I presume this question relates to an in-service component versus new construction? If so, typically, a pneumatic test for an in-service component is to check for leaks, nothing more. Therefore, I would limit the pneumatic test pressure to 25 psig to assure personnel safety.
ASME section VIII and B31.1 have guidelines for pnuematic tests in lieu of hydrostatic tests. See those documents for specific precautions and procedures. One might test to 1.15 times design pressure ( double check that, from memory) , but only after slowly buidling up pressure and checking leak down rate. Also, the metal and air temp must be above teh ductile brittle transition temp of all vessel materials and welds, and above the 70 F ASME minimum.
There are some services where it is not convenient to fill the vessel with water, such as underground H2 piping, yet is it essential to prove pressure integrity prior to fillign with a hazardous gas. In those cases the pnuematic test would be to the full 1.15 times design pressure.
If a hydrotest is not possible since the water is poisonous to the process, then consider another fluid, such as methanol which is not poisonous to the process. We've used this path several times.
I'd rather have a vessel fail under methanol pressurisation than under pneumatic pressurisation.
A pnuematic test stores much more energy and if the vessel fails, it will fail with explosive force.
One contractor I know had to do a pneumatic test of a very large vessel. They NDE'd the vessel to within an inch of its life and transported the vessel to the desert, set up significant instrumentation and remote operated equipment, and placed the vessel in a sunken area. They then set up a control room a kilometer or so away and controlled the test from the control room via remote operation and with vision provided by CCTV and binoculars from the control room.
Do NOT take pnuematic testing lightly. ESPECIALLY, if you have any QA/QC concerns regarding welding!!
For saftey reasons you need to look at local requirements for pressure systems. In some places any system over 6psi requires special saftey concerns. Other places use 15 or 25 psi as limits.
We test tubing at 150psi, 250psi and 450psi with air, but trust me there is a lot of saftey equipment involved.
= = = = = = = = = = = = = = = = = = = =
Corrosion never sleeps, but it can be managed.
The current edition of ASME VIII, Div 1 specifies a standard hydrotest as 1.3 times the MAWP stamped on the vessel nameplate. Alternatively a pnuematic test may be performed at 1.1 times MAWP. Both tests pressures are also modified for the difference in allowable stresses at test and design temperatures.
Your answer directly answers the original question from the PV code point of view, but is a pnuematic test safe?
Assuming I haven't stuffed up the calc, a garden variety 1.2m diam by 3m long pressure vessel, tested to 1500kPag has over 7.4MJ of stored energy. That's the same stored energy as about 1.7 kg of TNT.
[The formula is in Australia's inservice inspection standard AS3788 appendix D.]
[1 ton of TNT = 4.184MJ ref: www.answers.com/topic/megaton]
If you had a 1.7kg of TNT with a lit fuse, from what distance would you observe it??
IMO pneumatic testing requires careful thought, precautions and where possible, avoidance.
Is it safe? Of course not,it is extremely dangerous. Why do you test? To disclose major defects,yes?
Can it be made safer,ie:an acceptable risk?
I believe so with containments and isolation. I witness large vessels at air test pressures above 500psig. But they are in blast rooms,enclosed tanks,and/or isolated in all cases.
There are some good studies about this at the National Board and NASA. Do a search of this website.
The concern related to pnuematic test has to be placed in the context of how safe is the equipment when pressureized with the regular process gas at design pressure.
Pressurizing a newly built tank with air to 1.1-1.15 design pressure while special safety precautions are implemented is no worse than the case where, for example, the tank is pressurized to 1.1 times design pressure during a safety valve lift event while using the normal process gas. In those latter cases, there are no special safety precautions being observed, many operating technicians are gleefully ignorant of any danger, and the process gas could have other safety issues ( ie hydrogen, hydrogen sulphide, flourides, etc.)
If you are replacing an existing component, or repairing existing piping you may be able to get away with an inservice test with AI approval, that is system pressure. There are also other NDE options in lieu of hydrotest.
If you are procuring a new component have it tested off-site then at site you just need to test tie-ins, it is usually acceptable to AI to do this if hydrotest is complicated.
I'm a bit concerned with your line of thinking. Lets talk about context...
The gas service process pressure vessels we all gleefully walk around most likely passed an initial pressure test (hydro or pneumatic) at fabrication and then have further proven themselves in service to have acceptable pressure integrity by not blowing up!
