U2-(g), appendix 46

[Vrpeng-tips]

Hello,

Unit - Div.1 U-Stamp heat exchanger
Most of the design of the unit can be done using div.1 rules (Shell, Nozzles, head,etc..)
For a fixed tube sheet div.2 part 4.18 applies. And design can be completed using DBA.

In a rare case when we have tube sheet geometry out of scope and there is need for thickness validation, do we absolutely have to do DBA per div.2 part 5?
Or we can do the FEA and validate the T.S thickness and stress per div.2 part 4?

Thank you in advance.

 

[Vrpeng-tips]

correction.
For a fixed tube sheet div.2 part 4.18 applies. And design can be completed using DBR.

 

[Vrpeng-tips]

Hello Experts,

Any insight on this would be greatly appreciated. Thank you.

 

[TGS4]

I'm not 100% certain of what your actual question is. Please restate your question.

 

[Vrpeng-tips]

Hi TGS4,

Thank you for your response.

For Div.1 ed.2023 fixed tubesheet heat exchanger, all components are designed using Div.1 design-by-rule methods. However, for a non-circular fixed tubesheet which also acts as a support, Div.2 Part 4 design-by-rule formulas cannot be used. Does this mean Div.2 Part 5 (design-by-analysis) is the only option?

Before in ed.2017 the U2-(g) allowed accepted methods ( satisfactory to AI) and my company used FEA to derive the tube sheet thickness using allowable criteria set in UHX 13. (i.e 1.5S limit for design load cases and 3S limit for operating load cases).

Hope this is clear.
Thank you for your time and consideration.

 

[IdanPV]

Haven't the rules for heat exchangers design moved to VIII-2 already?

 

[Vrpeng-tips]

Hi IdanPV,

Thank you for your response.

Yes, the rules for HX have moved to div.2 part 4. Which is design by rule method and shall be used only for a flat circular tubesheet.

Question 1- For a non-circular flat tubesheet (used as support, no saddles) we MUST use div.2 part 5?

Thank you.

 

[Vrpeng-tips]

Not sure if I am being clear or not. Let me rephrase my questions & background.

fixed tubesheet heat exchanger, config. b, tubesheets integral with shell, extended as a flange and gasketed on the channel side.

Q.1 Is it valid to assume the non-circular tubesheets to be a circular and design the thickness using sec.viii div.2 part 4 para 4.18.8?

Q.2 if the above assumption can not be used, my only option is sec.viii div.2 part 5?

Q.3 if at all for any situation we end up using FEA to “design the thickness” of a component, do we must abide by sec.viii div.2 part 5 rules? i.e checking for all failure modes. Not just elastic analysis.

I ask these questions because, the U2-(g) has been improved in 2019 along with the Mandatory appendix 46. And app.46 tells to use either part 4 DBR or part 5 DBA. Which i believe only allows the use of FEA if doing DBA.

Before 2019, if you were to use U2-(g), you could use the FEA to “design the thickness” without following sec.viii div.2 part 5, of course with the acceptance from the ASME Inspector.

Here is what my interpretation is,
1. Design a component using Sec.viii div.1. Do not use FEA.
Example, simple pressure vessel.

2. Design a component not covered by above and/or if code permits using Sec.viii div.2 part 4. Do not use FEA.
Example, radial nozzle in a cylindrical shell.

3. Design a component and/or validate loading condition not covered by above (1&2) using sec.viii div.2 part 5. Use FEA.
Example, radial nozzle in a cylindrical shell subject to external loadings. Also, pipe support attachment to shell in a pressure vessel subject to seismic.

Hoping to get some clarity.

Thank you.

 

[IdanPV]

Hi,

Keep in mind that you do not have a Division 2 vessel, but a Division 1 vessel which needs to be desgin by the design rules of Division 2 (because of the change in Part UHX).
That doesn't mean your vessel is a Div. 2 vessel.

If there are no design rules in Div. 2 than U-2(g) applies and Part 5 can be used (among other options).

A.1 - No, that doesn't seems reasonable to me.
A.2 - No, you have other options, see U-2(g). But, if FEA is chosen then Part 5 shall be used.
A.3 - See App. 46.

 

[Vrpeng-tips]

IdanPV, I really appreciate your reply. It actually made me RE-READ the code. I also read your other post https://www.eng-tips.com/viewthread.cfm?qid=514178 where it gave me more clarity (not to jump straight to FEA part 5 and look for possible alternatives as mandated by code).

