pneumatic test 6

A local company making flexible connections for the automotive and aircraft industries, did their pneumatic tests under water. In your case and not knowing anymore about the conditions imposed on your testing procedure, you could use an antifreeze solution, light weight oil or other low freezing point solution to do the pneumatic tests.

What the heck is the problem? I've done pneumatic tests with air, nitrogen, or methane to 62 barg. All were code compliant and done very safely (without submerging the lines).

If you have attached a 3.5 barg design pressure valve to your 31 barg system, then the system is now 3.5 barg design pressure. You can't test a low pressure valve to a medium pressure with gas or with water. Something has me quite confused here.

David Simpson, PE
MuleShoe Engineering

Law is the common force organized to act as an obstacle of injustice Frédéric Bastiat

I think that he meant 35 barg instead of 3.5 barg.

thank you all for replies and sorry for some spelling mistakes above since i was in hurry.
chicopee: i do not see any prblem in testing itself but the pressure will be applied on the continuation line is too high compairing with DP for the line...
let me simplifiy the issue, i have line with DP 31 barg welded to a glob valve with DP 31 barg and the a contiuation line welded to the valve has a DP of 3.5 barg.
i tried to split the two lines into two test packages to test them with the valve in close position but for some reason(i believe he does not trust the valve)the client insist to test them as one package with the valve in open position..the material for both lines are the same(only DP is different),can we test them toghether under 34.1 barg????????

thanks again

it was zdas04 not chicopee

raptor,

In this case the client is correct. Testing against a closed valve is not normally allowed, but you might be able to make a case so long as the deign pressure of the valve (and it's seats and seals) is more than your 34.1 barg. However it's not a good way to work and shows a lack of forethought about how you were going to test this pipe. What you need to do if testing against the valve is not allowed, is cut off the 3.5 bar rated pipe d/s of your valve, weld a cap on, open the valve and then test at 34.1. Do the same thing for the 3.5 barg pipe then weld the two back together as a golden weld.

Next time start planning the hydrotest BEFORE you weld the pipes together...

For a pneumatic test, just make sure everything is well tied down and have an exclusion zone of 10m to 15 m during the test. DO NOT do any leak testing with soap at this pressure.

My motto: Learn something new every day

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

An alternative to LittleInch's reply would be to weld an ANSI 300 flange downstream of the globe valve and put in an insert blind for the two tests.

I would be pretty nervous about a spec break from ANSI 300 to 3.1 barg with only a globe valve between them. The globe valve will only give you any dP at all while flowing, when you stop flow the pressure across the partially open valve equalizes very quickly.

David Simpson, PE
MuleShoe Engineering

Law is the common force organized to act as an obstacle of injustice Frédéric Bastiat

That's another option - basically any form of positive physical isolation is required for the test

I didn't go into the design aspects - too many unknowns.....

My motto: Learn something new every day

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

". . .exclusion zone of 10m to 15 m during the test"

Try about 100 to 200 meters. In a catastrophic failure, large pieces may be thrown even farther than that. At 200m, I've got time to step out of the path of something comming my way. Yes, the risk of a serious failure is low, but the consequences are massive. And yes, I have pneumatically tested big-bore pipe at 35 bar/550 psi, and small-bore at 2300 psi. And I was worried.

FYI, at these pressures any pinhole leaks [like weld porosity] will be screaming. After reducing the full test pressure at least 10% and leaving someone to watch the guage to make certain that the cold gas you are testing with doesn't warm up and increase the pressure, then walk the line and listen for leaks. NEVER walk a line at full hydro pressure - unless you let me take out a life insurance policy on you.

