TomBarsh
Structural
- Jun 20, 2002
- 1,003
In a recent article I discussed some common reasons why COMPRESS may increase the shell thickness to a value that may not be apparent at first (see
For example, the required thickness of a cylindrical shell might be, say, 0.368" but COMPRESS enforces the nominal thickness to be 1/2" (0.50"
rather than the 3/8" (0.375" that one might expect based on 0.368" required thickness.
Of course, these cases involve situations where the allowable stress (and required thickness) is a function of the nominal thickness. This is not encountered for tensile stress but may be encountered for compressive stress or buckling considerations.
There is also another situation in which COMPRESS may increase the thickness over what might be expected.
The Wind Code dialog provides an input to limit the deflection due to wind. By default this is 6" per 100' (5 mm per meter). Limiting the deflection may be desirable for process considerations. For example, a tall process tower with many trays may encounter problems if the vessel (and trays) deflect too much; this is not necessarily true for all cases, only the process designer knows for sure. Refractory lining might also require such a limitation on deflection. The designer can also "switch off" the wind deflection limitation on the Wind Code dialog.
If the nominal shell thicknesses determined on the basis of required thickness for circumferential stress (internal or external pressure) and longitudinal stress (weight, pressure, wind, and seismic) is such that the deflection due to wind exceeds the allowable limit, then COMPRESS will increase the thickness of the skirt by one "standard thickness increment" and recheck the deflection. If still excessive, then the thickness of the lowest shell course is increased by one "standard thickness increment" and the deflection is rechecked. This process is repeated until the top-most shell course has been increased. If the deflection is still too much, then the whole process begins again at the bottom and repeats until the deflection limit is met.
The "standard thickness increment" is 1/16" up to 1" thickness, then increments by 1/8". In metric units the increment is always 1 mm.
(COMPRESS also allows the designer to specify a user-defined thickness increment (Set Mode Options dialog, Options tab). If this value is very small, say, 0.001", then the wind deflection calculations may be so lengthy due to the number of iteration cycles that it can appear that COMPRESS is locked up in an infinite loop. So use that user-defined thickness increment carefully.)
Several comments are in order:
1- COMPRESS does not include the straight flanges of the heads in the process of increasing thickness to limit the deflection. The extent to which increasing the thickness of the straight flanges will be effective in limiting the deflection is very minor. It seems better to let the thickness of the head straight flange be governed by whatever other considerations apply to the design: it makes little sense to have the dished portion of the head be governed by pressure to a relatively thin value and then force the straight flange to a much greater thickness. Thus designers may encounter situations where the straight flange of the bottom head is, say, 0.50" versus 1.25" for the bottom shell course. This can be accommodated by properly detailed transitions in thickness as per ASME Code.
2- There are no recorded calculations for the required thickness based on wind deflection. Unlike stress (or required thickness) calculations there are no detailed calculations for deflection. But the final result, the calculated deflection, is reported in the Wind Code. It can generally be confirmed that the actual deflection is about equal to the specified deflection limit.
If there is any question if the wind deflection governs the shell thickness, this can be confirmed by turning off the deflection limit and changing the shell thicknesses to a lower value (use the Global Change and enter "0", COMPRESS will select the required thicknesses), then run the Code calculations. The shell thicknesses will not be increased, at least for purpose of deflection limit. COMPRESS will report a Warning message that the deflection exceeds the "standard" value of 6" per 100' but this can be ignored.
A third effect that can result in increased shell thickness is the "Bergman" check, per ASME paper 54-A-104, for combined effect of external pressure and longitudinal compression due to weight plus wind or seismic. But this will have to wait for "Why does COMPRESS increase my shell thickness? [Part 3]"
"Although this forum is monitored by Codeware it is not intended as a venue for technical support and should not be used as the primary means of technical support."
Tom Barsh
Codeware Technical Support
For example, the required thickness of a cylindrical shell might be, say, 0.368" but COMPRESS enforces the nominal thickness to be 1/2" (0.50"
rather than the 3/8" (0.375" that one might expect based on 0.368" required thickness.Of course, these cases involve situations where the allowable stress (and required thickness) is a function of the nominal thickness. This is not encountered for tensile stress but may be encountered for compressive stress or buckling considerations.
There is also another situation in which COMPRESS may increase the thickness over what might be expected.
The Wind Code dialog provides an input to limit the deflection due to wind. By default this is 6" per 100' (5 mm per meter). Limiting the deflection may be desirable for process considerations. For example, a tall process tower with many trays may encounter problems if the vessel (and trays) deflect too much; this is not necessarily true for all cases, only the process designer knows for sure. Refractory lining might also require such a limitation on deflection. The designer can also "switch off" the wind deflection limitation on the Wind Code dialog.
If the nominal shell thicknesses determined on the basis of required thickness for circumferential stress (internal or external pressure) and longitudinal stress (weight, pressure, wind, and seismic) is such that the deflection due to wind exceeds the allowable limit, then COMPRESS will increase the thickness of the skirt by one "standard thickness increment" and recheck the deflection. If still excessive, then the thickness of the lowest shell course is increased by one "standard thickness increment" and the deflection is rechecked. This process is repeated until the top-most shell course has been increased. If the deflection is still too much, then the whole process begins again at the bottom and repeats until the deflection limit is met.
The "standard thickness increment" is 1/16" up to 1" thickness, then increments by 1/8". In metric units the increment is always 1 mm.
(COMPRESS also allows the designer to specify a user-defined thickness increment (Set Mode Options dialog, Options tab). If this value is very small, say, 0.001", then the wind deflection calculations may be so lengthy due to the number of iteration cycles that it can appear that COMPRESS is locked up in an infinite loop. So use that user-defined thickness increment carefully.)
Several comments are in order:
1- COMPRESS does not include the straight flanges of the heads in the process of increasing thickness to limit the deflection. The extent to which increasing the thickness of the straight flanges will be effective in limiting the deflection is very minor. It seems better to let the thickness of the head straight flange be governed by whatever other considerations apply to the design: it makes little sense to have the dished portion of the head be governed by pressure to a relatively thin value and then force the straight flange to a much greater thickness. Thus designers may encounter situations where the straight flange of the bottom head is, say, 0.50" versus 1.25" for the bottom shell course. This can be accommodated by properly detailed transitions in thickness as per ASME Code.
2- There are no recorded calculations for the required thickness based on wind deflection. Unlike stress (or required thickness) calculations there are no detailed calculations for deflection. But the final result, the calculated deflection, is reported in the Wind Code. It can generally be confirmed that the actual deflection is about equal to the specified deflection limit.
If there is any question if the wind deflection governs the shell thickness, this can be confirmed by turning off the deflection limit and changing the shell thicknesses to a lower value (use the Global Change and enter "0", COMPRESS will select the required thicknesses), then run the Code calculations. The shell thicknesses will not be increased, at least for purpose of deflection limit. COMPRESS will report a Warning message that the deflection exceeds the "standard" value of 6" per 100' but this can be ignored.
A third effect that can result in increased shell thickness is the "Bergman" check, per ASME paper 54-A-104, for combined effect of external pressure and longitudinal compression due to weight plus wind or seismic. But this will have to wait for "Why does COMPRESS increase my shell thickness? [Part 3]"
"Although this forum is monitored by Codeware it is not intended as a venue for technical support and should not be used as the primary means of technical support."
Tom Barsh
Codeware Technical Support