Hydrotest failure

[jtseng123]

Here is something from other company for lesson to be learned:
Quote:
"Every now and then you get reminded why you hydrotest and why you try to minimize how often you do NDE in lieu of hydro.
The fabrication, consisting of 24" sch. 30 (0.562 wall) pipe & fittings, was in the process of being hydro-tested when the failure occurred. There was approximately 80 linear feet of piping included in the test. The piping was to be pressurized to
2160 psi. When the test pressure reached 1740 psi, there was a loud noise, similar to a gunshot.
4 workers were impacted by the water pressure release. Two were thrown under a truck – One received a bad contusion on his head when he hit the under-carriage of the truck, and the other worker received a severe laceration to his cheek. A third worker lost several teeth when he was hit in the mouth by test equipment which was propelled by the rush of water. Luckily, the fourth worker was not injured.
This is a very sobering situation considering the number of hydro tests we administer during Construction. It is very easy to get lulled into a false sense of security after looking at hundreds of hydro & pneumatic tests. It is easy to begin thinking, “I've looked at a thousand tests - nothing's going to happen..." It is easy to become complacent with hazards you have in your job every day." unquote.

See attached photos:

 

[chicopee]

Was that elbow made in China?

 

[C2it]

The pressure seems high for the thickness .... Which Code and what material was used ?

Based on 2160 psi test pressure, assume the design pressure is 1440 psi ....

Seems to me that if this were designed to B31.3 with no corrosion allowance or tolerance deducted and an allowable stress of 20ksi, the minimum wll would be 0.84 in. or 50% greater than stated above.

Maybe it was much stronger material ?

 

[moltenmetal]

It sounds like the people testing this had trapped a lot of air, otherwise there should not have been sufficient energy in 80 ft of 24" pipe at 1750 psig for test equipment (and workers) to be propelled by the jet of water. Even in a system which is perfectly free of air, there is stored energy in the pressurized media and in the pressurized equipment, so a reasonable exclusion zone around the equipment is sensible- but something sounds very wrong in this case.

 

[stanweld]

Having routinely hydrostatically, burst tested pipe, there is enough energy released to cause the injuries described, depending on the location of the test personnel relative to the burst location.

 

[SNORGY]

This hydrostatic test failed as well, apparently at fairly low pressure.

One really shouldn't be conducting hydrostatic tests unless one knows what one is doing.

 

[JStephen]

One moral is, you should videotape your hydrotests, rather than photographing the results afterward.

 

[MJCronin]

I agree with C2it....... the test pressure seems way too high for 24"NPS/0.562" wall piping of almost any commmon material....

Is there a mistake somehow with these numbers reported ?

A quick review of the allowable stress tables of B31.3 shows no material (to me) that would be acceptable for these conditions.

Please tell us the material of the piping system and the Piping Code of Record for the project

I believe that there also was a material failure on the Chinese elbow intrados....but this is co-incidental.

Anyone ?????

 

[Engdoitbetter]

An interesting and dismal post, indeed.

With regard to released energy, I performed a short calculation with Bernoulli using SI units:

speed = sqrt (2*1740/14.5*100000/1000) = 155 m/s

which seems to me fast enough to throw a man.

Regards,

Stefano

 

[1gibson]

Is it common practice to hydro without pipe supports? Can't help but think at first glance that some supports could have limited the pipe straightening at the elbow.

 

[moltenmetal]

This one has apparently been circulating for a while.

http://www.safteng.net/index.php?option=com_content&view=article&id=1119&Itemid=178

We're not the first people to be suspicious that something is wrong with this report- see the comments below the article.

The speed of the water is less relevant than the volume of water that must be lost in order to dissipate the energy stored in the elastic stretching of the pipe and the elastic compressibility of the water. In the absence of substantial trapped air, this volume is very small. By my rough calc, the amount of water necessary to take the steel all the way to the onset of yield would be about a cubic foot, with perhaps another whole cubic foot or two for the compressibility of the water itself. Propelled out of a crack, it will not be coming out in a nice coherent jet- it will definitely spray. The likelihood that this small amount of water hits several people directly or via "thrown testing equipment" with sufficient force to throw the people any distance would seem very low indeed- UNLESS there were a LOT of trapped air to propel a vastly larger amount of water.

 

[stanweld]

The piping in question appears to be a component to be installed in an underground pipeline and would be tested at 90% of SMYS per B31.8 (possibly). Pipe could be X60, X65 or X70. The wall thickness would seem more correct with those assumptions. Over forty years ago (before OSHA), we metallurgists used to enjoy dousing Ross carrier drivers in the steel mill when busrst testing 12" to 16" ERW X52 pipe. While there was not much air entrained in the pipe hydrotest water, the energy released was enough to lift the explosion roof and lift the water over 20 feet above. If an adult male were sitting on that roof, he would have been thrown and injured.

 

[weld7777]

We had 12X10 TEE, SA 234, blow a crack @ 1100psi, in the area of bottom of branch run, Nitrogen test, we were very lucky no one was hurt, nor was equipment damaged. The TEE was from Malaysia, so China was not the only one that will have this happen. I believe that is why NDE cannot replace pressure tests.