Pipeline Test Head 2

[blackegl]

Recently we had a customer approach us to design a test head for them to be used for hydro-testing pipelines. Now, they want to have the ability to reuse the test head over and over whenever they hydro-test a pipeline. So here is an example of what I am talking about.

Pipeline operating pressure: 1440 psig
ANSI Class rating of flanges on pipeline: 600#
Hydro-test pressure of pipeline: 2160 psig

Now if we design a test head and use 600# flanges, prior to it leaving our facility we would hydro-test it up to a 2220 psig. Can we take that same test head out into the field and hook it up to the pipeline and complete a hydro-test at 2160 psig, essentially taking the flange up to hydro-test pressure twice? So I guess my basic question is, can I take a flange up to hydro-test pressure multiple times or is it a one and done type thing? If I can take them up to hydro-test pressure multiple times is there a recommended number of times prior to switching out to new flanges? Also is there a difference, if the flange on the test head it permanently connected to say a valve and the flange is never disconnected verse the flange that connects the test head to the pipeline since that flange would be installed and disconnected each time a pipeline is tested? Or on the other hand, should the test head be designed with 900# flanges if we want to take the flanges on the test head up to the 2220 psig range more than once and then use a 900# to 600# spool piece to connect the test head to the pipeline that could be swapped out each time or after a couple uses? The test duration will be between 4 and 8 hours each time.

Thanks

 

[zdas04]

If you calculate the stresses, you'll find that your test head is around 30% of SMYS (probably 15% of actual yield stress). I would expect 50-60 million cycles at that pressure to be required to develop any fatigue risk. Build it with ASME B16.5 Class 600 flanges and forget about it. I've used similar test heads hundreds of times (certainly dozens, I really haven't kept track) without any sign of problems.

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

 

[blackegl]

Thanks zdas04,

What had me raising the question was the fact that ASME B16.5 Class 600 flanges have an operating pressure of 1480psig up to 100F vs the 2225psig hydro pressure up to 100F. There must be some good reasoning for having these two pressure designations and I am not sure what that is.

 

[BigInch]

Well I don't think it is very ppropriate to use a component that will actually be in operating-working condition when it is supplying pipeline test pressures at 2225 psig, when its true working pressure is really only 1480. Every time you test a pipeline, it is actually only the pipeline that you desire to test, right? Test components should also be supplied at their working pressure, which here appears to be 2225. Pipeline test pressure is actually working pressure for those testing component things.

The same logic suggests that the hydraulic pump you use to raise the pressure to 2225 could also have an actual rating of 1480 and be run up to 1.5 x its working pressure to reach the 2225 test pressure. If you had a test pressure set valve designed to release pressure higher than test pressures, it could also actually only have a 1480 rating???

 

[blackegl]

BigInch,

That is the way I looked at the problem, was if the test head is intended to operate at the 2225psig I figured we should go with the Class 900# flanges, however on the other hand, that same pipe line will be tested many times during its life up to the hydro pressure. I am unsure what the correct answer is, so that is why I asked the question, to see if there is some code or reference that I have overlooked that my shed some light on what the correct way to do it is.

Thanks

 

[BigInch]

INGAA 4.2.7 says,
Verify the pressure ratings of hoses, fittings, gaskets, and other manifold materials.

4.8 Test Manifold Construction

It doesn't precisely say that these things should have pipe test pressure as their working pressures, but I think it alludes to it, why else would they say what they do in 4.8.2, 4.8.7 and 4.8.9


Do you have material certs allowing pressure to 1.5 x working pressure?
Of course that is not applicable to pipe being tested, so what else would they be talking about if not the test manifolds and other test assembly components themselves?

http://www.ingaa.org/File.aspx?id=18981

 

[zdas04]

This discussion is nuts. It would make some sort of twisted sense to use thicker pipe that has an MAWP greater than 2250, same with the pump. An ASME B16.5 Class 900 flange has a different bolt pattern than a Class 600 flange so I guess to carry this case to its logical conclusion you need a Class 900 to Class 600 transition spool that you throw away every time? Give me a break.

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

 

[BigInch]

zdas please see 6.2.1 -> design pressure for pipe and components used to fabricate test headers shall not be more than 80% SMYS at test pressure.

http://documents.dps.ny.gov/public/...cRefId={347CDC53-D5E5-42B1-9A3A-58E9CCD3C01D}

Nothing prohibits welding a 900# flange on the pipe to be tested to make the connection, which could easily be blinded off, or cut out completely after testing.

 

[LittleInch]

I'm with BI on this one.

