Hydro test longer than required in API 598 1

[MoReSo]

Hello,

2" A105 forged valve failed during the hydro test as revealed a pinhole defect. the required time per API 598 is 15 seconds while the test we done was done for 10 minutes (40 times more) at the required pressure (1.5 MAWP). the test was done just on the valve separate from the piping system

The manufacturer doesn't accept as he mentions that it failed because it tested for longer time.

However I say that this valve is going to be welded in a piping system per ASME B31.3 which requires 10 min. So if the valve failed during individual test then it will fail in the piping system as well. no matter how long the test is per API 598 as the valve will be tested for 10 min in the piping system later.(it might be lower pressure for linger time in B31.3 as the MAWP for B31.3 is different to the MAWP in API 598 ????)


My question is: is this acceptable to hydro test a valve for much longer than required time specified in the API 598.
If the answer for above question is NO then what happens to this valve during piping hydro test as this valve is inline welded.

Does the manufacturer allowed to reject the test because of longer time?

please advise.

Materials and Welding Eng.

 

[zdas04]

The codes specify a maximum MAWP, not minimum. So if one code calls a system 700 psig and another code calls it 650 psig, you would be within both codes if you set the MAWP at any number less than or equal to 650 psig. I've had many situations where the strongest component in a system could have had a 1440 psig MAWP, but some other component or consideration caused me to call the system 280 psig. Perfectly legal and acceptable under the code. What I'm trying to say is that the design engineer defines the MAWP with guidance from the code. Not the other way around.

As to test duration, the code specifies a minimum duration for a test. If API 598 says 15 seconds (it's not a code I've ever used, so I'll go with your understanding), then if you stop the test in 10 seconds it is not a valid test, but if you run it for 24 hours and then get a failure it is still a legitimate API 598 test. The manufacturer is just trying to blow smoke up your pant legs. I would: (1) contact him via my lawyer to get the valve replaced; (2) never, ever, ever purchase another of his products; (3) take every opportunity I could find to make his name and shady practices widely known.

Finally, the valve is warranted against manufacturing defects. A pin-hole leak is a manufacturing defect. There is no way that he can claim that your 10 minute test caused the hole, the test just discovered it. The manufacturer is TRULY an idiot for even starting down this road. He can't win if you push it (and you should push it), he is just trying to intimidate you into buying another valve. You really need to see how fast he'll crumble when he gets a strongly worded letter from your attorney.

David Simpson, PE
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. —Galileo Galilei, Italian Physicist

 

[MoReSo]

Thank you Zdas04.

I think you are right. but at the same time manufacturer is right too as they did the test and no leakage observed and I believe they followed the valve requirements of API 598. The manufacturer had not been asked to comply with ASME 31.3 requirements.

Why duration of hydro test in API 598 is so short in relation to in-line hydro test duration? the reason might be the difference in MAWP. API 598 ask to run the test at higher pressure and shorter time while in ASME B31.3 the MOP is lower and this allows longer time for the small valves to pass the test. Please advise if this is not the case.

Materials and Welding Eng.

 

[btrueblood]

"Why duration of hydro test in API 598 is so short in relation to in-line hydro test duration?"

Because manufacturers pay their employees by the hour. I.e., there is a trade-off in terms of quality (smallest detectable leak in your case) and inspection/test duration, where time = money. As Zdas said, all (good) manufacturers realize their inspection/testing programs are not perfect, and cannot detect every possible leak, and most will not even blink if you tell them you have a leak - they'll replace or repair it under the warranty terms.

 

[zdas04]

The manufacturer is NOT right. Not in any way, shape, or form. The fact that the test they did failed to demonstrate the manufacturing defect is a failure in their process, not proof for all time that the leak was not a manufacturing defect. Pin hole leaks can get plugged up with dirt, grease, salt, gunk, paint. When pressure is applied, that crap starts moving, but it can take days to migrate far enough to create a leak (if I was the manufacturer I'd be delighted that the leak showed up in the shop before the valve was welded into the line). 10-15 seconds was obviously too short to allow the gunk to migrate. No big deal. That is, no big deal until the manufacturer stupidly made an issue of it. His only reasonable response to a valve leaking to atmosphere during client's acceptance test is "Sorry for the problem, a new valve will be on the truck in the morning". That is the response of all reputable valve manufacturers. For your guy to say that a multi-minute test somehow invalidated the warrantee is obscene.

