ASME B16.6 Joint Separation Factors of Safety and Preload Joint Analysis 1

[DarcyDude]

Hello all,

New to posting so apologies if I mis-step. I found a few posts on this topic but none reached a satisfying conclusion before being closed (satisfying to me anyway!) My question is theoretical and is sort of two fold:

1) Does anyone know any resources citing the factor of safety to flange separation/leakage for a well installed ASME B16.5 flange pair? Reviewing the ASME doc it seems that design is stress/deformation driven, and not given explicitly in terms of separation. Notionally, from reading on the forum, the factor of safety is "small". From my short career I know it is at least 1.5 because I have witnessed a few hydrotests. Any resources, studies, or past experiences anyone can point toward that might narrow this down?

2) This is obviously a function of preload on the joint. This may be my inexperience showing, but at companies I have previously worked flange bolt torque values/procedures were captured in a company standard (which I was not involved in generating). Is there any consensus standard laying out bolt size/torques for given ASME B16.5 flange pair and gasket type? Or in the absence of a company (or gasket manufacturer) standard must an engineer perform a preloaded joint analysis?

Again sorry if this has been touched on, is in the wrong area, or if there is a reference I should have already reviewed! Please point me in that direction if any of the previous are true!

Thanks!

 

[XL83NL]

1. I remember a topic some years ago here which included a link to an old paper on how B16.5 once was created. Your answer may be in there, but maybe the answer is much simpler and other members can comment. I’m pretty sure though there’s no such as an FOS for flanges.

2. ASME PCC-1.

Huub
- You never get what you expect, you only get what you inspect.

 

[DarcyDude]

Thanks XL83NL for the PCC doc reference! So it does look like some joint analysis is generally required.

Re reading my question, my first item was maybe a bit unclear. I understand that the ASME design may not specifically dictate a FOS to joint separation, but if we were to theoretically pressure test an assortment of these flange pairs, what pressure would we expect to see at leakage? 2X the rated pressure? 3X?

Wondering if there is any info out there to get a better understanding of the joints robustness. Not that I hope to expose flanged connections to direct or equivalent pressure in this range, but just for my own knowledge.

 

[LittleInch]

Darcy,

I'm not sure you'll find what you're looking for.

Appendix A of ASME B16.5 (not 16.6 as per your title...) might help, but the key para in 16.5 I think is this one

"A flanged joint is composed of separate and independent, although interrelated components: the flanges,
gasket, and bolting, which are assembled by another influence, the assembler. Proper controls must be exercised in
the selection and application for all these elements to
attain a joint that has acceptable leak tightness. Special
techniques, such as controlled bolt tightening, are
described in ASME PCC-1"

In other words it's complicated and you can't find any one part of the elements which will give you a simple, here is a limit X, we won't go beyond 0.Y times X

Also you need to define "failure". A leak is often not a big issue, but physical failure / yielding of the components are. It's perfectly possible to find flanges exposed to 2 to 3 times their rating which have sprung apart, leaked and then sealed again when the internal pressure reduces. Have they "failed"?? is one drop a leak or do you need 10 drops or a hundred?

So I don't think there is an answer to your second question as this will depend on how much you tighten the bolts. Just enough to stop them leaking at the test pressure or to some pre defined limit of say 90% of the bolts yield strength? I would expect a well tightened flange to handle at least 2 x rated pressure without leaking "excessively", but that's just me.



Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[DarcyDude]

LittleInch,

I think that helped shed light on the situation. I haven't gotten a chance to dig deeply into PCC-1 yet but it appears, as the paragraph you cited from 16.5 alludes, that depending how you navigate PCC-1 you could end up with different joint preloads. And thus different flanged joint systems would experience different margins to separation. So no one size fits all answer, would have to analyze each joint to get a theoretical figure.

