Ansys Asme Pipe Stress Analysis - fast question 1

[Jos007]

Hi all

please check the following link:

http://karducci.altervista.org/PipeStress.pdf

as you can see it's a pipe system modeled with shell181 elements subject to internal pressure (only)
i need help about the membrane stress i get at the nozzle intersection
- it should be considered local membrane stress so the allowable stress limit should be 1.5S, is it correct?
- or maybe, since it's a very high geometry discontinuity location, i can consider the allowable stress limit equal to 3S?

i have been increasing the main pipe and the nozzle equivalent tickness to low down the membrane stress
but the solution i'm getting out is not cheap and hard to build

any nice tip would be very appreciated
thanks

 

[Jos007]

let's try to make it easier to answer:

1) the primary local membrane equivalent stress @ the nozzle intersection is higher than the allowable( > 1.5S = Sy)
2) the total local equivalent stress (membrane + bending + peak) @ nozzle intersection is lower than the allowable (< 3S = 2Sy)

which of the two above governs the design?

 

[TGS4]

Jos007 - unfortunately, evaluating the results
If a finite element analysis to the piping or BPV Code is slightly more nuanced than your questions imply. It appears that you are trying to perform an evaluation in accordance with ASME Section VIII, Division 2, Part 5, correct? Probably you have arrived there by apply the "unlisted component" rules if ASME B31.3?

What other loads are there on the component? You will need to evaluate those, too. Be glad to help you but I would need to know the whole picture... Are you expecting any cyclic loading? How about vacuum conditions or anything else that might result in compressive loads?

 

[Jos007]

@TGS4

- ASME Section VIII, Division 2, Part 5
- Internal Pressure only (self balanced)
- Not even constrained (only simmetry condition)
- No fatigue check required

imagine you build such a FE model, and imagine you get the results above described:
- would you consider it safe?

 

[Jos007]

1) the primary local membrane equivalent stress @ the nozzle intersection is higher than the allowable( > 1.5S = Sy)
2) the total local equivalent stress (membrane + bending + peak) @ nozzle intersection is lower than the allowable (< 3S = 2Sy)

which of the two above governs the design?

 

[Jos007]

OK, thanks.

let's try again:

1) the primary local membrane equivalent stress @ the nozzle intersection is higher than the allowable (PL > 1.5S = Sy)
2) the total local equivalent stress (membrane + bending + peak) @ nozzle intersection is lower than the allowable (PL + Pb + F < 3S = 2Sy)

i think that's not safe, PL should be < 1.5S
is it correct? if not explain it, please

thanks

 

[TGS4]

You are mixing two different failure modes.

If you are actually following the rules of ASME Section VIII, Division 2, Part 5 (2013 Edition), then you will know that for satisfying Protection Against Plastic Collapse, the local membrane equivalent stress should be less than Spl (which is the greater of 1.5S or Sy). if that is not met, then your have not satisfied this failure mode.

For satisfying Protection Against Failure From Cyclic Loading: Ratcheting, then the RANGE of primary-plus-secondary membrane-plus-bending should be less than Sps (which is the greater of 3S or 2Sy). Whether or not this is met, you still need to satisfy the other failure modes.

And, for that configuration, you would still need to satisfy Protection Against Collapse From Buckling and Protection Against Local Failure.

On the last one - Protection Against Local Failure, your geometry is one that is susceptible to such a failure mode. And a shell model is insufficient for making such a determination (the issue is in the inside corner of the lateral).

 

[Jos007]

thank you very mutch TGS4

on a solid model, from stress linearization along the main pipe and the nozzle tichness i'v got almost same stress values obtained from the shell model
the shell model is perfect to approach the problem, but for more accurate and detailed results the solid model is mandatory, i agree

one more question if possibile:
- imagine i make the stress linearization on the main pipe tickness and i find the results are not good
- one way to fix the problem may be to add a sleeve plate on the main pipe around the nozzle
- now, to evaluate the benefits, i repeat the stress linearization on the same path as before?
- or i should take the new path crossing both the tickness of the main pipe and the sleeve plate?