Duwe6,
I've got a few points about your post
[ul]
[li]Explosive decomprssion. A few years ago a summer intern at NASA did a calculation of the amount of energy potential in a multi-mile, big inch pipeline at very high pressure under a test compared to atmospheric pressure. His conclusion (published on NASA letterhead) was that in an explosive decompression the entire energy (MT of TNT-equivalent range) would be explosively vented. His conclusion was utter nonsense. Communication within a gas system is limited to the speed of sound, so the initial decompresison is limited to about one joint of pipe. The explosive decompression is ounces of TNT-equivalent, not mega-tonnes. When you look at the case studies that people keep trotting out for this discussion, the big-time failures (e.g., the top half of the vessel upside down on an upper floor of a structure) are all consistent with limited accesses to energy (the vessel that everyone points to was not part of the test, they were testing against shut valves and the pressure built up in the vessel until the stress in the vessel walls caused it to fail, you never would have tested to failure on purpose). Exclusion of 10-15 m is actually excessive. If you did get a catastrophic failure with shrapnel the bits of steel would be traveling at near-sonic velocity (call it 450 m/s) and you really think you could "get out of the way" in 200 mS? You've been watching too much TV. [/li]
[li]Pinhole leaks. Flow is a function of dP and flow path. A pinhole leak in a pneumatic test has a big dP, but a very small cross-sectional area and relatively long flow path. I've seen weld porosity leaks that barely spit, certainly didn't sound like sonic flow, because it wasn't.[/li]
[li]Full hydrotest pressure. I always walk the line (at least part of it, I generally have other people walk the rest of it because I don't walk very fast) at full hydrotest pressure. It hasn't killed me a single time. Upwards of 100 tests. Keeps not killing me. Guess I won't be buying the insurance policy in your name.[/li]
[li]Pressure increase with temperature. You're kidding, right? If I start a test at 80°F and the sun shining on the pipe heats it to 90°F (almost impossible in a buried line, barely possible in an above-ground line), that is about a 1.8% increase in pressure (900 psig to 917psig) . On the other hand if I have a line full of water at 900 psig test and it heats from 80°F to 81°F then I've gone to 1000 psig, now that can be dangerous.[/li]
[/ul]

I live in the Rocky Mountains. Every test I do has more elevation variation than I can handle with a hydrostatic test. I often have major battles with clients over the attitudes that you are expressing in your post. They don't have any basis in fact, it is just fear and superstition. When my clients buy into that fear and superstition we are left with zero options. No way to segment a test to take the elevation differences out. No way to test prior to installation (nowhere with several miles of level ground and it is really hard to get a drag section to match field bends). Fear and superstition prevents a pneumatic test. Usually they (I won't sanction this procedure and I leave) end up with the test gauge at the bottom and never testing the top at all, not even a leak test.

Everything we do as Engineers has some amount of risk. I write test procedures that (so far) have a 100% track record of never hurting anyone and never spilling contaminated fluids (and I have had several tests fail without any projectiles). I don't want an unsupervised new hire to write a test procedure. But I've gotten good results from supervised new-hires writing good procedures. The problem comes in when we (as Engineers) can't pull ourselves away from meetings long enough to do the Engineering, so we rely on standard prohibitions against pneumatic tests, threaded connections, etc. "Standard prohibitions" should be an anathema to us all.

David Simpson, PE
MuleShoe Engineering

Law is the common force organized to act as an obstacle of injustice Frédéric Bastiat

Fair enough, I keep thinking of buried pipe. Most piping facilities have a pressure test area wish is suitably protected and this is 1" pipe, but the point is well made.

My motto: Learn something new every day

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

Dave, this is a pneumatic test. Temperature rise is enormous. Where did the 'pneumatic gas' come from? Comes out of 2300 - 3000 psi portable cylinders. Then the pressure is decreased by putting it into the pipe - causes a major drop in temperature. And that self-refrigerated gas also has to pass thru the high-pressure regulator; regulator gets cold and its diaphram gets stiff. Now you turn your back on that regulator as the flow slows down when the pressure in the pipe reaches the setpoint. Regulator rewarms, diaphram gets less stiff, more pressure is delivered. There is the first pressure 'overshoot'. Now, the autorefrigerated gas in the line warms up. Causes a major increase in pressure. These mechanisms are probably why B31.3 tells us to increase the pneumatic pressure in 10% increments - lets the temperature equilibrate, keeps from having unexpected pressure jumps.

Now step back and consider the whole system. It took the contents of one or more high-pressure nitrogen cylinders to bring this system up to pressure. Incorrectly handled gas cylinders are damn dangerous, due to the massive ammount of stored energy and the potential for a catastrophic release. All we did was move that stored energy from the cylinder(s) into an untried pipe system. That full ammount of energy still exists, and it wants out. One [very] bad weld, one flawed pipe seam [s/s sch40 is usually a seamed item - much cheaper] and something goes BOOM. The odds of this happeming are very low, but the consequences are high.

Anybody can walk an 8-inch wide beam 2-ft off the ground. Simple & safe. Now move that beam 100-ft up. Still the same probability of falling off, but the consequences of a failure just became unacceptable.

Hydro's are very safe - very little stored energy due to an incompressible fluid. Pneumatics are a different animal. Bet that you are old enough to remember split-rim truck tires. They killed and maimed a bunch of mechanics; with only an internal volume of a few ft^3 and a pressure of 90-120 psi.

