compress or pvelite? 1

[rcbao2000]

I ran into a problem. Same vessel designed with both sw came up different results! Major different is in the leg design with seismic load.
Can any one offer an opinion? compress or pvelite?

 

[LSThill]

rcbao2000 (Mechanical)

You need to review the following:

ARCHITECTURAL INSTITUTE OF JAPAN
DESING RECOMMENDATION FOR STORAGE TANKS AND THEIR SUPROTS 1996 EDITION – Earthquake Resistant Design of Storage Tanks and Their Supports -

Regrds

Leonard Stephen Thill

 

[TomBarsh]

The answer is not "either/or" COMPRESS or PVElite. Both may be correct, the differences being due to different inputs, different conditions being reported, etc.


What you ask is not a trivial question with an easy answer. Some of the most difficult questions I have answered are of the type "I have two different COMPRESS vessel files and they give me two different results. Why?" There may be legitimate reasons for the purported differences depending upon the situation.

In some cases two different builds of COMPRESS will provide different results. This may be due to a bug fix in the later build (it is good to get bugs fixed). Or it could even be due to a new bug in the newer build (fortunately, this happens less often).

In some cases the difference between the same COMPRESS file run in two different builds may be the result of a minor change made elsewhere. An example comes to my mind:

Some time ago there was a case where a vessel was adequately designed in one build of COMPRESS, but the anchor bolts were inadequate when run in a new build of COMPRESS. This was a relatively tall, skinny tower for which the vortex shedding analysis option had been applied.

In the older build of COMPRESS the critical wind speed was only fractionally greater than the design wind speed, less than one mph difference. But this was sufficient that COMPRESS did not perform calculations for the vortex shedding condition (no need to do so if the wind speed at which this would occur is greater than the design wind speed).

In the newer build of COMPRESS, a slight change had been made to a weight calculation. This resulted in literally a few pounds difference in weight. This was sufficient to change the natural frequency of the vessel so that now the critical wind speed was fractionally less than the design wind speed. COMPRESS proceeded to perform calculations for the vortex shedding case, which resulted in the anchor bolts being increased in size.

Which analysis was "correct"? (rhetorical question!)


More frequently, users send files to us that they claim are the same but provide different results. More often than not, there is some difference in the inputs: design conditions, dimensions, corrosion allowance, etc, that accounts for the differences. I have even seen totally different wind and seismic codes in the two "identical" files; of course they will provide different results in this case.

There are numerous settings and switches available in COMPRESS that will affect how the calculations are produced as well. Some of these will greatly affect how the calculations are performed. For example, several ASME Code Cases and Interpretations are available on the Codes menu. Selecting these can make the difference between a design passing or failing.

All these comments pertain to working with just one software application. Comparing calculations produced by two different sources (COMPRESS and another commercial or in-house software application, hand calculations, spreadsheet, etc) will be even more difficult. This can require painstaking detective work to step through to discern the differences.

I have been asked a few times to comment on such differences. In most cases the differences resulted from an "apples to grapes" comparison. Both sets of calculations were correct but were reporting different conditions. In a recent example I believe one software had a Code Case applied but the other did not, this made a huge difference in a flange calculation. No, it was not the "easy" one for the flange rigidity, I don't remember exactly what it was but it required tracking by hand through all of the complex calculations of Appendix 2.


So, you have your work cut out for you. The best approach is to treat the calculations as hand calculations and "check" them. This will help you understand the calculations and identify the differences.

Tom Barsh
Codeware Technical Support

 

[TomBarsh]

That said, compare the weights and the natural frequencies. Also confirm that the input criteria for the seismic code are similar.

Then investigate the calculations for the support legs. These all come from the AISC Steel Design Standard. Compare the axial loads and bending moments, and allowable stresses applied in these formulas, then trace these values back further if there are discrepancies.

Tom Barsh
Codeware Technical Support

 

[rcbao2000]

Thank you Leonard. I can not get the paper from ARCHITECTURAL INSTITUTE OF JAPAN. I will check into it when I have chance.

Thank you Tom. I found difference in my problem. As you said, Compress uses AISC and PVElite uses WRC107 to calculated supporting legs. The results are "apples and grapes" different! Both results are correct, but which one is more economical and which one is more conservative?

Does anyone have more in-depth knowledge on both standards?

 

[LSThill]

rcbao2000 (Mechanical)

Design Recommendation for Storage Tanks and Their Supports ,2004 COST 2,520 YEN

http://www.aij.or.jp/eng/publish/index.htm

Maruzen Co., Ltd.
International Division
P.O. Box 5050 Tokyo International 100-3191 Japan
Fax: +81-3-3273-1044 Tel: +81-3-3273-3234
E-mail: export@maruzen.co.jp

Regard

Leonard Stephen Thill

 

[TomBarsh]

rcbao2000,

COMPRESS uses WRC bulletin 107 to calculate stresses in the shell at the leg attachment point. The rules of the AISC ASD Specification are used for analysis/design of the legs themselves; ie: the legs are treated as a "column" in the parlance of structural steel design. The legs (columns) will generally have both an axial compressive load and a bending moment at the top of the leg (results from the eccentricity of the load at the top of the leg). See this document for more discussion:

http://www.codeware.com/support/faq/1156.html

Whether you do the vessel design by hand or use COMPRESS or other software, the same principles must apply: (1) Analyze the local stresses in the shell at the leg attachment (WRC-107 or similar), (2) analyze/design the leg itself as a column (AISC, etc), and (3) analyze/design the leg base plate (AISC, etc).


Tom Barsh
Codeware Technical Support