Bellows Testing 3

[Dyabolikal]

Hello

I am a Valve technician that has been tasked with updating our companies work procedure for safe testing of PSV's. We currently refer to ASME I & VIII, API 527 & AS1271 for testing criteria's. I am seeking out information in regards to the testing of bellows, as this seems to be something that changes from workshop to workshop. Is there official documentation eg Standards that cover the testing of bellows on a test bench or related testing device.

I get the impression that this is one of those grey areas where there are unwritten industry guidelines that explain the test criteria but nothing official from a regulatory body such as ASME. If this is something that is a manufacturer to manufacturer thing, I will chase information from our OEM but I would prefer a standard over that.

Any assistance would be greatly appreciated.

Thankyou for your time.

 

[The Obturator]

There is no known industry standard for bellows testing for Pressure-relief Valves.

ASME VIII Will give you the minimum requirements to be met for a PRV, but will not detail any bellows requirements. API-520 etc., Will just advise on the application and use of bellows.

Pure bellows design is a manufacturer proprietary thing and will differ with manufacturer, including the bellows vendor(s) they use.

The PRV manufacturer will design a bellows with minimum requirements based on criteria such as maximum and minimum bellows tube size, material, relative spring rate to PRV spring, free length and any manufacturer free issued parts such as the disc holder and flange etc. There will also be a specified pressure rating and from which the bellows vendor will produce a test procedure for approval and acceptance by the PRV maker.

The bellows unit gets delivered to the PRV manufacturer normally with test and material certification and that's about it

What you are trying to determine is what tests are applied thereafter. Right?

During PRV assembly there are typically no other special tests done on the bellows. However, the integrity of the bellows is checked. That is to say, checked for any leakage.

Besides the set pressure and leakage tests, there is also a 'Shell Pneumatic Test'. This is where the body, bonnet and cap are pressurised to the same nominal pressure (typically 7 barg/100 PSIG) via the outlet flange and all seals and joints tested for tightness. This is a test actually mandated in ASME VIII as checking the integrity of the secondary pressure zone. When it is a bellows PRV the bonnet is pressurised at the same time through the bonnet vent with no more than say 10 PSIG pressure differential between the body and bonnet chambers. Note that the test does does not apply to open bonnet design (or where the cap is not gas tight).

Straight thereafter, is the bellows integrity test. This is simply applying a nominal 1 barg/15 PSIG through the PRV outlet and checking the bonnet vent for any leakage. Any leakage is cause for failure. Whether you apply 1 bar , 10 bar or higher does not matter, the bellows is ruptured somewhere and has failed. That's it!

Note also that not all manufacturers follow the above for bellows PRV's, but will pressurise body only for both shell test and bellows integrity.

Of course the PRV assembler will visually inspect the bellows for any apparent damage or crack etc., before assembly, but apart from the integrity test that is all that is done by a manufacturer. It is doubtful the manufacturer would supply the bellows vendors test report as this contains proprietary information.

The bellows vendor only carries out a test on bellows only. It's a one time thing only. Same as doing hydro test on PRV body. You do it once and that's tt.

So from a maintenance point of view, I would suggest you follow the bellows integrity test which may or may not be in PRV manufacturers maintenance manuals.

*** Per ISO-4126, the generic term 'Safety Valve' is used regardless of application or design ***

*** 'Pressure-relief Valve' is the equivalent ASME/API term ***

 

[Dyabolikal]

Thank you for the reply

I figured that would be the case, as there was no quick links to standards when looking online. I have seen a couple manufacturers test documentation, one suggested testing to 30 psi (200 kPag) but that test had the bellows removed and tested independently from the valve, and another document suggesting 20 psi as per normal body test procedures.

It is known... well for at least the companies I have worked for, that low pressure is the way to go as we do not want to damage the bellows through over pressurisation, but also it has been suggested that the bellows are tested to the relevant backpressure on the man plate to prove integrity under normal backpressure conditions. The reality though is if the bellows have sustained damage it will show during the bubble test.

The other question I have are due to durations. The work procedure that I am adjusting states a 3 minute wait while doing the bellows test. I find it quite strange they had suggested that length of hold, I mean I understand that they want to wait for the spring chamber to fill with air if the bellows are leaking, but I doubt it would take as long to do that. Would 1 minute be enough or is the document on the right track.

Our procedure also talks about a balanced bellows test where the port in the bonnet and the outlet is pressurised with a maximum 15 psi differential to test seal integrity between the body and the bonnet, I would also presume the cap seal as well. I have seen this test performed before and has always confused me in regards to its relevance. I understand that the seals are there for a reason but if the bellows fails in service then it would leak from the port in the bonnet, does that not just nullify the test, or is there something I am missing.

I could imagine that the bellows could leak from the top-works where it interacts with the body/bonnet seal is that what this test is trying to determine. If that is the reason for the test then its either the seal or the bellows and trouble shooting that would prove difficult without either completely replacing the body seal or the bellows, both of which are not favourable options in a maintenance operation as spares cannot be ordered in bulk and the lead times usually can be from 6-12 weeks

The standard body test or as you called it the 'Shell Pneumatic Test' we do perform on conventional valves and that's separate from bellows testing. I do have a question in regards to that though, as shell testing is also apart of our work procedure. As our company purchases our standards I am not sure if the ASME standard I am using it the up to date version. I cannot to to SAI Global to check as ASME wants you to buy of them. The secondary seal test under section UG136 (off the top of my head I think its UG136) states the pressure to be 200 kPa but from reading above you have stated it to be 700 kPag. I wouldn't mind clarification on the discrepancy as I don't want further arguments in the future from techs in my company, and if the ASME standard I am using needs to be updated I will inform our QHSE team.

