ASME BPVC VIII-2 Elastic-Plastic analysis 2

[Paulettaa]

Dear All

I am running an elastic-plastic analysis of an equipment as per ASME BPVC VIII-2 Ed. 2021. I have some questions hope you could help me.
1- I am using Ansys workbench. As I understand in material properties for FEA true plastic strain vs true stress shall be the input data. Based on the procedure in Annex 3-D which quantity needs to be considered as true strain in the FEA? is it Gamma1+Gamma2?

2- If an elastic stress method is applied and joint efficiency for a given component is less than one then I multiply the allowable stress by the value of joint efficiency. What about elastic-plastic method? Do I have to increase the load factor or decrease the thickness or something else?

3- The load factor beta, depends on the class of component in division 2. What if I want to run an elastic-plastic analysis for a division 1 component? Do I need to take beta to be 3.5?

Warm Regards

 

[IdanPV]

2. Note that PART 5 assumes Joint Eff. of 1.

3. Yes, see ASME VIII-1, Appendix 46.

 

[TGS4]

1. The plastic strain is indeed gamma_1 + gamma_2.

2. Idan_PV is correct - partial weld joint efficiency factors are not permitted in Part 5. If you are coming in from ASME Section VIII, Division 1 through Appendix 46, this is also clearly spelled out.

3. Idan_PV is also correct on this - this is noted in Appendix 46.

 

[Paulettaa]

Thank you IdanPV and TGS4 for helping me.
1- Does this mean that the true strain for true yield stress is equal to e[sub]p[/sub] (e[sub]p[/sub] epsilon[sub]p[/sub])? I am asking since in 2021 edition the code has been modified to consider true strain to be linearly proportional to true stress when Gamma1+Gamma2 is less than e[sub]p[/sub]. This is important since in order to define a multilinear plasticity firs you need to define a true stress where the true plastic strain is equal to zero.
2- Does this mean that if you want to design a component based on the rules given in Div.2 Part.5 it shall have JE equal to one? I have seen documents from various vendors through years applying Part.5 with JE's less than one by simply modifying the allowable stress by a factor equal to JE. So for example: Pm < SE where E is the JE instead of Pm < S. My judgement says this is quite justifiable. The second problem is that in Appendix 46 the joint efficiency is not equal to one. The code is just asking for full radiography requirements of UW-11 and UW-12. Well clearly there are welds with full radiography and joint efficiency less than one. See for example table UW-12 weld type number 2. What then? If the code wants to rule out these type of weld from design by analysis it shall simply say JE equal to one not full radiography.
3- In design for Div.1 there are two sets of material properties defined in II-D for some Austenitic materials (most notable SS304) where it is possible to take a higher allowable stress by increasing the criteria for yield from 0.66Sy to 0.9Sy and these materials are given a note G5 in material properties table of 1A. This helps decreasing thicknesses in designs where some straining might be tolerable. It is reflected (maybe unknowingly) in Appendix 46 for limit load analysis where it asks to take yield stress equal to 1.5S. Here again the thicknesses might decrease if the material properties is so selected with note G5. However, if elastic-plastic method with beta equal to 3.5 for all materials is used all of a sudden this option is not working anymore. Is it the intent of the code that by using elastic-plastic method using note G5 material properties be not available?

Thanks again

Regards

 

[Morcuse]

The choice of 0.66Sy or 0.9Sy relates to modifying one possible governing option of the allowable stresses in Div. 1. The other option still applies to stainless steel materials with a higher allowable stress: tensile strength / 3.5.

Therefore, the application of an elastic-plastic method with beta equal to 3.5 continues to be consistent, even for stainless steel materials allowed to use closer general membrane stresses to yield strength, since it is understood that there continues to be an acceptable margin of design from the material tensile strength.

 

[TGS4]

1 - the true stress-true strain relationship in Annex 3D provides the full (elastic + plastic) strains. Typical FEA software only asks for the plastic portion of the curve - hence the use of only gamma_1+gamma_2. You use the stress when gamma_1+gamma_2 is equal to epsilon_p to be the proportional limit - and although the plastic strain there is (by definition) epsilon_p, you enter it as zero in yrou software. Does that make sense?

2 - I was the technical project manager within the ASME Committees for trying to incorporate E<1.0 into Part 5 - just like I had done for API 579-1/ASME FFS-1. However, that action ultimately failed because the consensus of the Section VIII Committee (both Sub-Group Design and the Main Committee) was that if the designer (engineer) wanted to use DBA, then they needed to ensure that sufficient inspection was performed to ensure that a weld joint efficiency equal to 1.0 could be assured.

I certainly see your point about Table UW-12, Type 2 joints having a weld joint efficiency equal to 0.9 for full radiography. However, it was the judgement of the Committee that if the details of this joint type were included in a DBA, then the "effective" joint efficiency would be apparent embedded in the DBA. Hence, the wording that is currently in Appendix 46.

One thing to remember about Weld Joint Efficiencies - one of the biggest issue with them is that there is no technically-justifiable basis for the number in the Code. Rather, they are based solely on the judgement and experience (empirical basis) of the committee members at the time they were introduced. They are, effectively, a penalty factor for not "looking" for problems in the welds.

