Some Compress questions

[McJe]

While evaluating Compress, I came accross some issues, which I think are not 100% according to ASME...
But, I'm pretty new at this, so I might be wrong too.

(1) A nozzle in a head:
To calculate the minimum head thickness, the corroded values of Di and h are used, and the thickness is calculated according to UG-32. But, D/2h isn't 2 anymore for corroded elliptical heads. The formula of 1-4(c) should be used with the new K-value.

(2a) Nozzles with increased thickness (calculations for internal and external pressure, on shels and on heads):
The second formula of the A1 calculation is:
2*(t+tn) * (E1*t-F*tr) - 2*tn * (E1*t-F*tr) * (1-fr1)
In case of increased thickness, the value of tn isn't the same in the first and the second occurence. The first occurence defines the limits of the reinforcement, and should be the nozzle neck thickness (if L<2,5*tx , fig UG40-e1)
The second occurence of tn evaluates and recalculates the available material in the nozzle part, and here the value of tn should be the increased thickness.
Compress uses the increased thickness in both cases. This allows material outside the reinforcement zone to be included in the calculations.

(2b) Nozzles with increased thickness (calculations for internal and external pressure, on shels and on heads):
According to ASME, a nozzle with increased thickness needs to be calculated in 2 parts:
1 - with the formulas for A2, only the nozzle neck thickness is considered.
2 - with the formula for A5, the increased thickness must be calculated.
Both formulas have different limits of reinforcement (calculation of te in formula A5)
Compress seems to assume that the increased thickness can be included in the calculations of A2. This gives large differences with my manual calculations.

(3) I found an error in the calculation of the shell thickness for external pressure, when calculation nozzle reinforcements, when there are stiffening rings. Depending on the place of the nozzle and the ring, a wrong shell thickness is calculated. This error is easy demonstrated in Compress, but since I'm still evaluation with the demo version, I cannot save the example.


Best regards,
Jeroen.

 

[TomBarsh]

I suggest that you contact Codeware directly to receive direct answers to your questions. This forum was not intended to be only Codeware providing technical support.

(1) COMPRESS calculates the head ratio and determines the K factor separately for the new and the corroded conditions, then uses the formulas from Appendix 1 to determine the head required thickness. See

http://www.codeware.com/support/faq/1011.html

COMPRESS also provides a switch to force the use of the UG-32 formulas for the heads for both new and corroded conditions. This can only be applied if the heads meet the geometry limitations in these paragraphs. This keeps the people in Alberta happy.

(2) Apparently you refer to Type 9 and Type 10 nozzles with a heavy barrel. COMPRESS provides these Types specifically for the purpose of accounting for the special rules of Figure UG-40 sketch (e).

I have gone over these a number of times and am not aware of any error. If you believe there is an error you should send a nozzle sketch and we will investigate.

(3) This is a limitation of the software. The issue is that when a cylinder has 1 or 2 (or more) vacuum rings on it, there may be multiple "effective lengths" for each span of the cylinder. Each section would then have a different required thickness. COMPRESS determines 'tr' for the cylinder as based on the longest span. This same 'tr' is used for any nozzle that is attached to the cylinder regardless if it is located in the longest span or not.

We discussed this some years ago and there may be a bug list item to consider the 'tr' that actually applies for the shell at the nozzle's location. It is a low priority issue though. In general this would have negligible effect on the nozzle's MAEP rating, assuming that the vessel is properly modeled by using separate cylindrical shells for each of the actual shell courses. Modeling a vessel with only a single cylindrical shell to represent multiple actual shell courses might encounter an overly-conservative condition.

"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

http://www.codeware.com

 

[TomBarsh]

For example, if one modeled a vessel that is 100' tangent to tangent as a single cylindrical shell and had a single lonely vacuum ring around 4' from the bottom head, then the cylinder will be designed based on required thickness for the 96'+ span (the '+' being the extra distance to the line of support at 1/3 the depth of the head). If a nozzle is located in the region below the vacuum ring, where the tr for vacuum will be much less due to shorter effective length, COMPRESS will still use the larger value tr for the nozzle reinforcing calculations. This is an extreme case. The problem will be minimized by modeling the total cylinder length as based on the actual lengths of the shell courses. If one is really concerned about this issue, then simply model a few extra &quot;dummy&quot; seams so that the nozzle is located in a cylinder whose actual length reflects the effective length of the actual construction.

&quot;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.&quot;

Tom Barsh
Codeware Technical Support

http://www.codeware.com

 

[TomBarsh]

Regarding (2), make certain that you are considering all of the reinforcing area available within the normal limit of reinforcement. It is incorrect to assume that the available reinforcing area is limited to the area below "te" (as though this was a physical limit). There are Code interpretations available from ASME on this question.


"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

http://www.codeware.com