And yes, a pinhole leak on a 550 psi pneumatic will 'scream' at you. Ask a boiler operator about the noise from a tiny high-pressure leak.

This same thread comes up all the time. However, I guess about all I would say is that in general one should probably best have a very good reason to substitute a pneumatic for a traditional hydrostatic test (I guess not only done with water, but sometimes maybe even other fluids if water is objectionable), and when one nevertheless does they best do your homework as to all the hazards and preventions involved. Trying to save a little money or bother in testing, or engineering of same, or for that matter in testing pipeline designs or profiles that could have been done a little differently to begin with that would allow for conventional hydrostatic filling and testing, may not in the long run be good enough reason if someone (heaven forbid) is injured in the testing. In this regard, the juxtaposition of the following two statements are striking (if I am reading same correctly):
“Exclusion of 10-15 m is actually excessive. If you did get a catastrophic failure with shrapnel the bits of steel would be traveling at near-sonic velocity (call it 450 m/s) and you really think you could "get out of the way" in 200 mS? You've been watching too much TV.”
If pieces of steel, or for that matter virtually anything else, happen to be propelled at a velocity of 450 m/s, I guess I have to ask how is an exclusion zone of only 10-15m “actually excessive”? I am aware that a 28-feet long chunk of the San Bruno pipeline (originally buried) was reported by many sources including found a hundred feet (>30m) from its original location (after the gas pressure in the line reported at had climbed to only “386 psi” shortly before). For whatever it is worth also I have also seen one poor unfortunate soul was per the report at and others reportedly/tragically killed within his dorm room “350 m away from the explosion” during pneumatic testing (I do suspect however this incident involved exposed as opposed to buried piping).

Believe it or not - "Get out of the way" is easy, at 200m. Guys at rock quarries do it daily. They 'duck & cover' for the initial blast due to possible supersonic small rocks. Then they pop up like a Major League catcher, scanning the sky for big chunks that got lobbed into the air. zdas04 got the velocity right, the big chunks will be slightly subsonic, or slower. If anything stays in your field of view, not drifting up or down, nor left or right -- it is going to land right where you are standing. Take a couple of steps to one side and don't stay where that big rock intends to land. Easy for an old, fat guy at 200m. Not enough reaction time at 20m.

I've never done a test from cylinders. I don't know why someone would do that. If the test is small enough for cylinders to make sense, the logistics of a hydrotest generally make more sense. I would be very reluctant to do a pneumatic test out of bottles. If drying the line was the big issue, I might use liquid propane for a test that small (then carefully control the rate of depressurization to minimize phase-change cooling).

Hydrotests are not nearly as benign as people seem to think. There was a case here a few years ago where a company was doing a hydrotest on a line that ran over an 1100 ft bluff. They put their test equipment at the top. Pressured the top toward 900 psig. The bottom got to somewhere above 1300 psig and a valve parted. Unfortunately it was pointed towards a road. More unfortunately a small vehicle was passing the site as the valve came off. The road was next to a river and the driver discovered that Volkswagens will definitely float, just not indefinitely. Could have been a catastrophe, no one was hurt so it just pretty funny. Hydrotest water is a hazardous waste that is really hard on aquatic life. Way too much of it has been dumped in ditches, dry washes, and rivers. Failed hydrotests usually just make a bit of a mess. Sometimes they put people at risk through flooding confined spaces. Sometimes they spray water into switchgears. Sometimes the part that comes loose is not captured and can be launched with enough force to do damage.

When I've done air tests I've used on-site air compressors and my problem was getting the gas cool enough that I didn't have to add much make-up gas during the soak periods.

With methane tests I've been able to purchase pipeline gas to run the test and the temperatures were basically ground temp.

For nitrogen tests I've used liquid nitrogen trucks. The way you get liquid nitrogen up to pressure is by heating it. On any inert gas test, the first warning in the procedure is that you have to monitor injection temperature and that it must be within a very narrow band around current ambient temperature. I have seen several accident investigations and seen several accident reports on failures associated with inert gas and every single one of them has had a root cause of "incompetent Engineering" or "Failure to follow procedure".

There simply is no way to take the risk associated with pressures approaching SMYS to zero. You can get them close enough. A person wearing fall protection can walk that 8-inch beam 100 ft in the air time after time without significant harm coming to them--procedures are the "fall protection" of static testing.

