Interpreting FEA Stresses - Primary or Secondary? 1

[Crompto29]

I have a Rectangular door that is to be welded into a Ellipsoidal Head.

I have completed an FEA analysis but have been unable to confidently catagorise the stress in the head around the outside corner of the door frame and have therefore had it contracted out. They have since returned their analysis results which agree with my own.

I am trying to interpret the results and was hoping some of you might have some feedback. The FEA VON Mises Stress results are attached.

The Highly stressed region is under membrane and bending stresses.

What I would like to work out is if the membrane + bending stress is considered as Primary where I can use

PL + Pb < 1.5S

or is it a Secondary Bending stress where I can use

PL + Pb + Q < 3S

And why?

The results have a Von Mises Stress which is less then 1.5S therefore I am assuming the stress has been catagorised as Primary.

But it is quite localised and next to a discontinuity therefore why can't it be considered as Secondary? Is more then sight and experience required to establish Secondary Stress?

If it is Secondary is it mandatory to do a much deeper linearisation analysis across the thickness of the component to establish if there is a peak component?

As I will be placing someone elses results in my report I would like to be able to check or understand the results.

I have been using ASME VIII:2007 Div2 as my main reference.

Is it mandatory to run the analysis again for hydrotest conditions to check that PL + Pb < 1.43Sy?

Any feedback would be greatly appreciated. I notice a few FEA pressure vessel queries today. It is for a Hyperbarc Chamber that is to be pressurised once a day.

 

[jte]

Crompto.

You neglected to mention the loading which is important to this discussion. However, just guessing here, I’d say that your answer is “neither.” The hot spot which you are concerned with is neither primary nor secondary, but a peak stress. You will be evaluating this critter for fatigue, right???

jt

 

[TGS4]

Have you linearized the stresses through the thickness in accordance with Annex 5.A?

Also, please pay special attention to the note in 5.2.1.2

For components with a complex geometry and/or complex loading, the categorization of stresses requires significant knowledge and judgment. This is especially true for three-dimensional stress fields. Application of the limit load or elastic-plastic analysis methods in paragraphs 5.2.3 and 5.2.4, respectively, is recommended for cases where the categorization process may produce ambiguous results.

I believe that your situation may fall into this category. And, as jte says, don't forget fatigue - fatigue may govern the design.

 

[prex]

That localized stress is clearly due to the mutual constraint of adjacent structures and is therefore secondary in nature. More correctly the membrane+bending component at the corner is secondary, the balance being peak.
In a general situation like yours, I would proceed as follows:
-check the limit for general membrane: this is normally obtained by checking the thickness of the head without the opening, and by checking the opening reinforcement (using the longest side of the rectangle in this case, however this issue might be more complex, depending on the position and the proportions of the opening with respect to the head)
-check the limit for local membrane: the highest membrane stress will likely be at the spot in the figure, at the intersection of the neck with the head
-check the limit for membrane+bending stresses, that, as above, are to be classified as secondary stresses; note that secondary stresses need be checked as a range, so, if for example you had both internal and external pressures, the stress should be calculated for the total pressure excursion
-obtain the peak stress from the model, if sufficiently detailed, or by tables of factors, and check fatigue, if relevant.
Anyway, if, as you state, the maximum stress is lower than 1.5S, there shouldn't be any problems, but of course you need to explicitly check all stress limits per code in your report, or you can also classify all your stresses as primary, state that there are no secondary nor peak stresses, and if you can so satisfy the limits, you are OK.
Concerning the hydrotest, I would say there is no obligation of checking the limit, unless there is some special condition (e.g. hydro in horizontal position of a tall vessel) that could cause the limit being not satisfied.
However note that:
-there is no primary bending in your structure (unless you decide to classify the secondary stresses as primary!)
-as I assume you are calculating with elastic stresses, even if you wanted to explicitly check the limit, you wouldn't need to rerun the FEA, the hydro stresses would be simply calculated by linear extrapolation.

prex

http://www.xcalcs.com

: Online engineering calculations

http://www.megamag.it

: Magnetic brakes and launchers for fun rides

http://www.levitans.com

: Air bearing pads

 

[Crompto29]

Great Help. Thank you.