 

[TGS4]

The limits have to be met everywhere. You also have to make sure that if you classify something as local, then you have to meet the locality criteria as well (1.1S at a limit of sqrt(r*t))

 

[Jos007]

Evening All,

here:

http://karducci.altervista.org/Ansys_Pipe_Stress_Solid.pdf

you can find my stress analysis of the solid model regarding the failure check against internal presure (only) according to ASME VIII div2 2013
My biggest doubt is about the stress categorization, i'm not sure to consider the nature of the bending stress @ the nozzle intersection correctly:
as you can see in the "report", i assume the primary bending stress Pb = 0 (always) @ nozzle intersections, because it's a strong local discontinuity region
so, to check the stability of the item, i execute the following checks:

considering A516 Gr65 @20°C (Sy/Su < 0.7)

PLASTIC CHECK:
PL + Pb = PL < Spl = Sy = 240 MPa #5.2.2.4 step 5

RATCHETING CHECK :
PL + Q < Sps =2*Sy = 480 MPa #Figure 5.1

FAILURE CHECK:
(σ1 + σ2+ σ3) < 4*S = 598 MPa #5.3.2

am i correct?
thx for your time!

 

[TGS4]

A couple of comments:

A) You need to ensure that you are performing the Protection Against Plastic Collapse checks using the Design Pressure, and not the hydrostatic test pressure.

B) Your assessment that, at the intersection, there is no Pb is appropriate.

C) When performing the ratcheting check, you should be using the operating load ranges. See

http://becht.com/blog/asme-section-...ic-analysis-discussion-collapse-vs-ratcheting

D) When performing the Protection Against Plastic Collapse and the Local Failure check (4S), you need to be using the load case combinations described in Table 5.3

E) You are using ANSYS. Please note that the linearization scheme in ANSYS is NOT compliant with the requirements of 5-A.4.1.2 Step 2 (a).

F) Your Figures make it next-to-impossible to see if the SCLs are appropriate. Please refer to Annex 5-A for SCL guidance.

G) The Local Failure check is for primary membrane-plus-bending principal stresses only. It does not include secondary stresses nor peak.

 

[Jos007]

HI TGS4

thx a lot for the answers

about the comments:

A) I did that too, but with the Design Pressure i get accettable values, with the test pressure i have not. I can't find load combinations with the test pressure for an elastic analysis. Does that mean i have to switch to an Elastic-Plastic Analysis or a Limit-Load Analysis to check the item against the Test Pressure? is it mandatory?

B) ok Cool. Do you think it's right to assume Pb = 0 along the longitudinal side too? please check the following image to see what i mean:

http://karducci.altervista.org/Pb.png

since it's close to a geometry discontinuity it should be correct

C) ok, i will check it carefully

D) ok

E) ok, i read about it

F) I have re-uploaded the file, now it should be easier to check the chosen SCLs:

http://karducci.altervista.org/Ansys_Pipe_Stress_Solid.pdf

G) ok, btw if i do the check considering the secondary stresses too it shoud be conservative, do you agree?

H) i have a little question: considering a material having the ratio Sy/Su < 0.7 => we have Spl = Sy according to 5.2.2.4 step 5.
The Sy value of the material at the ambient temerature is easy to get, but at the design temperature, where do i get the Sy(T)? i mean the Yield Limit at the Design Temperature
On ASME Section II, Part D, Table 5A i can only find the S(T) value, but assuming Spl = 1.5*S(T) should be wrong according to above. What should i consider?

thanks a lot for your time!

 

[TGS4]

A) The test pressure is not part of the required loads to be considered in an elastic analysis.

G) That would be conservative, yes.

H) Sy(T) is obtained form Table Y-1 in ASME Section II, Part D

I don't have time to check your SCLs right now - perhaps later today or tomorrow.

 

[Jos007]

ah, i have the part D but it's incomplete, that's why, np i'll get it asap!

if you find the time, please answer me about the point B) too

thank you very much

 

[TGS4]

Pb really only exists in flat plates. So, your assessment that Pb=0 at that location is reasonable.

Since you are using ANSYS anyway, and the geometry is rather complicated, why don't you do an elastic-plastic analysis?

 

[Jos007]

i will try the elastic-plastic analysis as soon as possible and i will post my results

thank you