Pneumatic Test 1

[mcm1209]

Guys
I am working for a pipeline construction company.
I have been in the process and pipeline services since 1999.

I joined this company to start a hydrotest division.
We are being asked by our customer to conduct a pneumatic test of 7 miles of 20" pipeline.
Test pressure is somewhere in the neighborhood of 1300 psi.

I am very opposed to this but my company wants to push forward. The customer has given us the green light.

Oh yeah
We are doing this test within the week.

I need hard facts to stop my company from doing this. I have been looking for info on the net but have not been able to find something concrete.Or facts so i feel better about this.

I did find
"437.4.3 Only allowed for piping systems operated at 20% or less of SMYS"

Need help

 

[MJCronin]

When, exactly, should I be watching CNN for "BREAKING NEWS" ??????

 

[dcasto]

Come on, The pipeline is designed to operate at less than 20% stress, the test will be 22%. That's a piece of cake and not dangerous.

 

[BigInch]

Turn on CNN 15 years from now when corrosion finally completes its job.
Latest Pipeline Incident data shows
Most common cause of an incident: Corrosion
Highest Loss Potential: Third Party Damage from excavation
The Surprise Dark Horse: Failure from Instrument Malfunction, failure to operate or operates incorrectly
Of all Valves & Hardware caused incidents: Relief Valves Failures are the most common

The last two are especially worrying from a safety in design perspective.
Overpressure (Operating pressure > MAOP at the time of the incident) hardly appears on the list at all.

Independent events are seldomly independent.

 

[chicopee]

I like to know the outcome of the pneumatic test.

 

[MJCronin]

Some of my favorites.....

http://www.tedpelling.com/news/Pneumatic Shanghai LNG Terminal.pdf

http://webwormcpt.blogspot.com/2009/07/pneumatic-test-failure-in-mississippi.html

http://www.stepchangeinsafety.net/incidentsdiscussions/incidents/incident.cfm/incidentid/188

http://www.safteng.net/index.php?option=com_content&view=article&id=2037&Itemid=4

Some good pressure testing information:

http://www.hse.gov.uk/pubns/gs4.pdf

 

[zdas04]

MJCronin,
I keep seeing those same incidents. Fatalities are a big thing and need to be studied and understood. It bothers me that so many people use these examples to indict pneumatic testing.

The first link was REALLY short on details, but I've only seen brittle failure of pipe at seriously depressed temperatures. I've seen several incidents over the years where nitrogen was introduced to CS piping at temperatures significantly below 0F. That doesn't often end well, and the failure is always described as "brittle fracture" like in the link. Most of these that I see I assign a root cause of either bad procedures, no procedures, or not following procedures.

The second link requires a subscription that I don't have. I can't comment on those incidents.

The third link above is titled "Fatal accident during pressure testing". A PSV on the nitrogen supply was not properly captured. The valve lifted and the tailpipe pinwheeled. I've seen the same thing happen on pressure vessels and compressors. Most of the time it just ends up breaking the elbow off. Attributing this fatality to "pressure testing" feels quite disingenuous.

Thank you for the fourth one. I use one of those pictures in my Engineering Course. The details I had were really vague about how that much pressure built up in the tank. Now I know. A really crappy isolation plan (terrible Engineering) allowed testing of the pipe against some number of closed valves with inadequate protection of the downstream systems. Clearly not an indictment of pneumatic testing, but an example of an Engineer who shouldn't be in responsible charge of a can opener.

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.
The plural of anecdote is not "data"

 

[LittleInch]

I think the key point from those incidents is that the failure doesn't have to be the pipe being tested, but other associated equipment which because of the stored energy leads to large movement of pipes (and vessels!) or items being ejected like a bullet or artillery shell.

Of course procedures, isolation, weld testing, temperature control can prevent such incidents, but you wouldn't have seen the level of destruction in these pictures. Not sure about the brittle failure in the Chinese weld, looked too neat for that and more like weld failure, but I'm no metallurgist.

