Vacuum Vessel Design and FEA 2

[tpwmech]

I am currently working on the design analysis of a large vacuum vessel that I have designed using COSMOS Works FEA software. This is first large vessel of this type that I have designed and I have a few questions regarding the ASME sec. VIII div. 1 specifications.

It is a cylindrical aluminum vacuum vessel with a 70" ID and is 14' in length with a number of large ports. The shell thickness, end flanges, and stiffener rings meet the ASME code requirements. The ports meet the 'material removal must be replaced' requirement as well.

My question is regarding the ASME maximum allowable stress for the material in regards to the FEA calculations. Does the FEA calculated max stress simply have to meet this requirement or do they need an additional factor of safety as well? If so what FOS should be included on top of ASME's conservative maximum allowable stress values?

Also, the FEA naturally shows the highest stress (9700psi) at a very small point on one of the ports. The areas directly around it and on the other port edges show a stress of approx 4000psi, while the maximum stress in the cylindrical shell is approx 2500psi. It is also not taking into account the reinforcement provided by the large fillet welds. Can this very small point be disregarded in the context of the chamber as an erroneous value given by the FEA or must the factor of safety be computed using this value?

Sorry this was so long.
Any help will be greatly appreciated.
Thanks,

Doug

 

[TGS4]

Question #1 - Why are you using FEA for a Division 1 vessel?

Question #2 - Does the vessel in question meet ALL the requirements of Division 1?

Question #3 - Have you heard of Division 2?

 

[prex]

Assuming you are not required to code stamp this vessel (otherwise answers #1 and #2 of TGS4 would apply), I concur anyhow with TGS4: you are complicating your life by doing a FEM analysis on this vessel.
Assuming you are using shell elements, with such an analysis you see bending stresses that are unrelevant to your situation (ASME VIII Div.1 explicitly recognizes that secondary stresses need not be determined under its coverage).
Also you are not mentioning the allowable stress in compression for the shell and the heads: FEM does not help you on this, unless you want to perform a buckling analysis, but in that case remember what the latins said: 'errare humanum est, perseverare diabolicum'.

prex

http://www.xcalcs.com

Online tools for structural design

 

[jte]

tpwmech-

Question #4 - How do you define "large"?





prex-

(ASME VIII Div.1 explicitly recognizes that secondary stresses need not be determined under its coverage).

In reading UG-22(h) and UG-23(e) I'm not quite so sure about that…

jt

 

[JStephen]

Reading that, it sounds like the discontinuity stresses are not the membrane-plus-bending stresses of which it speaks.

 

[prex]

jte,
read UG-23(c) towards the end.

prex

http://www.xcalcs.com

Online tools for structural design

 

[jte]

prex-

So UG-23(c) states in part that It is recognized that high localized discontinuity stresses may exist in vessels designed and fabricated in accordance with these rules. Insofar as practical, design rules for details have been written to limit such stresses to a safe level consistent with experience.

It does not state that such "high localized discontinuity stresses" need not be evaluated. Then it goes on to provide limits borrowed from Div 2 in UG-23(e). Stresses caused by thermal loadings are usually classified as secondary stresses. UG-22(h) seems to make it clear that such stresses shall be considered. Note that "considered" does not mandate calculations - engineering judgement and prior experience may suffice, but if the engineer considers the thermal loading potentially significant then these secondary stresses must be considered.

I guess I don't understand your point...

jt

 

[prex]

Well jte, the wording you cited tells exactly what I mean: the code recognizes that high local stresses may exist, but that they need not be determined, as they are meant or presumed to be acceptable simply by following design details enforced by the code.
Also: it is incorrect to state that thermal loadings are usually classified as secondary: the classification of stresses is always a matter of being in accordance with a mandatory technical rule. An example of primary thermal expansion stresses? The stress in the shell and in the tubes of a heat exchanger due to differential thermal expansion (when relevant of course).

prex

http://www.xcalcs.com

Online tools for structural design

 

[zjliang]

Doug,

If there is no fatigue issue, do not worry about the horrendous peak stress, very localized concentrated stress. However, regardless of what cases, the stress at one spot (either nodal stress or elemental stress) shall NOT be directly compared with the Code Allowable stress at all. Linerization is always necessary to classified the total stress to membrance and bending stress, then check with code. See details in Appendix 4 in Div 2 Code.

Prex and jte,

Here I have some comments on the consideration of secondary stress in Div 1 Code.

Why Div 1 Code is so "popular" and easy to use, because it uses higher allowable stress and more detailed requirements to limit the discontinuities. However, the compromise is so-called "engineering judgement" has to be made in some special cases, especially for handling those loading in UG-22. But where is the engineering judgement from? My understanding is either from more detailed analysis like FEA or from the past experiences.

 

[JoeTank]

Zjliang,
You make refernce to Div 1 in your post. Since Div 1 does not use higher allowable stresses, I wonder if you really intended to make reference to Div 2?

Joe Tank

 

[zjliang]

Thanks for your reminding, Joe.

I meant higher safety factor 3.5 (revised), but lower allowable stress. Although quite close to div.2 limit (3.0), it still cause substential difference in some particular cases like higher pressure, thicker vessels.

Anyway, thanks again.

 

[jte]

zjliang-

As I said... UG-22(h) and UG-23(e) are included for a reason. Div 1 is also more popular since its easier for a user to buy (they don't need to provide a PE stamped User's Design Spec) and easier to repair/alter (that PE thing again, and number of contractors willing to play with repairs to Div 2 vessels). But even with Div. 1 you still need to evaluate things which aren't specified in the code. Show me a designer who limits stresses to S when he's evaluating piping loadings on a nozzle and I'll show you one who's wasting material.

With the additional material savings I'd expect that the new Div. 2 will become a bit more popular.

prex- I don't follow how your example of tube/shell differential expansion is not self-equilibriating and thus results in primary stress...

jt

 

[prex]

jte,
that those stresses are primary is out of discussion, I think: just look at how TEMA treats them.
The reason is in the exact wording for the definition of secondary stresses in Div.2 App.4: A ... stress developed by the constraint of adjacent parts or by self-constraint of a structure.
Here a structure must be interpreted as a single structure with uniform structural behavior (my interpretation of course): this limitation is necessary because the stress developed in a structure by the expansion of a different one is not necessarily self limiting. This is the case for shell and tubes of a fixed tubesheet HE; another example is a piping and its supports: expansion stresses are secondary in the piping, but become primary in the supports.

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

 

[jte]

prex-

I think we'll have to agree to disagree! I'll choose to underline a different section of your quote: A ... stress developed by the constraint of adjacent parts or by self-constraint of a structure.

I'll also quote from the definition of primary stress, 4-112(g): ... A thermal stress is not classified as a primary stress...

Having said that, I agree with your commentary about piping (see also 4-138 and

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

jt

 

[TGS4]

Unfortunately, discussions such as these about what is primary/secondary are bound to continue until people decide to abandon elastic analysis and go with elastic-plastic.

prex, for your case, I think that if you were to perform an elastic-plastic analysis, you would find that there are almost no "stress" limit on the stresses that you discuss - but it takes an elastic-plastic analysis to discover it. We can argue all we want about what's primary/secondary, but the definitive answer for every individual problem can be found in an elastic-plastic analysis. The paper that jte refers to covers that discussion adequately.