Minimum Skirt Thickness et al 1

[shopper1732]

It is my belief that, per code, skirts on vertical pressure vessels must have a minimum thickness equal to the vessel wall. True?
Also, that unless thicker than vessel wall, OD of skirt must match OD of vessel. True?

 

[codeeng]

The skirt thickness is calculated based on the vessel weight, wind, seismic plus any other external loading. Generally it should be 1/4" minimum just from a fabrication point of view. Skirts less than 3/4" thick normally have the OD match OD vessel. Over 3/4" have centerlines match.

 

[jte]

shopper-

False to both. The code does not specify design details for skirts.

There's no physical reason a vessel's skirt thickness should be governed by the internal pressure in the shell. A little common sense would be good here. As for matching OD's vs matching centerlines - again, just a question of load transfer and induced moment. For heavy vessels with relatively thin skirts, it makes sense to provide a load path (mid wall of shell to mid wall of skirt) without a moment arm.

The guidelines codeeng provided are reasonable. You'll find most company standards have something close to what he stated.

jt

jt

 

[carthago]

A skirt or a saddle is a structural item.

Thickness should be checked as follows:

Skirt bending stress=
[Weight/Area] +/- [Overturning Moment/ Z]

Thus the [ +/-] means 2 checks have to be performed, one on each side of the skirt, for at least these cases:

Check for wind.
Check for seismic.
Check for empty vessel.
Check for vessel operating).
Check for vessel under hydrotest (in which case , the
wind is lower).


Notations:

Overturning moment = due to wind and due to seismic.
Both not acting concurently.

Allowable to be used are:
S_hot from Div. II , for the tensile case
B for compressive stress

Area= 3.14* (R_av) * t = skirt area
Z = 3.14* t* (R_av)^3 = Skirt modulus of inertia

with:
t= (Skirt thickness) - (Skirt tolerance on thickness plate)
R_av = skirt average radius.

Make sure that there enough material compensation and that skirt access holes are fully compensated.

-----------------------
Note also that skirt design temperature is < then vessel design temperature.This means that S and B and the Young modulus (used for calculation of deflection) of the skirt are a bit higher than the ones used for the vessel.

--------------

It happened, that for tall columns and in order to avoid vortex, that the column own frequency has to be improved (e.g. increased), in this case a THICKER skirt than the one given above is necessary.

 

[carthago]

Omission & Correction:
Area = TWO *3.14* (R_av) * t = skirt area

 

[roca]

On thick wall vessels / columns as a general rule of thumb the skirt thickness is generally taken as 0.25 x Shell Thickness. This can of course be reduced if proven by calculation. Centreline of skirt and shell plates should be in line.

I just had a quote from a vendor who specified a skirt thickness of 55 mm - the same as the shell wall thickness. A bit OTT I think

 

[Shmulik]

Just adding my 2 pennies …
Be aware of restriction marked in UG-4(b): material for nonpressure parts, such as skirts, supports, baffles …. , need not conform to the spec. for the material to which they are attached …. The allowable stress values for material not identified in accordance with UG-93 shall not exceed 80% of the max. allowable stress value permitted for similar material in Subsection C.