My motto: Learn something new every day

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

 

[BigInch]

If only they'd been making methyl isocyanate gas using that not so dangerous process, they wouldn't have killed 5000 and injured 500,000 people.

When there is a less dangerous way to do something, that is the best way.

If only I'd been hydrotesting at the time...
Do you really want to say that every night before you go to sleep.

Independent events are seldomly independent.

 

[zdas04]

I live in the Rocky Mountains. More lines have failed from elevation changes (test point at the top, pressure at the bottom of the hill twice test pressure) than have ever failed from pneumatic tests. It is all about responsible risk mitigation.

There are many systems on the flat that were designed for dry gas that have no earthly way to get all the test water out after the test. We never are able to kill all the microbes in any test water that we use. Ever. So that bit of water that you just can't get out grows into a septic environment that can have accelerated pitting and rapid failure. Do that in a Cyanide plant an you can go to bed every night saying "if I had only refrained from introducing that water into the system because of my irrational fear of pneumatic tests ...".

Good Engineering is all about managing the risks you cannot avoid. To demonize pneumatic tests is the exact opposite of that goal.

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.
The plural of anecdote is not "data"

 

[BigInch]

Actually it's about avoiding the risks that you can.

Independent events are seldomly independent.

 

[LittleInch]

Zdas04,

With the greatest of respect, failure during a test due to lack of consideration of hydrostatic head is simply truly terrible engineering.

The real issue, IMO, is that not enough people give real thought and discussion about the potential impact of a failure of a system under test. There is a thought process which just regards testing as a necessary evil to get the certification, not what is safest or the best in the long term.

Corrosion due to microbial action in the test water can be very serious, especially when significant quantities are left in hot locations. Removal of test water is again something often overlooked and can lead to problems with drying the pipe and preventing corrosion.

As BI says, it's all about balancing and controlling risks as to what the"best" solution is in any one place and set of circumstances.

My motto: Learn something new every day

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

 

[zdas04]

Risk avoidance is a means of risk mitigation. But if we avoid all risks then we just sit in a basement and quiver. There are other means of minimizing the mix of risks that always exist.

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.
The plural of anecdote is not "data"

 

[BigInch]

The real problem is that risks are not evaluated properly.
Many parameters are not understood and wide generalizations are drawn. Many variables simply are not considered and many others are considered independent when they are not. If one valve fails, do you really think that the valve next to it has the same failure probability. The facts are that it is almost always assumed to be equal to some standard number for that type of component, however if one valve fails it could just as easily be a symptom of improper maintenance that is applicable to everything in the whole plant, dramatically increasing the probability of failure of each and every item. In other words, the events are not independent and there is a much higher probability of cascade events. This is evident in just about every major incident. The probability of the several events that come together to produce an "incident" if evaluated independently, would give you a zillion years of trouble free operation, yet people get killed and damage is done.

When you do an air test, have you considered that marijuana has recently been legalized in Colorado... for example. Texting while walking has dramatically increased the risk of accidents of pedestrians crossing the street. The numbers in the book are not correct, are not evaluated correctly and do not consider the ever changing environments. Calculate the probability of a nuclear plant problem, and a tsunami that floods the pumps, because somebody thought that the probability of a tsunami big enough to wash over the dike and shut down the cooling pumps was too small. It would never happen in a zillion years, but it did. Please explain how that was possible. Do HIPPS systems have the same probability of failure in Iraq as they have in the North Sea? Probably not because there is no concept of maintenance in Iraq. Complex systems, by definition are hard to understand, so therefore very difficult to analyze at all. Some risks are better left avoided when there are other means available to accomplish the same objective at less of a risk, even they might cost more. Theoretically you could figure the costs x risk balance of losses and make a mathematically justifiable decision, but nobody has any real numbers to do so. They usually work in the imaginary matrix of probabilities, most drawn from the perfect world of NASA or nuclear plant data, and of independent events. The present method of risk evaluation is complete nonsense.



Independent events are seldomly independent.

