Hydrotest vs. Testing with Air 10

[clooney]

Dear gents,
I have not come across any work instructions or specs in our office establishing a guideline on "deciding when one should consider testing a piping system or pipe line by Hydrotest vs. Pneumatic Air Test?" or any other method for that matter. Can anyone direct me in finding the answer? I have been checking with the more experienced people in our office but, despite all their past achievements, none of the seniors were able to give me a straight answer or at least guide me in my search . Thanking you in advance.

Farzad

 

[mgp]

I agree with moltenmetal and others that a hydrotest should always be preferred unless there are good reasons not to do so - there may well be.

In 1995 there was an incident in a british construction yard where three workers were killed during a nitrogen pressure test. Here's the story as I have heard it.

The project was an offshore topside gas facility.
The pipe to be pressure tested was an 18" gas line - in order avoid wetted piping, reasons were good to use nitrogen for pressure testing instead of performing a hydrotest. For bolted joints, Grayloc type clamp fittings were used.

Somewhere on that line a valve or something had been temporarily removed during the test and the line was blinded off with a blind hub and a Grayloc clamp. A couple of meters upstream of that line there was a check valve and further upstream was the pressure indicator used for the hydrotest. The check valve was not removed as nitrogen was being filled from the upstream side, however with the line being blinded off downstream there was a considerable volume trapped between the check valve and the blind hub.

The pressure test was performed, the system was depressurized using the PI, but because the PI was located upstream the check valve the trapped volume remained pressurized to some 180 barg (2610psig) without anybody noticing it.

The workers went back to reassemble the removed valve. First unbolting the clamp - it stuck so nothing happened.
The poor guy then hit the clamp with a hammer to it loosen it and that was the end for him and to of his colleagues.
The nearest guy virtually dissapeared as he was blown to pieces.

It must be said that my information is second hand as I have it from several colleagues or ex colleagues who at the time worked for the company that designed the module. All I could find during a google search was this:

http://www.wwne.demon.co.uk/ww2000/d00-4.htm#letter

One can argue that this incident was due to human error, but it surely demonstrates that pneumatic pressure testing should be done only with extremely high safety precautions.
Safety cost money, so if there are not very good reasons for using air or nitrogen - use liquid.

Whenever we do pneumatic testing - e.g. for instrument air, nitrogen (much smaller pipes and lower pressure) - we have to evacuate the whole area etc etc. during the test. This is a legal requirement.

Just my comment

Regards
Mogens









.

 

[SMF1964]

Hydrotest vs. pneumatic test: Let's start on the reasons for performing the pressure test to begin with - to assure that there are no leaks.

If you do this via pneumatic test and you hear 'hissing', that tells you only that the system as a whole has a leak somewhere. You must then go look for the leak.

If you do this via hydrotest, then your leak will be readily visible because it's got a wet spot.

As long as we're talking horror stories: had a station air compressor fail at a flexible hose (2" diameter). The hose burst like a balloon, sending at least three pieces of the steel flying, along with some wire filaments from the braiding. One of the steel pieces caught the mechanic on the jawline, missing his jugular/carotid artery by less than 1/2" (1 cm, to those of you metric) but severing the major nerves in his face so that he cannot smile on one side of his face.

If you want to see the consequences of pneumatic vs. hydrostatic related failures first hand, pop a balloon with a pin and pop a water ballon with a pin. The difference will be readily apparent.

 

[zdas04]

I've got a whole book full of piping failures in gas service. I don't think that using gas as a test medium is a significant portion of the list of failures. That is because in a pneumatic test there are a list of things that must be checked prior-to and during the test. Had the workers in the example above followed those procedures (and if the engineer designing the test had called out the proper procedures) they wouldn't have been killed that day.

Gas is more concentrated source of energy than liquid. Any time you use it under pressure you need to use appropriate care.

David

 

[dwedel]

zdas04,

I agree with you on the safety involved with pressure testing. Hydrotests can fail spectacularly.