I used to work in steam plants and have seen the mechanical operators that went looking for a 600 psig superheated steam pin=hole leak with their hands. They're usually called "stubby". Superheated steam has a whole lot more specific enthalpy than ambient temperature air or nitrogen. Even at that the pin-hole leaks I've seen were noisy, but not sonic. The rule of thumb from control valves is that if the "hole" is smaller than 1/6 the flow path length you will not have choked flow (e.g. for a pin-hole leak the flow path is the wall thickness and a hole diameter would have to be greater than 1/16" in a 3/8" thick pipe to get choked flow). 1/16" hole is pretty damn big.

Yeah, those split-ring rims were fun. I saw a tractor tire take out a rafter in a shop once when I was in High School, The tech was filling the tire with antifreeze. Energy is energy. What matters is the duration of the release event.

David Simpson, PE
MuleShoe Engineering

Law is the common force organized to act as an obstacle of injustice Frédéric Bastiat

A 1", sch 40 ss line requires nearly 10,000 psi to rupture. And I've tested 1/2" to 4", sch 40 lines to rupture to so determine. 35 barg doesn't produce enough stress in hoop to to cause rupture even with a very long, deep crack. The pneumatic test is quite safe to perform. You will need to assure that your blinds are properly sized and attached. I have seen instances where the blind shot off. The volume also has not been calculated and the stored energy is a function of pressure times volume. If the volume is low, even the NASA calcs will permit closer to a 10 meter or less exclusion radius.

Zdas is absolutely correct regarding "explosive decompression" There are a number of petrochem companies that use a length of 100 ft at the test pressure to determine the exclusion radius using the NASA calcs. Having said that, one of our customers required a 150 ft exclusion radius for a 55 psig test; the exclusion radius per NASA calcs for the full length of the test was only 30 ft.

rconnor,
Come on. San Bruno? Really? San Bruno was a fuel-air explosion on a pretty grand scale. Not explosive decompresison. The line split, gas mixed with ambient air and found an ignition source. Yeah, a 1.5 ton piece was thrown a ways. So what? When the US dropped an atomic bomb on a couple of Japanese cities in WWII there were some pretty heavy pieces of steel thrown a ways also. Is that relevant? Yes, when you ignite an explosive gas (or liquid for that matter), it can explode. Who knew?

The Shanghai case is relevant, but that report is premature ("the cause of the failure is under investigation"). I've seen several of these kinds of failures and the cause is usually that they started the test with the pipe below freezing or introduced inert gas a cryogenic temperatures. That is a failure of Engineering (or following procedures) not a condemnation of pneumatic testing.

Of course there is significant energy available to cause harm during a test. That is why you have Engineers design static tests.

David Simpson, PE
MuleShoe Engineering

Law is the common force organized to act as an obstacle of injustice Frédéric Bastiat

"San Bruno was a fuel-air explosion"

That explains why a well-buried line made such big blast. FAE's can be HUGE.

The thing that needs to be considered here (and all pressure tests - pneumatic and hydro), is a proper analysis and consideration for the failure events possible and what can / needs to be done to mitigate against the effect. The failure rate in pressure tests is fairly low thankfully, but nowhere near low enough to be in the "negligible" region

Hence pneumatic needs to consider things like crater size for a large buried line, what connected flanges, spools, end caps, valves, tees etc could potentially "fly off" and in which direction, how secure the pipe is against movement in such an event, things like that.

Then you can demonstrate what you are doing is sufficiently low risk once you take the required actions (e.g. testing against a blast wall / barrier to prevent things hitting anyone if they fly of, restricting access to certain areas, having a suitable exclusion zone.)

Each system is different, so you need to look at each one separately, but with the same procedure thinking all the time - what if it fails here and how Do I prevent that failure injuring someone. Testing is designed to demonstrate strength or find weakness prior to operational service. Most of the time it does the former, but occasionally the latter.

When I was a student many years ago, we were testing plastic tubes repeatedly in a lab with water / nitrogen to use as rockets. One failed during a run scattering bits of plastic over a significant area (students were considered expendable in those days....). Only luck prevented any serious injury. After that they made us build a wooden frame very close to the tube to catch any subsequent explosion before the velocity was too high. A further failure saw the end shoot off and ricochet off an angle narrowly missing an observer. Both failures could have been prevented from creating injury potential by going through a proper failure analysis and review, but I was at the bottom of the learning curve in those days and practices like this not applied in university type situations.

So take note of all the posts above and individuals personal experiences and think about failures, do it properly and in all likelihood, your system will pass the test without incident or injury

My motto: Learn something new every day

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