Vertical vessels, skirt-supported

[JOHNPIA2]

Does anyone know a design method to check the stresses on the dished bottom head to skirt connection(Other than FEA, of course), for cases where the skirt diameter is less than that of the vessel head?

 

[Tankman650]

A very conservative method can be used when e is very small:

e = horizontal distance between cylinder shell center line and skirt cylinder center line (in.)
v = skirt vertival load #/in
t= head thickness in.
m= e*v ,in-#/in unit moment

bending stress in head fb= 6*m/t^2

add membrane stress to bending and check stress limits per Div 2.

For large vales of e , I recommend fea or elastic analysis.

Of course, this does not account for thermal.

 

[JOHNPIA2]

Thanks tankman650; is e half of the difference in diameters between shell and skirt? And what could be considered as a large value for e? How about the method shown in API 620, fig. 3-4?

 

[Tankman650]

Your understnding of e is correct.

You must use your judgement for the value of e. I tend to keep the bending stresses low to avoid the problem with secondary stress levels. If the skirt dia gets very much smaller than the cylinder dia. the skirt will punch on the head similar to a nozzle load.

API 620 does not address this loading in the free body diagram. The attachment of a support to the shell with overhanging portions is discussed in API 620 3.13.5.4. This paragraph refers you to papers given as a reference to ZICKS paper on support for a horizontal tank.