Can you share any of your stories? I'm affraid I can't share my hydrotest story.

dwedel

 

[Scipio]

One of the ways I look at it, if every pipe spool or pressure vessel was designed properly, fabricated per design properly, the original materials were cast or forged properly, and the joints made up properly, so there was no chance of a failure, there really wouldn't be a need of a pressure test. Since that's not the case, pressure tests are performed because occasionally, something's screwed up enough that the spool or vessel will fail pressure testing. If I have the option, I'd rather it fail during a hydrotest than pneumatic test.

Now, there are a number of reasons to justify pneumatic testing, like zdas04's elevation change (I'd do a pneumatic there, myself), but my default is a hydrotest. In fact, in my region (Alberta, Canada) the local regulators require hydrotests be performed unless there's adequate justification for alternate testing, and then they want to see detailed testing procedures to cover the checklists zdas04 mentions, which has to be prepared and stamped by an engineer prior to testing. There's an interesting article on the danger of pneumatic testing on page two of this pdf link:

http://www.absa.ca/newsletter/v07-2002/v7-iss2.pdf

The really interesting part is the equivalent energy table, which equates a 2 ft x 6 ft vessel of air at 500 psi to a pound of nitroglycerin.

 

[zdas04]

Sicipio,
That case study has been kicking around a while and it is pretty scary. While the arithmetic of energy storage is accurate, a vessel does not have the ability to deliver the energy as effeciently as a pound of nitro.

I once calculated the stored energy in a 20 mile, 12-inch static test to 700 psig with air. The numbers were in the kiloton range. The problem with the calculation is that by the time the gas even 100 ft from a failure sees the hole, it is a flow calculation instead of an explosive decompression calculation. I was unable to establish exactly what volume of gas would participate in the decompression, but I was able to determine that it is a fairly small subset of the total gas in the test.

David

 

[dwedel]

zdas,

I deal mainly with maintenance on the compressors, so I am not exposed to the pipe testing as much, however, I do need to have knowledge of the testing.

I was at a welding class, and they were talking about the hydrogen crack in a pipe. If the longtitude crack in a pipe does occur, they said that the crack would propogate through the steel faster than the speed of sound in air. Thus the pressure in the pipe would not relieve itself and the crack would propogate through a hole length of pipe. Is this crack propogation faster than the speed of sound in water as well?

to everyone,

I guess, that I agree, that pressure testing with water is preferable to air. However, many of the posts here are emphasizing the explosive power of air, and it makes me feel that people are getting the idea that testing with water is "safe". I would emphasize that this is not the case, that a pressure test is dangerous no matter what the medium. All pressure tests should be treated with caution and careful adherance to protocol.

 

[moltenmetal]

You're right that a hydrotest is not "safe", it is "safer". That said, a properly-performed hydrotest is not just safer, it is ENORMOUSLY safer than a pneumatic test at equivalent pressure.

Crack propagation speed in metal is irrelevant to the safety difference between the tests. The volumetric expansivity of the test medium IS. A liquid's pressure decreases enormously with decrease in volume (i.e. it is nearly incompressible), and hence the amount of energy stored in a cylinder of compressed (subcooled) liquid is minor compared to the amount stored in a similar volume of gas at the same pressure.

Hydrotests are safer than pneumatic tests because even the sudden generation of a crack in a vessel or pipe will merely generate a leak or spillage rather than schrapnel and a devastating shock wave. Even the small deformations necessary to develop the crack will reduce the pressure in the vessel enormously if it is entirely filled with water.

A well-designed and executed hydrotest takes advantage of these factors. By carefully eliminating the trapped air in all components prior to carrying out the test, a hydrotest can become a very sensitive leakage test as well as a very safe test.

 

[bigskyguy]

I run pressure tests weekly (some time daily) on a variety of differant systems, I'll tell ya from experiance, I run with pnumatic tests at a lower pressure before I move on to full pressure hydrostatic tests. For a typical 150-300 PSI system that requires a hydrostatic test of 1.5 working pressure, throw a 50 PSI air test on it first. Main reason being, if their is a leak, it sure is alot easier to fix if you only have to release air pressure instead of dumping out water. I've argured this with a number of inspectors, my solution, stick to the specification but take precautions and use other means of testing that is more user friendly to repairs. And always put the gauge at the lowest part of the system....