The Loading is 6 Bar(g) internal pressure only. And Elastic Analysis.

You've answered my Primary Secondary Query.

With what prex has told me and my experience it seems that many chambers are designed so that the peak stress doesn't excede 1.5S therefore there is no need to check fatigue, or atleast fatigue is not an issue. (Am I right?) I've never seen Fatigue calcs on any new chambers. I've seen a few 30 year old chambers that have NDT testing and a refurbishment and are sent out for another 30 years. (with stricter maintenance schedule of course.)

 

[TGS4]

Look at the fatigue curves and judge for yourself whether or not, by limiting the peak stress to a certain value, if fatigue is a genuine concern or not.

 

[corus]

There is no such limit of 1.5S for fatigue stress. If the component is welded then classify the weld detail and check that the number of cycles is acceptable for the stress range.

As far as checking the category of stress, I don't really see the point of wasting time on that as your stress is below the lowest limit and hence regardless of the stress category, the calculated values are within the limits set out in the standard. Your comment that as the stress is below 1.5Sm, then it must be a primary stress, makes no sense what so ever, as the value of stress doesn't equate to the stress category. As an academic exercise though, I'd class the maximum surface stress at the corner as peak (as its a stress concentration). At the junction of a head or flange then the membrane stress (mean stress across the section) is local primary, and the bending stress is secondary, with membrane plus bending (ie. the surface stress away from the corner) is less than 3Sm. At the corner you'd have to calculate the linearised bending stress through the thickness to get the bending stress without the peak component,

corus

 

[Crompto29]

Yes, quite academic.

I have been using comman sense. By keeping the peak stress below 1.5S (or yeild) and knowing that the number of cycles in the chambers life is only 10,000 I would expect that Fatigue is not a concern primarily because only elastic behaviour exists.

However as the Bending Stress is Secondary, a simple fatigue check should be undertaken to confirm that fatigue is not a concern.

Possibly not the most efficient design method.

 

[corus]

Yor comment that the peak stress is below yield and hence there is no need for fatigue assessment makes no sense. The fatigue design curves for fusion welded pressure vessels make no reference at all to the yield stress of the material. In the pressure vessel code I have, the fatigue life can be 10,000 cycles for a stress range of less than 300 MPa, which is less than the yield stress for S355 steel, for example. I'd also look at the principal stress for fatigue rather than the Von Mises stress intensity, as the principal stress can in some cases be greater than the Von Mises.


Again, your comment that a fatigue assessment should be undertaken because the stress is secondary makes no sense at all. If you have cyclic loading, then regardless of the stress category, a fatigue assessment should be undertaken.
Even if the stress or the number of cycles appears low, I would include a comment in your calculations that you have considered it.


corus

 

[TGS4]

I'll have to disagree with corus on the fatigue analysis method - specifically which stress to use in the evaluation. The Code requires calculating the stress range using the signed component stresses. Then, the component stress ranges are used to calculate either a stress intensity (2006 and earlier) or an equivalent stress (2007 and later). That's what is required to be in compliance with the Code - anything else might be technically OK, but would not be in compliance with the Code. And, if the Code has the force of law (as it does in most of North America), then if you are required to follow the Code, and you don't - you are breaking the law.

Also, for the welds, I would make certain that you use either the Structural Stress Method for the Fatigue of Weldments, or use a Fatigue Strength Reduction Factor (Kf). Both of these are outlined in Article 5.5 of Part 5 (2007 and later Edition).

 

[Crompto29]

Thanks allot, I beleive fatigue is a factor and the current design isn't adaquate. I will ask for the design to be revised.

Corus's post seems to relate to PD 5500 which appears to use direct stresses in its fatigue analysis.