Wind Drift on a tank

[EfficientPuppy]

Any ideas on how to find wind drift on tank analytically without the use of an FEA? I have searched through API 650, API 620, and AWWA D100. I'm pretty sure its not in there. I've been doing tanks for a while, but you never know maybe I've missed it. I'm ready to break out ASME VIII, but I doubtful it would be in there.
Is an FEA the only why to figure this out?

 

[JStephen]

Just treat the tank as an assemblage of vertical beams of varying stiffness.
Shear deformation may be the major contributor to deflection rather than beam bending and both should be considered.
This can be carried out by spreadsheet using normal beam equations and shear deflection for each shell course.
Stiffness of the soil below can also be considered if desired.
In some cases, P-delta loading will contribute to the loading.
See Bednar's pressure vessel handbook for some information on calculating vibration due to vortex shedding in vessels.
For vibration issues in large-diameter loaded tanks, also see Veletsos' method of calculating effective mass of contents.
I'm not aware of any reason to calculate wind drift, other than to calculate natural frequencies. See Eq. 15.4-6 in ASCE 7-16 for that. For that purpose, it is simpler to assume shell and content masses are lumped at intervals rather than distributed.

 

[EfficientPuppy]

For for my own understand let me rephrase your suggestion in my own words.

Do you suggest that I model the wind drift by stacking a series of round columns on top of each other set to the diameter of the tank, each column representing the thickness of their course. Then I use that to determine deflection using both bending and shear deformation, Using P-delta loading as applicable?

This is fairly straight forward, I just wonder if that is accurate enough. I worry about the D/t ratio being so large that I worry that the tank will squish on the windward/leeward and bulge on the sides. I'm sure there is a better word to describe this behavior, squish and bulge seem like the wrong terms here. Is there a way to account for that analytically?

Vortex shedding is not what we are looking to evaluate here, there is concern from their side that the tank will contact nearby structures during a wind event.

 

[JStephen]

Where the deflection/vibration issues normally come up is on tall slender tanks, so the beam approach is reasonable there.
With more normal-size tanks, calculated wind deflection will be "small".
Gross buckling ("blow in") of the tank shell during wind loading is a design check in both API and AWWA codes. That is a concern primarily if the tank is empty or mostly empty. Elastic compressive buckling is another design check, although not normally a controlling factor for wind.
If the tank is full or mostly full, hydrostatic pressure will tend to keep it rounded out.
For a fixed-roof tank, the roof acts as a stiffener to keep the top round.
For an open-top tank, the tanks are furnished with a wind girder to keep the top relatively round.
I am not aware of any information on out-of-cylinder deflections during wind loading.
I think it was the paper "McGrath, R. V. "Stability of API standard 650 tank shells." (Correction, it is J. H. Adams, "A Study of Wind Girder Requirements for Large API 650 Floating Roof Tanks", API Refining, 197), that speculates on the origin of the wind-girder equations used in API-650 and AWWA D100. The assumed loading used in there could presumably be used to calculate corresponding deflections in a circular ring. But the loading itself is "assumed", so I don't know how meaningful the exercise would be.
You could calculate wind pressures via ASCE 7 or other sources and plug into FEM analysis and perhaps have a better guess.
I suspect the maximum deflection would be a vibration problem rather than a static deflection, though.