 

[mgp]

There was a thread about equivalent energy from a gas contained vessel which might be of interest:
thread794-26767

Regards
Mogens

 

[joeswoes]

My 2 cents as a chemical process equipment designer :

Looking for leaks ? low air pressure under water or with snoop detergent.

looking for safety/structural integrity ? 1-1/2X rated working pressure hydrostatic.

Rated working pressure pneumatic testing is extremely dangerous due to the compressability of gas vs. liquid.

 

[dwedel]

Moltenmetal,

We just had an accident with a hydrotest. Maybe there were some problems and some procedures weren't followed, I don't know as I was not there. But what information we are getting back, it sounded like it was pretty explosive. I cannot assume that it is enormously safer to perform a hydrotest verses air test.

Have a good day,

dwedel
Hotrod Big Engines!

 

[PAN]

Refer to ASME B31.3, a pressure relief device is necessary for pneumatic test. This should imply that pneumatic test is more dangerous than hydrostatic test.

In my company, the safety operation team will not allow us to perform pneumatic test if we can perform hydrostatic test.

 

[zdas04]

The best pictures I've seen of static test failures have all been vessels tested at unreasonable temperatures with water. Brittle failure of a 2-inch thick piece of steel is a really impressive energy release.

Maybe one of the reasons I'm such a vocal advocate for pneumatic tests is that I've seen too many people enter a static test with a binary mindset (i.e., "hydro good", "gas bad") and not specify minimum water or ambient temperature for a test, not specify liquid disposal technique, not specify an air-elimination period, etc. I once reviewed a hydrostatic test "procedure" where the "engineer" hadn't calculated hoop stress that the test would impose - when I calculated it at 175% of SMYS (at the bottom of a hill) he was a bit shaken and we re-wrote his document to make it an air test at under 30% of SMYS (same bulk pressure, no hydrostatic head). This guy was actually a pretty good engineer, but was suffering under a silly set of pre-conceived ideas that had been shoved down his throat by people who think "hydro good" and he thought the test was the easy part of his design.

Everyone is at least a bit scared of gas tests and they tend to be better specified. Fear is a powerful motivator.

David

 

[moltenmetal]

dwedel:

I don't know the test's conditions so I can't comment. But I can tell you this: there's a reason B31.3, ASME VIII etc. carry out a hydrotest at 1.1x rather than 1.3x or 1.5x design, and that reason is the energy content of the compressed gas in a pneumatic test. You can test small devices to failure quite safely if you are scrupulous about eliminating all the trapped gas and other means of energy storage- in fact this is routinely done to establish pressure ratings without carrying out finite element analysis etc.

We will test systems of limited volume pneumatically at low pressure (15 psig) pneumatically to get the gross leaks out prior to a hydrotest. If it's of really limited volume, we may extend that to full shop air pressure (100 psig), but by no means would I recommend this as a general rule.

David: we agree that thinking, NOT rules of thumb, are a prerequisite to safe testing. The assumption that hydrotests are intrinsically safe, regardless of design, can be FATAL. It's also possible to design and carry out a pneumatic test safely- it's done every day. Any test which ignores sources of pressure other than that supplied by the test pump or gas cylinder, or ignores secondary effects of the testing program, OR ignores the stored energy in the test medium, may be safe, but only by accident- regardless whether it's a pneumatic or hydraulic test.

But again, stored energy is the principal hazard in MOST tested systems- and most people are not testing long, thin pipelines where the energy release on failure is limited by pressure drop in the line. The energy release from the failure of a 5-gallon air receiver during a pneumatic test, or by accident due to corrosion, can be spectacular and can lead to injury or death and serious property damage. The same receiver tested to failure by a properly-designed hydrostatic test WON'T result in injury or death- in fact it may not even get anyone wet.

 

[dwedel]

Thanks for the reply Moltenmetal,

Your points make more sense in light of the relative size of the test. I had not considered that.

dwedel
Hotrod Big Engines!

 

[acoilfld]

Generly we hydro test when ever possible - for safety reasons and also because in B.C. an above ground hydro test is 1 hour, 4 hours for below ground, 24 hours for a air test.
I only air test fuel gas lines and engine oil lines. On a small air test we fill the line with compressed air and "bump" it up to test pressure with nitrogen.
Working with a pipe full of compressed gas is like handeking a pipe bomb - only larger.

A.C.