A newly fabricated vessel on the other hand has unproven pressure integrity and may have undetected defect(s) that will initiate a pressure vessel failure under test conditions. Under a HYDROtest, the consequences of failure are much reduced since water is an incompressible fluid, but if the vessel failed under pneumatic test conditions, the outcome would be very different - explosive infact.
The two situations [pneumatic pressure test vs an operating vessel in gas service] may have the same stored energy, but that's probably where their similarity ends.
Please don't try to pneumatic pressure test, either positive or negative, large API storage tanks with anything but water. I know of 4 deaths with pneumatic testing shop and site built API tanks. I have seen the smaller variety round the bottom out when tested with N2 instead of water, not anchored properly and few extra inches of pressure.
We always try to work around any pneumatic testing as such with such as jacket piping where Therminol Vapor is involved and water is very bad. A new pipe spool is heat cycled to our cleaning temperature and then the core pipe is hydrotested to the proper test pressure and the Therminol jacket is test with helium at 40 psig, normal hydro would be 125 psig, and sniffed. All jacketed components are He tested after the thermal burnout and bead blasting. After assembly the whole system is put under house vacuum and the decay cannot exceed 1# in one hour. If there is and indication of a leak that section is put under 25-35# He and some goes looking.
One some very large systems that operate at 150 psig prior to the introduction of the process, air and cyane, we pressure up with N2 to 70 # and watch the pressure decay. The leak rate allowed is based on an equivalent amount of Cyane vapor.
Even equipment design for high pressure air like Scuba Tanks require safe guards in filling and testing. Early on in development of Diving as recreational sport we had local dive boat that had several of the small surplus compressors used to fill tanks while traveling. One of the tanks ruptured and made the boat itself a popular dive site for several years afterward. You never know what’s going to happen.
No matter what you are testing or operating keep your hands in your pockets and not feeling for leaks. Embolism from gas, water, or oil to the extremities isn't a very pleasant sight.
Of course,if you have the option of an initial hydrotest, that is always the best solution.
There are cases where a hydrotest is not possible or practical. In those cases where a hydrotest is not possible or practical, a properly conducted pnuematic test is the correct manner to confirm structural integrity of the pressure vessel.
The dangers present with a pnuematic test are significant. Given that it is likely to be the first time the vessel has beeen strained to its design limits, the possibility that failure can occur is present. In addition to the safety protocols outlined in ASME sect VIII, other precautionary measures should be considered, including add'l NDT of critical welds, as well as secondary review of all fabrication QC documents. Finally, one needs to ensure that the most likely cause for short term catastrophic failure ( crack propagation with null ductility) be avoided by ensuring all testing temps are well above any ductile brittle transition temp for base and weld metals.
With all the prior discussion of the damage that can ensue from a failure during pnuematic testing, the damage is no greater than that of a failure during normal operation at desing pressure or during a releif valve lift event. It seems unbalanced to have all concerns focused on the initial pnuematic test yet ignore the fact that the same risks are present during everyday operation of tanks that are filled with compressible gases.
Davefitz said: "It seems unbalanced to have all concerns focused on the initial pnuematic test yet ignore the fact that the same risks are present during everyday operation of tanks that are filled with compressible gases."
I'll grant you that is your view, but the RISK of an initial pneumatic test failing is much much greater than the risk of an inservice process vessel failing.
The steps which mitigate the risk are the pressure test plus demonstrated safe operating history. Knowing that the vessel didn't fail during the test and hasn't failed since the test allows us to walk freely around vessels without much thought to the stored energy just behind the membrane of metal.
In no way should this sense of security lead to underestimating pnuematic testing's inherent danger.
The amount of stored energy from a compressible fluid might be the same, but the risks are definitely not.
robsalv,
I've inspected a number of inservice vessels and in many cases would take no comfort in their track record. At least with new construction I have a better chance to know what I'm dealing with. Don't get me wrong, all the points above, concerning the inherent dangers of pneumatic testing are well taken, I just think depending on a vessel's integrity because it's been in service for a while could be a big mistake.
It's a base assumption in my argument that the process vessel is fit for service - it goes without saying really!
I certainly wouldn't be putting back into service any vessel I found with serious defects I can tell you!
My point is, it's dangerous to have less respect for pneumatic testing just cause you don't have process vessels with compressible fluids spontaneously blowing up every day!
Boy, there's a lot of stuff flying around on this one!
My two cents... the pnuematic test is safe if all the rules are followed... or so the Codes imply. My former employers would never do a pnuematic if a hydro could be easily done. The pnuematic test was only used with the generous application of NDE well above the norm required by the Code. Obviously this was due to the higher risks.