 

[zdas04]

There are people who do a poor job of risk assessment and that is a bad thing. There are people who do a competent job of risk assessment and that is good thing. To say "risks are not evaluated properly" is painting with a pretty broad brush. I do a lot of pneumatic tests and feel that I understand the risks, understand the alternatives and their risks, and believe that the procedures I put in place and personally supervise limit those risks to appropriate levels.

My big problem is that the growing hysteria in the Engineering community against pneumatic tests makes it harder every year to get a test protocol approved. Last year I had a job with a 2200 ft elevation change (952 psig pressure difference between the top and the bottom when full of water) over 18 miles. We wanted to do a 900 psig test. The alternatives were (1) put the gauge on the bottom and accept that the top 120 ft would be "tested" at atmospheric pressure; (2) put the gauge at the top and accept that we were testing the bottom to 130% of SMYS; or (3) design a pneumatic test. I couldn't find a fourth alternative. Making up drag sections and testing in 60 segments left us with too many untested welds and too much pipe that was moved after test (really high risk). There wasn't an accessible location in the middle of the line to isolate and do two (or more) tests--this was really rugged country. I presented my evaluation and my client said "Yes, but a pneumatic test is just too dangerous". The contractor supported my plan and said that it addressed all of the potential for failure and he would much rather use my plan that try to do a hydrostatic test. I finally had to get the company's Engineering Manager to override the local Engineer. That created a lot of hard feelings and that company hasn't called me back, but we did the right test, safely, and with appropriate risk avoidance/mitigation. It chaps my butt that I lose a good client for trying to do the safest possible evolution. The reason for the problems was NOT that pneumatic tests cannot be done safely, it was the hysteria that is being created by comments like this thread is full of.

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.
The plural of anecdote is not "data"

 

[BigInch]

It's that clients have finally realized that it's harder to do business than before with the vast liabilities they face on a day to day basis. They have realized, like the company I work for now, that one wrong move can finish a $6B company and, fortunately, they prefer not to take that risk. If you have no choice, then everybody has to do it that way, or not do it at all. If there's a choice, then you have to choose the safest way, or .. not do it at all.

We just moved 500 people to Abu Dhabi, because it's too dangerous to have them on the road for an hour&half each way trying to make the commute from Sharjah.

The 1400 miles of BTC pipeline pipe in Turkey was transported around the country at a maximum of 50 KilometersPH 30 mph!

Our clients not only want to do it the safest way, they INSIST ON IT.
If the client want's to pay for it, what problem do you have, more work???
Our biggest problem is our guys that think they know better.

Independent events are seldomly independent.

 

[dcasto]

The topic is pneumatic test at low stress levels. Keep on topic y'all

 

[BigInch]

I don't think low stress level test has been established. There is no point using that pipe for a such a low op pressure.

We are being asked by our customer to conduct a pneumatic test of 7 miles of 20" pipeline.
Test pressure is somewhere in the neighborhood of 1300 psi.
Are there any people in the potential "blast" zone? Yes, line is close to residences
Do you cross any public roads? Yes, at least 10 road crossings

Wall thickness is 0.375
SMYS is 65,000 (X65)
Diameter is 20"
If DF is 0.5 (residences nearby), then Pmax = 1218
Just a test pressure factor of 1.1 would be a 1340 psig test pressure.
I don't believe for a second that this pipeline will run at 20%.



Independent events are seldomly independent.

 

[dcasto]

Then why does he recant on pressures and say the test pressure is 535

 

[BigInch]

Answer a question with a question:
Why isn't the pipe designed for 535?
Or maybe he's confused.
At this point I have more confidence in my calculations.

Independent events are seldomly independent.

 

[BigInch]

There was one report of at least one death occurring during a hydrotest involving connecting hoses being disconnected with trapped air inside. I'd not forget to mention that hydrotesting can be a dangerous enough operation on its own.

I'm looking for that report now. If I find it I will post the link.

Independent events are seldomly independent.