Yes - Vessel is div.1 Heat Exchanger.

1. As per Sec.Viii Div.1 UHX-1(a)(4) -

understood and abiding.
2. As per Sec.Viii Div.2 4.18.4(a) -

Since my tubesheet is non-circular, the design rules from 4.18 can not be applied. so reading 4.18.1 scope.

3. As per Sec.Viii Div.2 4.18.1(b) -

understood and reading 4.1.1.2.

4. As per Sec.Viii Div.2 4.1.1.2-

understood and as per above, I shall perform FEA in accordance with Part 5.
Please correct me if I am wrong.
Thank you all for your patience and help. looking forward for the response.

 

[IdanPV]

My understaing is that because you do not have a Division 2 vessel, but a Division 1 vessel.
For a Division 1 vessel you do not care about 4.1.1.2.1/2 and when no design rules are available you go to U-2(g).

This means: When design rules are neither given in Division 1, nor in Part 4 of Division 2, you apply U-2(g) and use another generally accepted method or FEA in accordance with App. 46.

 

[TGS4]

Thank you for the clarification.

In your step 3 - where you reference 4.1.1.2 - you need to stop and read UHX-1(7)(b) first.

UHX-1(7)(b)The rules in Part UHX cover the common types of shell-and-tube heat exchangers and their elements but are not intended to limit the configurations or details to those illustrated or otherwise described herein. Designs that differ from those covered in this Part shall be in accordance with U-2(g) in lieu of 4.1.1.2 referenced in 4.18.1(b).

UHX-1(7)(b) says that you shall go to U-2(g). Now, if you choose to use U-2(g)(1)(-a), which points you to Appendix 46, then you end up in the same spot. However, it doesn't preclude you using U-2(g)(1)(-c).

 

[Vrpeng-tips]

Thank you IdanPV and TGS4.

thank you for pointing out UHX-1(7)(b), my bad I didn’t read that far.

So I am left with,
Option one = U-2(g)(1)(-a), use of Appendix 46. Which give two further paths. Part 4 and/or part 5. Since Part 4 didn’t work, I can use Part 5.

Option two = U-2(g)(1)(-c), use recognized and generally accepted method. This is what we are using right now.
But i am afraid it is not a valid method (as of today). So we only perform elastic stress analysis. Analyze 4 design and 4 operating load cases (as per Sec.viii div.2 part 4 table 4.18.7 & 4.18.8) and check results against the stress limits per Sec.viii div.2 paragraph 4.18.8.4 steps 7,8,9etc. Allowable stresses are per sec.viii div.1 from sec.ii part D. This method is also accepted by inspection.

My understanding suggests that option one shall be used and that option two is not valid (as of today as U-2(g) has become more clear).
But I am finding it difficult to justify it. Am I wrong in my understanding?

Thank you for your precious time and input.

 

[IdanPV]

What make you think that the method of U-2(g)(1)(-c) is not valid?
Note that if Div.2 Part 5 is used, all the failure modes shall be considered, see 46-4(c).

 

[Vrpeng-tips]

I say the method is not valid because Div.2 part 5 is not being used.
1. Not checking all modes of failure.
2. The stress limits are not set as per Div.2 Part 5 table 5.3.
3. There may be other requirements to meet when using Div.2 part 5 (i.e weld details) that aren’t being discussed.

Question- How can you perform FEA to design the component without abiding to Div.2 Part 5?

 

[Vrpeng-tips]

OK. did some more reading and I think I have little more clarity.
So, FEA can be used under U-2(g)(1)(-c) to design component thickness if not covered by DBR, by performing only the elastic stress analysis and without strictly following Div.2, Part 5, but still referencing its stress limits, load combinations, and classification is a generally accepted engineering practice. And of course with inspection approval of this approach.

Users are free to use Appendix 46. 46-4 DBA, Div.2 part 5 to design a component thickness if DBR is not available and if they choose to do so. This can be seen feasible if one is trying to optimize the design for efficiency (both performance and manufacturing side).

question - have i got this correct?
My sincere apologies if I am stretching this and utilizing your precious time. just wanted to get some clarity.

Thank you.

 

[Vrpeng-tips]

Any feedback would be appreciated.
Thank you in advance.