 

[The Obturator]

For the bellows integrity test, I did earlier state 1 bar / 15 PSIG as a nominal test pressure, but the pressure needs only be low. 2 bar / 30 PSIG may well be specified by others. A low pressure will generally show up a leak path better than a higher pressure. A higher pressure can possibly seal up a leak path. Such a test is not in ASME and manufacturers have developed their own procedures. Its better to test the bellows as it is installed in the valve. That way you are testing its integrity and seals against the body and bonnet joints too during the shell pneumatic test.

I can't understand why some party has 3 minutes for a bellows integrity test. I can only think they have used the hold times, as for seat leakage testing in API-527, to cover themselves. In reality as you say, if the bellows has the slightest damage, it will show up straight away through the bonnet vent over which you would place a soap bubble.

Your 4th Paragraph, actually describes the Shell Pneumatic Test I referenced earlier. Most test at 7 bar / 100 PSIG. Its relevance is quite important. It is to test the integrity of the seals,- between nozzle and body, body to bonnet seal. set screw(s) to body, cap to bonnet and lever packing if fitted. ASME VIII UG-136 requires this stating it is a requirement for a PRV discharging into a closed system. This is called the secondary pressure zone. If you have a superimposed back pressure you don't want leakage of a potentially toxic fluid coming out the seals. The pressure to be used stated by ASME is for it to be at least 30 PSIG. I've attached an extract from current 2019 ASME VIII UG-136 on this. 7 bar / 100 PSIG is used by most manufacturers. Will not apply to open (regular) lever construction and open bonnet PRV's. There are some manufacturers that will not include the bonnet and cap test if the PRV is of bellows design. They say these are not secondary pressure zone parts (I disagree with that).

Note that ASME VIII UG-136 etc., is for obligatory requirements for PRV manufacturers to meet and not specifically maintenance, though some requirements are carried over. You should also be familiar with API-576 - Inspection of Pressure-relief Devices, which also highlights the requirements of the shell pneumatic test, but not the bellows integrity test.

The shell pneumatic test should be done after the set pressure (setting) teat. This should be immediately followed by the bellows integrity test for any leakage out of the bonnet vent. Finally the seat leakage test is performed - a damaged bellows would not allow any seat leakage to show up.







*** Per ISO-4126, the generic term 'Safety Valve' is used regardless of application or design ***

*** 'Pressure-relief Valve' is the equivalent ASME/API term ***

 

[Dyabolikal]

Thank you once again

You have been a great source of information, currently the discussion revolves around having the pressure at 30 psi across the board as this standardises all secondary seal tests. Our company are obsessed with keeping to the letter of what is stated (due to NATA accreditation) in the standards even though most understand that these are the minimum requirements and that exceeding them is generally ok to a certain point, case in point that they would insist that a 15 second test is valid even though they supply leakage rates set at a per minute cycle, I'm sure that we have all seen that before though.

I have decided to just go with at technicians digression in regards to hold times for secondary seal tests, its generally accepted that a minute is enough time but other factors may cause that hold time to extend by allowing the techs to be in control and not forced to accept a time in a procedure will ensure that the tests are carried out correctly.

So the pneumatic shell test is actually what our procedure states as a balanced valve test interesting, our normal back pressure and/or shell test only has the outlet pressurised if there is a port in the bonnet its normally plugged if the valve is of a conventional design, and if its a bellows design the test is waived in favour of a bellows test. Now seeing what you are saying we may be able to incorporate the test within our normal backpressure and/or shell test to include if the valve has a bonnet port this test should be performed in lieu of the standard shell test.

Your last point is what I would consider quite important given that it is drilled into techs that the order of tests should always be lift test, leak test, shell/backpressure tests bellows tests are generally left to the end but moving it up before the leak test would make a lot of sense. I could appreciate the counter argument would be well if the bellows test fails all tests are considered a failure but like you stated earlier if too high of a pressure is added it could cause a false pass to occur even though 90% of bellows are fine that 10% remaining is just too high a risk given that these valves are a last line of defence.

Once again thank you for your time

Dya

 

[The Obturator]

You should ideally do the shell pneumatic test on both conventional and bellows PRV's. ASME Clearly states to test the secondary pressure zone. What if the bellows should fail in service? Fluid would come out of the bonnet vent (which should be piped away if the fluid is toxic) but there would still be an amount of pressure in the bonnet. Both conventional and bellows type bonnets are tested like this by most manufacturers. You earlier stated that you already do the test by having no more than 15 PSIG differential between the body and bonnet. You simply do the bellows integrity test straight after with a lower pressure, by simply depressurising the bonnet and seeing if there is any leakage out of the bonnet vent. Doing this as routine and standard avoids any misunderstanding.

Note that not all manufacturers conventional PRV's have plugged bonnet vents. They are simply not drilled and tapped. Major manufacturers have a plugged bonnet for the conventional design where the same bonnet can be used for a bellows type.

*** Per ISO-4126, the generic term 'Safety Valve' is used regardless of application or design ***

*** 'Pressure-relief Valve' is the equivalent ASME/API term ***