David Simpson, PE
MuleShoe Engineering

In questions of science, the authority of a thousand is not worth the humble reasoning of a single individual. —Galileo Galilei, Italian Physicist

 

[LittleInch]

API 598 is a testing specification, not a valve specification. The theory behind the test is similar to that for pipes - the duration is minimised to allow mass production to take place without testing causing a major hold-up. The shell pressure test of a valve, like the pipe, is just there to find gross manufacturing defects, that's all.

The pressure used is just compatible with all main codes, e.g. ASME B 16.34 (valves) quotes 1.5 x class pressure for the test as does 31.3, though it tends to be a lower MOP.

zdas04 is completely correct when he says it is total irrelevant how long the test is. There are no elements which are time dependant. The excuse given by the vendor is complete and utter rubbish for a valve which is brand new.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[MoReSo]

Thank you for all of your explanations.

The manufacturer accepted the failure but two questions still remains:
1. Are we allowed to test the valves much longer than API 598 duration? if instead of 15 sec we test the valve for 10 min and valve failed after 15 sec by leaking (please forget the pinhole defect for now and consider it as general) then the valve is acceptable or not? the leak might happen due to loosening bolts or destroying the packing, as this 2" forge valve is not designed for that pressure(1.5*MAWP) for long time

2.If we test the valve and the valve passed API 598 for 15 sec; then I start debating and say that " it is too short! and ask the manufacturer for the answer of this question: what will happen when we weld it to the pipeline and want to hydro test the pipeline before commissioning? the pipeline hydro test will not be for 15 sec and will takes much longer?? is this a valid question.

I believe that the answer of the second question is: the MAWP when we hydro test the pipeline is much lower than the MAWP when we test the valve per API 598 so the valve can stand the hydro test (pipeline) for longer duration.

need your guidance!

Materials and Welding Eng.

 

[bcd]

Hey. I think you are missing the point. Maybe the valve supplier has made you confused.

API 598 is just a quick quality check. It is not perfect because the test duration is short, and this test in no way proves the valve is OK to use in your plant. Again, it is just a quick quality check that can catch major mistakes.

Many manufacturers recognize that a short shell test with water is not sufficient to show small pin hole leaks, so they do a test with lower pressure air first. Testing with air after exposing the valve to water does not work because the water can plug up the small pin holes so they do not leak during the air test.

Ultimately, you are going to put this valve into a pipe where it can be exposed to pressure for years, so any statement from the manufacturer that claims his 15 second test proves the valve is good is completely and absolutely wrong. The supplier must replace the valve...period. You should not allow any further discussion with them on this topic, and you should not pay their invoice until they replace the valve with a good one. And have them test the replacement valve with air pressure first before you accept it.

 

[LittleInch]

If you look at the test spec, it specifically has this test as a shell test and leakage through the packing is not a cause for failure or rejection. I repeat, as others do, the length of the test has NO IMPACT on the valve being tested to this specification. The valve is not designed to operate at that pressure which is why you have it open during the test and ignore leakage from the packing, but in terms of strength the shell or body is designed to withstand the hydro pressure without mechanical failure or exceeding the Yield stress.

If you design it right then the test pressure of the system can easily be very close to 1.5 X class rating - just depends on your code and design pressure.

The API 598 test should ensure that your valve won't fail catastrophically (break apart), but 15 seconds is nowhere near long enough to detect a pin hole leak in the casing. Exactly the same philosophy applies to pipe - the tests there last 10 seconds and do the same thing - i.e. find gross defects only.

The failure rate of these things is generally low enough that the occasional cut out and replacement of a faulty valve / pipe is worth it compared to the increased cost and time if every valve / pipe was tested for, say, 5 minutes each.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[zengx]

MoReso,

What's your valve rating? and what's your system MAWP?

Mohammed Diab.
Inspection Supervisor - Saudi Aramco.

 

[gustorf]

You may also look at it from another angle:
There is nowhere mentioned, when the 15 seconds have to START.
So it's up to you during testing to ask for 10 minutes stabilizing of pressure, before you start counting the 15 seconds....