In this scenario I'm wondering specifically about excessive leakage (although other failure modes exist as you pointed out). I have worked on some projects (outside of ASME piping systems) where preloaded bolted joints were designed using methodology in NASA TM 106943. If the joint in question was providing a seal it was common to report a margin to joint separation (where load on a portion of your seal you had overcome preload). So in these circles anyway excessive leakage was thought to theoretically occur at joint separation.

Maybe one day I'll cave in and calculate the margin to separation for a joint designed with PPC-1 using the NASA methodology. Real life is obviously a totally different animal because, as your pointed out, bolt preload after assembly can be pretty variable unless rigorous controls are in place.

Thanks!

 

[XL83NL]

1. I remember a topic some years ago here which included a link to an old paper on how B16.5 once was created. Your answer may be in there, but ..

In below thread, go to BJI's MLPS-post (12 and counting) to find that paper.

https://www.eng-tips.com/viewthread.cfm?qid=454965

Huub
- You never get what you expect, you only get what you inspect.

 

[LittleInch]

The problem I hve with joint separation is tha that isn't real.

Any gasket or ring needs to have a certain force / pressure for it to seal between itself and the metal surface of the flange. If you think about a dripping flanged joints, it's often already under a lot of compressive stress and in no way could it be said to have "separated"

so that pressure is the critical one, but finding out what that is can be quite difficult.

You might try the welding, bonding and fasteners forum for a similar query as there are some people there much more knowledgeable about this then me or can point you in the right direction.

A conversation with a gasket vendor if you can find the right person might be a goldmine of information

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[DarcyDude]

LI,

You are definitely correct, I was not thinking about the scenario properly. Last night I started looking through ASME BPVCE SECTION VIII DIV 1 APPENDIX 2 (RULES FOR BOLTED FLANGE CONNECTIONS WITH RING TYPE GASKETS), which seems to handle the behavior you mention with factor m and y where y is the gasket seating pressure (I believe the pressure you referred to in your post). Looking further the major gasket manufacturers publish values for m and y. So using the ASME method you could define a pressure where you no longer have the needed compression in your joint for the given gasket material.

Caesar II has an application for for the BPVC SEC VIII DIV 1 APP 2 calculations. So I tried inputting the values for a b16.5 flange, with a given gasket material, and a pressure at the max B16.5 rating for said flange to get more insight. This predicted that you would exceed bolt and hub stresses FAR before reaching this reduction in compression. As I interpret the output you are breaking bolts before decompressing the gasket. I triple checked all my inputs, but I am still really skeptical of that result. As you pointed out, and from accounts from others, expected flange behavior should be to leak excessively prior to a mechanical failure.

So maybe I'll have to do the calculations again by hand to get better insight. Thanks for the recommendation for the other forum and gasket manufacturer. I'll try poking around there as well... but need to balance getting some real work done haha

 

[LittleInch]

It's a well known fact that ASME b16.5 flanges fail when you look at ASME Viii app 2 calculations.

But there are billions of asme b16.5 flames in use around the world so they do work!

Now why you're breaking bolts I can't say but maybe the asme viii is using more bolt force than pcc1 to get it to seal.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[BJI]

I think you would benefit from getting a copy of WRC 538, it is the basis for the development of ASME PCC-1 Appendix O. Appendix 2 calculations are only suitable for sizing a flange from a strength perspective, it makes no assurances of leak tightness. The ‘m’ and ‘y’ values do represent operational and seating factors for gaskets, but are only used in the context of flange sizing (design bolt load), it should not be used for leakage based design with actual assembly bolt loads. Target and maximum gasket stress values are not published in Appendix O but are published in many papers, from gasket manufacturers and some test data is available online. WRC 538 provides a method to determine more appropriate flange limits, allowing for localised yielding without exceeding flange GPD limits. However, for standard B16.5 flanges the bolt limits are already published in Appendix O.

The NASA method is pretty basic, only really suitable for clamping forces, couldn’t be applied to piping joints with ring gaskets. There are many published joint flexibility methods that are capable of this type of analysis, if you really want to go down that path.

