Water Tank - Roof Analysis Under Out of Plane Loading - Methods (AWWA D100)

[SGMIPE]

I am concerned about the localized out of plane loading into tank roofs from antenna wind loads (multiple 700-800 lbs. point loads). The tank roof doesn't have any rafters, stiffeners or supporting structure. The roof diameter is about 27' and plate thickness is 3/16". AWWA D100-21 is not very specific about roof plates with out of plane loadings. Method 1 (Membrane Analysis Technique)From section 3.4.3.1 seems the closest to evaluate the tank roof at the area in question. The other alternative approach would be to look at the localized stresses and compare them to Von-Mises stresses. What would be the best way to evaluate Water Tank Roofs subjected to out of plane loading?

 

[le99]

How would you design a large roof hatch with steel plate covering? THe design approach should be the same.

 

[JStephen]

The first approach is to move any significant loads to the shell to avoid the issue.
I think Section 3.4.3.1 is written for buckling due to uniform loads, not localized loads, and won't really be useful for point loadings.
Consider using WRC 107/297 approaches treating the cone as an equivalent cylinder if the t/R ratio falls within the range covered.
Or using the line-load approach in Bednar's pressure vessel handbook.

 

[SGMIPE]

Thank you for the response JStephen! Moving the loading is not an option in this situation. I had a feeling that the allowable local buckling compressive stress was meant for designing shells subject to uniform loading from gravity and winds. But given that there is no specific sections under AWWA D100-11 or D100-21 for out of plane I was inclined to use that section to evaluate. Reviewing the line-load approach in Bednar's pressure vessel handbook seems a lot more applicable to the loading situation. Some of the other newer AWWA codes (AWWA D107-16) allow the use of Von-Mises stresses to evaluate tank shells, however, this did not get carried over to AWWA D100-21.I will compare the results of using the line-load approach recommended with the other methods and see where they stand.