As far as I'm concerned the design pressure (or MWP if you prefer) of the test header is the test pressure. That's what it's seeing as its duty time and time again and hence should be designed for.

You don't need any special code issue, you simply need to apply the correct design pressure to the correct components.

In this case the pipeline has a design pressure of 1440 by the sounds of it, the test head has , IMHO, a design pressure of 2160 and hence a test pressure higher than that (depending which code you certify it to).

I would make the connection to the test header via a welded connection which is then cut off and either welded to the next section or to a flange (golden weld).

I can understand that test heads are often simply rated to the same as the thing it is testing, but I personally wouldn't allow that.

BTW which pipeline code are you working to? at 1.5 x MOP I can only assume you're B 31.8 location class 3 or 4?

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

 

[zdas04]

I don't know who you guys work for, but if I proposed welding a Class 900 flange onto a built system, then cutting it off and welding a Class 600 flange onto the system after the test my clients would not only fire me by might very well have me committed.

I've seen tests to 150% of MAWP on Class 600 systems done with Class 300 test equipment and no one could figure out why I objected. Over the entire life of the test manifold that the OP proposed the manifold will be pressurized about 50-100 hours. Not concerned with creep strength. It will probably be pressurized and depressurized 25-50 times before it is lost or cut up into something else. Not concerned about fatigue strength.

People do a lot of sketchy things upstream of a test and nearly always get away with them. When something does break we all run around trying to find root causes, write a couple of scathing reports and then once again get the hell out of the way of progress. That is the world of gathering anyway. Jurisdictional lines may have more consequences, but I've seen test equipment go from my gathering system to both Kerr-McGee and Williams interstate pipeline jobs without modification.

The farthest I would go is to build a Class 900 test skid and a 900/600 (and/or 900/300 and/or 900/150) transition spool. And I would use that transition spool until it got lost or repurposed.

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]

zdas04,

Who proposed welding a #900 flange? Weld the test end on then cut it off and finish with whatever you're tieing in, often a weld end valve or another piece of pipe.

I accept there is a lot of "getting away with it" that goes on, but you seem to be assuming everyone is building their test ends out of sch160 pipe - "you'll find that your test head is around 30% of SMYS (probably 15% of actual yield stress). " Not always the case.

If that's the case then the pipe is good for a higher class so nothing lost there. Connecting flanges actually built for the pressure they are working to - seems like a good plan to me.

Anyway, I'll take your last para as some general level of agreement with me and BI. We all know flanges are pretty good for lots of pressure, so if the only "weak point" is the connecting flange to the system under test I'll take that.

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

 

[blackegl]

Little Inch,

The design code for the pipeline that the test head will be used on will vary due to the fact that they want to use it on various pipelines that they work on. It just so happens that the first run that they want to use this on is 1.5 x MWP.

I as well have seen some shady things done, however I don't what to be the engineer that is responsible for some type of accident. It seems as though the way to proceed is to use the 900# flanges on the test head and then create a 900# to 600# transition piece. I think it is better to error on the side of caution even though it seems as though the industry norm of these items is to have them built on site by rig welds with whatever material they have on hand.

Thanks

 

[zdas04]

Nothing prohibits welding a 900# flange on the pipe to be tested to make the connection, which could easily be blinded off, or cut out completely after testing.

I would make the connection to the test header via a welded connection which is then cut off and either welded to the next section or to a flange (golden weld).

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]

zdas04 - OK, I missed that from BI, but I think we're done on this one.

blackegl - I think you're going the right way. I'm always happier standing next to a test end with 2200psig on it knowing it has been tested to 2,800 or higher rather than exactly at the same pressure it is "working".

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

 

[BigInch]

zdas, I think it is common sense engineering. If you told any piping engineer to design a header that would see well over 1480psig every time it was in operation I'm pretty sure I can tell you what you would get. It doesn't matter what it is being used for, pipe design is pipe design. I'm sure you are aware that sometimes a 600# rated pipe actually blows up at even less than 1480, never mind at test pressure. Would you actually risk that happening on your testing equipment and possibly killing one of the test crew?

Myself, I suppose that I actually would use 2 x #600 flanges to make the connection and call both of them part of the pipeline, then blind off the one actually welded to the pipeline, tossing the other one in the salvage bin. But that's all. The rest all #900.

 

[dcasto]

Next are we going to toss out a paddle blind on a test because it saw its max pressure 1 time? Zdas is spot on! The life cycle of a a flange cycle is around a million times, now, the bolts and gasket? The bolts a few hundred thousand, the gasket a few times, toss them

Oh, does this mean all hydraulic systems need to be used a few times then dumped?

 

[BigInch]

Not if they are properly designed.