Failure theory used when performing an FEA in a ASME VIII-1 vessel 1

[FFEA]

I am performing an FEA to qualify a nozzle subject to external loads for an ASME VIII, Div. 1 vessel. I am using NozzlePRO, which follows the rules of ASME VIII, Div. 2, Part 5, using the allowables for a Division 1 construction (Table 1 and Table 1A from Section II, Part D).

The problem is that the AI says that we should use the maximum principal stress because the Division 1 allowables are based on Rankine's theory. On the other hand, ASME VIII, Div. 2, Part 5 states that the von Mises equivalent stress shall be used (5.2.2.1).

The question is, which failure theory should be used in order to get a design "as safe as" those provided by the rules of Division 1? What is the common industry practice?

Vessel general data:
Code of construction: ASME VIII, Division 1, 2015 Ed.
Material: SA-516 Gr. 70/SA-106 Gr. B
Design pressure: 1 MPa
Design temperature: 70 °C
Cylindrical shell diameter: 1500 mm
Shell thickness: 12.7 mm
Nozzle OD: 508 mm
Nozzle thickness: 12.7 mm

Thank you in advance.

 

[TGS4]

Follow the VIII-2 rules, using the VIII-1 allowable stresses. That is sufficient.

The Code Committee is drafting a Mandatory Appendix on this exact subject. The guidance that I have above is in line with that Appendix. You need to address all of the failure modes listed in Part 5. If there is no allowable basis in VIII-1, then use the basis in VIII-2.

The issue that your AI brought up is a red herring. It has been rejected by the experts in the Code Committee.

 

[FFEA]

Thank you very much TSG4.

And do you know why using the maximum principal stress has been rejected by the Code Commitee? Could it lead to non-conservative results?

 

[DriveMeNuts]

Federico,
If you rely purely on the maximum principal stress only while the second principal stress is also positive then that is fine. This is essentially the Tresca stress.

However, if the second principal stress is compressive then there are shear stresses which fail more easily than a purely tension stress. This means Rankine is non-conservative for tension-compression stress systems making it inappropriate for complex stress systems. I think your AI is misguided.

I'm sure if your AI understands vessel design then Tresca theory would be acceptable. It produces identical results as Rankine for stress that have Principal stresses which are all in tension and produces more conservative results from Rankine where there are locations with tension and compression principal stresses (shear stresses).

Most FEA applications can effortlessly be set from Von Mises to Tresca which is only slightly more conservative than Von Mises. If you use Tresca, you could tell your AI that Tresca is exactly the same as Rankine theory with the exception that it is more conservative than Rankine in areas of shear stress, therefore you are complying with Rankine.

U-2(g) only says that the equipment needs "to be as safe as" div 1 rules. No where does it prevent you form using a failure theory other than Rankine. Your AI is possibly over reaching thir level of expertise.

 

[FFEA]

Thank you for your answer DriveMeNuts.

That's exactly what I explained to the AI, but he insisted that comparing the von Mises equivalent stress against the ASME VIII, Div. 1 allowables would be as "mixing apples and oranges".
I agree with your point that I could use Tresca in order to comply with Rankine, but I think it is more appropriate to use von Mises. As it is stated in ASME PTB-1-2014, Page 149, "The maximum distortion energy yield criterion is used in VIII-2 because it matches experimental results more closely and is also consistent with plasticity algorithms used in numerical analysis software".

 

[Vikko]

ASME VIII Div 1 is DBF (Design By Formula) only !!!

If you want to do FEA/FEM/Design BY Analysis, you could follow ASME VIII Div 2. Or EN 13445-3 Annex B (for plastic limit analysis) or Annex C (for the classical elastic (Hooke Law) analysis).

Tresca and Von Mises are almost the same, just Tresca is a bit more conservative (while Von Mises formula for the equivalent stress is a bit more complicate). Both ASME VIII Div 1, and many other codes, for shell and tubes always use a Tresca formula (similar, not exactly the same formula, for the joint efficiency, and other minor assumptions) for the equivalent stress in the cylinder material under pressure.

Rankine shall not be used for steel! But only for brittle materials, like bricks, stones, ceramic, glass, maybe cast iron, etc.

 

[TGS4]

The DBR portion of VIII-2 also uses max principal stress, but combined loading uses von Mises.

Red herring.