As mentioned the B16.5 flanges often don’t pass the Appendix 2 sizing calculations. There is no one factor of safety, even if you could eliminate the many variables associated with joint assembly. Relative flange flexibility is not consistent across all sizes and classes, and the bolt to gasket area ratios can vary considerably. The bad actors are usually pretty well known, typically the configurations that are under-bolted.

If you increase the pressure or apply increasing external load, you will always see leakage before rupture. In fact, excluding any degradation to the bolting, you should not be able to rupture a bolted joint under any ‘normal’ pressure excursion. Even under high bending loads, some joints won’t leak before the attached piping buckles. You should be able to find some leakage testing covered under ASME publications.

 

[LittleInch]

billions of asme b16.5 flames

Ha, the joys of typing on a phone. The edit function has gone so just to correct this I really did mean "flanges"...

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[DarcyDude]

Hello BJI,

Thank you for the detailed response. I have not had a change to get into find detail, but after an initial skim through WRC 538 seems like the best resource I have yet come across for answering these questions.

If the ASME method is only suitable for flange design from a strength perspective, in your mind what is the preferred method for checking for leakage? The methods I am aware of are the Equivalent Pressure Method (which most seem to agree is highly conservative) and NC-3658.3 (Which is less conservative than Equiv. Pressure, but also seems to be an approximation, so unsure how it compares to test results). Are there other popular methods I am overlooking?

But seems we are reinforcing that a) ASME B16.5 flanges will leak before burst b) ASME B16.5 flanges do no provide a consistent factor of safety to leakage and c) determining factors of safety to leakage would require analysis or testing for the specific joint configuration.

I will look into ASME leak testing. That seems like the most direct answer.

Thanks!

 

[DarcyDude]

to clarify "If the ASME **Appendix 2** method is only suitable for flange design from a strength perspective..."

 

[DarcyDude]

Sorry it appears XL83NL may have already provided some guidance here:

https://www.eng-tips.com/viewthread.cfm?qid=350904

I will try to get my hands on PVP2013-97814

 

[DarcyDude]

So reading through the recommended documents PVP2013-97814 and WRC 538 both provide good insight into design and limitations of flanged joints, and talk through all elements (flange, bolting, gasket, etc.). Short reads and very helpful. Interestingly both were written by the same fellow. Both (especially WRC 528) give insight into the methods behind ASME BPVC SEC VIII DIV1 APPENDIX 2 and ASME PPC-1. WRC 528 focuses more on bolt loading of B16.5 flanges, while PVP2013-97814 focuses on external moments on b16.5 flanges. Both provide some FEA studies of these joints. After reading both I have a much better understanding the concepts at play.

Also was able to locate some test data on AMSE B16.5 flanges in the below publication. Its a little dated, but appears for steel ASME B16.5 flanges (depending on bolt load) one can expect a leakage pressure 2-4 times the rated pressure of the flange.

https://www.osti.gov/servlets/purl/7246900

Thanks friends! I feel satisfied we identified some good resources for understanding these joints as well as some notional margins to joint leakage!

 

[TGS4]

BJI - extremely well put. Thank you - couldn't have said it better myself. LPS for you!

DarcyDude, you have been directed to some excellent resources on the subject. I see that you noted that the author of the PVP2013 paper and the WRC Bulletin are one in the same. Look for additional papers written by him - he truly is an expert in the field.

I will note that the methodology in PVP2013-97814 has been incorporated into ASME Section VIII, Division 1 in UG-44(b) and Division 2 in 4.16.12.

 

[BJI]

I wouldn't call anything by Rodabaugh and Moore dated, any of their work would be very relevant today. Regarding the previous question, you should be able to find flexibility formulas published by these authors online.

The other author, Brown, has published these types of flexibility formulas in WRC 510. It could also be worth having a look at the method published in EN 1591. Other references are available but may be more difficult to source.