Proper method to check the flange leakage 1

[farzam]

Hi All,
In order to check the flanges of the piping system I use the Pressure Equivalent method(concerning leakage).Now I face a problem: Since the margine between the design pressure(19barg) and working pressure of flange 150#(19.6barg @50C) is low most of the flanges have been failed. If i try the ASME VIII App.2 to check the flanges some of the flanges will be failed(even i do not apply loads).Regarding the pressure-temp. 150# flange is adequate, but i wonder how should i prove to the client that the flanges have no problem concerning the leakage due to external loads.

Thanks in advance

 

[C2it]

Your method is not wrong if it works for you, but it's very conservative. Either use equivalent pressure + internal pressure then do an ASME VIII Div 1 App. 2 stress check or the newer ASME VIII Div. 2 (2007) stress check, which directly incorporates external loads.

 

[farzam]

C2it,

I donot want to follow App.2, some of the flanges can not meet the App.2 requirements.ASME B16.5 flanges are ok based on the Pressure-temp. only to check the loads I'm looking for the proper approach.

Thanks

 

[TGS4]

If you are using spiral-wound gaskets, then this paper may be of some use to you:

http://asmedl.aip.org/getabs/servle...819000059000001&idtype=cvips&gifs=Yes&ref=no.

The papers referenced in that paper are also useful to your issue.

Long and short of it is, with external bending moments and the equivalent pressure method, for many situations, you could probably go to twice the rating pressure with design internal pressure + equivalent pressure.

 

[racookpe1978]

It appears you are trying to prove a flange rated for 19.6 bar can be used successfully for pipes carrying 19.0 bar, then find many fail when checked against real world pressure, weight loads, strain, and deflection.

Shouldn't you go to a higher-rated flange? I suspect your client is right: You are too close to the limit, and your pipes are not lying flat on the floor just subject to static pressure.

 

[TGS4]

racookpe1978, these flanges are often more than enough to handle additional loads. The "overly" conservative approach is to limit design pressure + equivalent pressure to the rating pressure. There have been numerous FEA studies and the above-noted paper that say otherwise.

I've been down this road before. The ability of B16.5 flanges to withstand additional loads to internal pressure is well demonstrated in the literature.

 

[C2it]

Farzam,

You did not read my suggestion. ASME VIII Div 2 )that's Div 2, not DIv 1 App. 2) handles external loads directly.

If you still have problems then try EN1591 / EN13445 App G(same) which essentially does a mini FE calc of flanges, bolts and gasket as an elastic body. You will need software to do it, it's seriously complex.

 

[TGS4]

One caveat on my earlier post - the bolt load must be sufficient. What constitutes sufficiency? I would say that a minimum bolt stress of 50ksi is a start. However, for problem flanges such as NPS 8, Class 150, you might want to look at bolt stresses in the 80ksi range (all of these are target assembly bolt stresses).

Review Division 1, Appendix S and PCC-1.

Also, C2it, Division 2 does not provide for a method to evaluate flange leak-tightness. Part 5 is good for stresses, but there are no rules for leak-tightness. Of course, the same thing could be said for Division 1, Appendix 2. (BTW, the stress evaluation calculations in Appendix 2 don't correspond to anything physical - it's only a DESIGN process, not an evaluation process).

 

[C2it]

TGS4,
I agree that ASME flange methods are stress based, but they do make an attempt to control strain with the rigidity factor. As I said, EN1591 is superior in that it addresses both stress and strain and the elastic body as a whole.

How would you define an 'evaluation' process as compared to a 'design' process ?

 

[TGS4]

C2it - we are in complete agreement. I am a BIG fan of EN-1591 myself.

I define a "design" process (such as in Appendix 2, Division 1) as a method that will result in a robust design. The stresses and other values calculated in the procedure may or may not have reference to anything physical (such as stresses). Having performed FEAs for B16.5 flanges that fail the Appendix 2 calculations, I can say with a high degree of certainty that the "stresses" that are "calculated" in Appendix 2 do not even remotely resemble the stress distribution pattern in the flange. Not even close. However, if you design a flange from scratch using the Appendix 2 methodology, you will end up with a rather robust design - experience says so.

An "evaluation" process is one where you perform calculations AND the results of these calculations, such as stresses and deflections, actually correspond to what's physically happening, and relate to a failure mode that needs to be protected against. As demonstrated by this paper -

http://asmedl.aip.org/getabs/servle...2819000051000001&idtype=cvips&gifs=Yes&ref=no

- flanges don't actually fail due to overstress. Rather they fail due to leakage. Now, they leak due to deformation, which is related to stress, but only insomuch as it relates to exceeding yield and having plastic deformation.

Does that answer your question?