Nozzle loads, primary or secondary. 1

[jrjones]

Okay, before I get in trouble, I know that there are a couple of threads out there on this topic...but as yet I am not satisfied. I am trying to analyze a nozzle (specifically the nozzle wall) using the 2007 ASME section 8 div 2 (S8D2) analysis.

First of all, my nozzle configuration is made up of a forging welded into the shell, that attaches to a conical reducer, that attaches to a 8" weld neck flange. The forging into the shell is roughly a 12" pipe. This is done because of a cold wall to hot wall transition.

Anyway, the stresses in the region between the small end of the reducer and the weld neck flange hub are pretty high on one side. I compare the local membrane stress to 1.5S (where S is the at temperature allowable) and I am okay. I want to compare Pl and Pb (local membrane and bending) to 1.5S as well. If I do this, I don't pass. When I read table 5.6 from S8D2 it guides me to treat bending as a secondary stress (Q). Then I could compare the Pl+Pb+Q<3S.

So, I know you are wondering...the loads that are giving me a problem are due to piping expansion. I just don't know if I believe that the thermal expansion loads will shake down such that their influence on the nozzle stress can be categorized as secondary. If the thermal movement was very small, maybe...but "categorically" I'm not sure (I think there is a pun there!).

I'm okay with secondary stress categorization in the nozzle to shell junction, but not sure about the nozzle wall.

Can anyone set me straight on this categorization stuff.

 

[LSThill]

jrjones (Mechanical)

Reference you need to have WRC Bulletin 429

L S THILL

 

[TGS4]

For starters, in the pipe wall, you will not have Pb, which can only exist in a flat plate. Through-wall bending is (almost) always secondary.

If you want to validate this, perform a limit load analysis. If you want a better explanation of the primary/secondary categorization, you need to see the "Criteria" document of Section VIII, Division 2 (from the 60's). Let me know if you can't find a copy.

 

[jrjones]

I found the WRC bulletin and will read up...I don't have access to the older S8D2...TGS4, do you know where I could find the "criteria" document you referred to..?

Thanks for the help.

Ryan

 

[prex]

jrjones,
the whole picture is not so simple as you put it. You need to separate the different stress contributions in order to classify them (or perform a limit load analysis as suggested by TGS4 to overcome everything).
As you state, you are in a zone of transition between a cone and a cylinder (flange neck).
As far as the stresses due to internal pressure are concerned, besides the general membrane circumferential and longitudinal stresses, you have the stresses due to the constraint of adjacent structures (cone and cylinder, not to mention the flange ring) with different stiffnesses. This constraint gives rise to membrane and bending stresses (TGS4, this is not a flat plate, but bending stresses exist indeed!). Both of these are secondary in nature, but the membrane part shall be generally classified as local membrane to avoid excessive distortion, the bending portion is certainly secondary.
As far as the stresses due to thermal expansion of the connected pipe are concerned, you get:
-under the axial load a membrane longitudinal stress and again membrane and bending stresses due to the interaction
-under the transverse moments essentially membrane stresses only, that combine locally to other stresses according to the orientation of the moments
-under the transverse loads and the twisting moment essentially membrane stresses only.
Now the point, that will possibly surprise you, is that all of these thermal stresses should be classified as primary, except for the part due to the cone-cylinder interaction (so again the bending stresses are secondary). This is because you don't know under what kind of deformation those stresses will relax, unless you perform a combined analysis including the vessel, the pipe and the supports.
The resulting whole picture corresponds to what you are doing: the bending (the through thickness bending, not the bending across nozzle hollow circular section!) is secondary.

prex

http://www.xcalcs.com

: Online tools for structural design

http://www.megamag.it

: Magnetic brakes for fun rides

http://www.levitans.com

: Air bearing pads

 

[bvi]

The responses are correct, the through thickness bending is a secondary stress regardless of the source of the load on the nozzle, but I wanted to address another part of the question. The nozzle does not know whether the loads being applied are due to pressure, weight, or thermal expansion, and local deformation of the nozzle will almost never singificantly reduce loads due to piping thermal expansion. All loads on the nozzle are treated the same, as external mechanical loads. There is an error in some commerical software (e.g. CAESAR and NozzlePro) that treats the membrane stress in the nozzle and shell due to thermal expansion loads of the piping as a secondary stress. In my opinion, this is an error. It was a misinterpretation of the old Div 2 Appendix 4. I think the new Div 2 is clearer in this regard.

 

[jrjones]

Thanks for the discussion. I did have an issue with calling the nozzles stresses secondary. My issue is that I question myself because when I read the 2007 section 8 div 2, bending in the nozzle wall is to be treated as secondary. This is in the table for a description that says for all external loads and moments on the nozzle. So this would say to treat the bending as Q and compare Pl+Pb+Q<3S.
This seems to contradict what makes sense and what has been said above. I especially like the way BVI put it, "...and local deformation of the nozzle will almost never singificantly reduce loads due to piping thermal expansion."
This seems correct to me and would dictate treating the stresses as primary (<1.5S). But I am confounded by what S8D2 says. The 2004 code and the 2007 code don't really seem to agree.

 

[LSThill]

jrjones (Mechanical)

WRC BULLETIN 429 - FEBRUARY 1998
This work is an othgoruth of previous PVRC work, which defines the problems and issues involved in assessing stress results for three-dimension (3D)finite element analysis (FEA) in terms of the ASME Boiler and Pressure Vessel Code stress limits in the design by analysis Section III (Class 1, NB) and Section VIII, Division 2. The intital project presented sic short-teams recommendeaber, four areas for additioan consideration, and number of issues requireing long-research. The current (1) relation between failure mechanisms ans the ASME Code stress categories, (2)the approparate stresses for each category, (3) the appropiate locations for assessing each stress categroy, and (4) the appropriate stress for obtaining the membrane plus bending stresses. Although the relationship are interrelated, they are developed individually and then synthesized. To enhance understanding to the foru area oc consideratin, 11 geometris are defiened and evaluated through discussion and finite element results. Each of the four areas of consideratin is discussed in detail and two-dimensinal(2-D) azisymmmetric and 3-D example geometrise are presented. "Recommended Guidlines," which Sections VI
II and III of the may use to update appropriate protions of their code.

Yes: responses are correct! but WRC BULLETIN 429 - FEBRUARY 1998 is the histiory for the update on the 2004 and 2007 code.

 

[TGS4]

jrjones - contact me off-line and I will set you up with a "fair-use copy". Posting it here would definitely infringe on the ASME copyright...

The new and old VIII-2 are in agreement , the new one uses better language. Except for the situation of a flat plate, bending stresses are secondary. Period.

This example is the "perfect storm" that highlights the difficulty in classifying stress. Even the concept of stress classification (particularly what is primary and what is secondary) is somewhat antiquated. Check out how these two papers evaluate a similar issue:

http://store.asme.org/product.asp?c...Papers&category_name=&product_id=PVP2004-2613

and

http://store.asme.org/product.asp?c...apers&category_name=&product_id=PVP2005-71535

In the end, I would just perform the limit load analysis and be done with it.