Hi,

The typical approach to qualify local stress at gross structural discontinuities (these including fittings/intersections, welded attachments) is to use ASME VIII-2 Part 5 criteria.
So, the answer at your question (3) is DEFINITELY YES!

If your elastic analysis analysis revealed secondary stresses above twice yield strength, then it appears that ASME VIII-2 Part 5 requirements related to elastic stress analysis and corresponding classified stress qualification are not fulfilled.

In such case, you may employ elastic-plastic analysis to demonstrate the discontinuity & loading suitability. However, be aware that elastic-plastic approach consists basically in several analyses:

(1) (Elastic-Plastic) Limit-Load Analysis, or, alternatively,
(2) Elastic-Plastic Stress Analysis,
to demonstrate that plastic collapse risk is negligible. These analyses are aimed to replace the Elastic Membrane (general, local, bending) Stresses qyualification (e.g. Pm, Pm+Pl, Pl+Pb stresses). Be aware that some load multiplication factors (1.7...3.0 - see Part 5 requirements/tables) are to be considered as "Load Factors".

(3) Ratcheting Assessment - Elastic-Plastic Analysis: this analysis should be employed if Secondary stress (Pl+Pb+Q) and/or Secondary Stress Range (DELTA Pl+Pb+Q) exceed(s) SPS Limit (usually twice Yield Strength). In such case, you need to apply MINIMUM 3 full loading/unloading cycles to demonstrate that plastic strain is self-limiting and there are no significant/unacceptable plastic/remaining deflections - see applicable ASME VIII-2 Part 5 Section, where the requirements are clearly stated.

So, to conclude, if ELASTIC STRESSES do not match the required limitation conditions, the ELASTIC-PLASTIC analyses MAY be employed, BUT with some multiplication load factors and repeated loading/unloading cycles.

I would also advise you if you already did not do that, to review the FEA Model with respect to ASME VIII-2 Part 5 and/or sound engineering requirements. I am talking about fillet welds areas. Probably you used Shell Elements. Just simply model the surfaces intersections and meshing the surfaces by Shell Elements is not sufficient to get realistic results. See the requirements stated by ASME VIII-2 Part 5 in this subject. I would also suggest to use PRG FEA Software package (FE PIPE, NOZZLE PRO) whose typical templates include the suitable meshing algorithms for welding areas.

I hope these considerations may help.

Happy Holidays!


Best regards,