3 - this effect that you mention was fully understood within the Code Committees when Appendix 46 was implemented (I wrote 46-4, so I can confirm this with 100% certainty). Using the EP method within Protection Against Plastic Collapse allows you to take advantage of the full elastic-plastic capability of the material. One of the cautions that I would have for you for using the Limit Load method for demonstrating Protection Against Plastic Collapse, is that this method is incapable of demonstrating any serviceability criteria. This serviceability criteria is critical for determining whether or not a design can "tolerate some straining"/deformation. So, yes, go ahead and use the Limit Load method to set your thickness, but to check whether or not you have excessive straining/deformation, you need to use the EP Method.

Is it the intent of the code that by using elastic-plastic method using note G5 material properties be not available?

No - not at all. The allowable stresses are not used in the development of the EP stress-strain curve.

 

[Paulettaa]

Thank you TGS4 for your explanations.

the consensus of the Section VIII Committee (both Sub-Group Design and the Main Committee) was that if the designer (engineer) wanted to use DBA, then they needed to ensure that sufficient inspection was performed to ensure that a weld joint efficiency equal to 1.0 could be assured.

I must say that it would be better to think about incorporating E < 1.0 in DBA since it is not always a matter of special design or care regarding sensitive equipment that encourages DBA. Sometimes, other reasons like the geometries outside the scope of DBR forces the designer to do something about it and most of the times they have nowhere else to go but to use DBA.

Anyways, with many thanks for helping me I had other questions regarding elastic-plastic method for buckling analysis. In Part.5 5.4.1.3 the capacity reduction factors are as follows:

(a) For unstiffened or ring stiffened cylinders and cones under axial compression

does this also include compression due to moments (like those arising in seismic loading?)

(b) For unstiffened and ring stiffened cylinders and cones under external pressure

does this include combination of axial compression due to seismic load and external pressure?

(c) For spherical shells and spherical, torispherical, elliptical heads under external pressure

For all cases it is either no stiffeners or ring stiffened cases. Does this mean if there are vertical stiffeners applied in an equipment, then elastic-plastic is the only way to go?

Last, I perform an eigenvalue linear buckling analysis and give the mode shape deformation results (scaled to the fabrication tolerances) as imperfection to an elastic-plastic nonlinear analysis and run the calculation. If the FEA converges does that mean that protection against both plastic collapse and buckling are achieved and no extra calculation for protection against plastic collapse is required?

 

[TGS4]

Paulettaa - I don't disagree with you. However, the consensus of the Code Committee was otherwise. However, non-unity weld joint efficiencies are addressed in API 579-1/ASME FFS-1 for fitness-for-service use.

To address your other questions:
Axial compression 5.4.1.3(a) does NOT include compression due to moments. Unfortunately.
External pressure 5.4.1.3(b) does NOT include the combination of axial compression and external pressure.

EP is the only way to go in these situations. Likewise with vertical stiffeners.

The buckling rules are going to be updated in the 2023 Edition. It's a just a little bit more detailed than what you listed (eigenvalue buckling analysis and scaling the eigenvector as an initial imperfection), running the EP buckling, but that's still basically it.

Unfortunately, the Plastic Collapse and the Buckling analyses are completely separate (with their own design margins), so you will need to perform the individually.

 

[Paulettaa]

Unfortunately, the Plastic Collapse and the Buckling analyses are completely separate (with their own design margins), so you will need to perform the individually.

Thank you again for these helpful comments.
Still I cannot understand what is the difference between design margins in EP analysis for plastic collapse protection and buckling protection when they both use the procedure in 5.2.4 and use the load factors as presented in table 5.5. the beta factor in table 5.5 is independent of the mode of failure (buckling vs plastic collapse)

TGS4, This is where I cannot understand the difference between the requirements of 5.2.4 and EP analysis for buckling.

(c) Type 3 – If a collapse analysis is performed in accordance with 5.2.4, and imperfections are explicitly considered in the analysis model geometry, the design factor is accounted for in the factored load combinations in Table 5.5. It should be noted that a collapse analysis can be performed using elastic or plastic material behavior. If the structure remains elastic when subject to the applied loads, the elastic–plastic material model will provide the required elastic behavior, and the collapse load will be computed based on this behavior

When the procedure is the same, the load combinations are the same, load factors are the same and acceptance criteria is the same then the design margin has got to be the same. The only difference that I see here is the introduction of geometrical imperfection as a perturbation for buckling analysis.

 

[TGS4]

Paulettaa - please contact me offline (my contact info can be obtained by clicking on my name). My assertion that the two analyses are separate is based on the 2023 rules. I am happy to share that rules with you before the official publication.

 

[Paulettaa]

TGS4
That would be so nice of you. I sent you an E-mail through my G-mail contact. I am also a member in RDG of BPV VIII. However, I do not know if it is possible to make contact through C&S Connect tools. If it were possible it would have been great.
Anyways thank you so much for helping clarifying code issues especially Div.2 related design problems.