd want to do some manual verification of this effect to demonstrate that it actually works properly. Simple stuff:
a) A comparison of a single rectangular column subject to varying axial and lateral forces versus a similarly loaded column that is built up from plate elements.
b) A similar problem with a HSS circular or rectangular tube.
2) Even if this works perfectly, I'm pretty certain it's still a perfectly Elastic analysis. This may not be appropriate for a tank. Though, as I think about it, the more "slender" the elements, the more likely an elastic analysis is to be correct.
3) I wonder if STAAD can do an eigenvalue buckling analysis? My guess is that it can. This may give you a better idea of the elastic buckling loads.... to which you would have to add some safety factors.
4) I still think the proper thing to do is likely to follow some industry standard that gives you hand calculation methods for developing an allowable stress. Granted, my thinking may be a bit dated.... based on a "pressure vessels" course I took in the late 1990's.

 

[JoshPlumSE]

AkshikaS -

I just played around with ETABS plate / wall elements and the geometric non-linear analysis is taken into account for them. Though, it is interesting to test it with both in-plane and out-of plane behavior. That being said, it is out-of-plane behavior that you're primarily interested in.

Note: Even if you do this in ETABS, you still have the same issue about elastic buckling vs real buckling. And, I still tend to think that the stresses you get from your analysis should probably be compared to a hand calc of "allowable stress" based on an appropriate code. Relying on an analysis for this can work too. But, I'm more comfortable with this when I know what the "correct answer" is based on code provisions or such.

 

[ak.t]

Dear JoshPlumSE,

I did check with PCA co-efficient method too. The moments are within limits. But still i want to understand the comment I received in the right way. Here it is "check tank manually and to check as per the first principles as in staad it concentrates the moments according to nodal deflections to assure that the moments considered are ok"

Dear WARose,
I attached here the staad model. Do i require to consider P-delta for this? Kindly lemme know.

Thank you

 

[WARose]

Dear WARose,
I attached here the staad model. Do i require to consider P-delta for this? Kindly lemme know.

Offhand I'd say "no". (With a 5 inch wall, approx. 5' high.) But it would also depend on how much this thing moves under loading. (I didn't run it.)

That aside, you could probably check the allowable axial load via the empirical method (for walls) in ACI.

 

[ak.t]

Thank you, WARose.