Valve leakage 5

[pipexp]

Dear experts,

I have a special problem with respect to valve that are used in offshore oil and gas industry. A perticular manufacturer supply us valves complying all the testing and specification requirement. But it is observed the all those valves start passing within a very short span of time. Therefore, I would like to know what are the things that is required to review during finalisation of purchase specification.

It shall be of extreem help if you kindly help me in this regards.

Thanks.

 

[bcd]

Sounds like the valves passed factory leak tests, but quickly leaked once put into service. If so, the problem is not that something was missed during finalization of the purchase. The problem is that the proper effort was not made before purchasing.

Keep in mind that the factory tests are not intended to prove the valve is going to work in a particular application. They are a quality check that verifies the valve was manufactured and assembled correctly. Work needs to be done upfront, before purchasing to verify the valve is the best choice and will work in the application. You can request special testing that may duplicate the service conditions (to the best that is practically possible), and learn if the valve will work. But then it is far too late and a waste of everyones time if it doesn't pass this test. It is also much more expensive to do special tests. Just do the work upfront before purchasing.

bcd

 

[dcasto]

Again is the valve specified the correct one for your service. My rule of thumb for any valve is that it is just a piece of pipe with some Cv value, they all leak at some point. So, what is your leak rate requirements and what do you have in place to make sure YOUR OIL & GAS always meets the EXACT same specification you put set when you specified the valve?

Do you tell the vendor that there maybe some sand? How about water? Is it always 100% gas or can there be free liquids?

There are ANSI B16.104 specification a leak calss I,II,III,IV,V. This gives leak rates at different delta Pressures ect..

 

[BigInch]

What basic valve specification did you use?
Were they API 6D valves?
Was the valve model one that has been used in similar sub-sea service?
Was any special testing such as an extended high pressure leak or strength test specified in addition to that required under the basic specification?
Was any inspection of the valves done at the factory?

http://virtualpipeline.spaces.msn.com

 

[pipexp]

Yes, it is API 6D.
NOT SUB SEA VALVES BUT TOP SIDE PLATFORM VALVES.
No additional testing done.

 

[BigInch]

Probably should specifiy a longer and/or higher pressure factory leak test for any valve made by that mfgr in the future (if you ever buy valves from that vendor again).

http://virtualpipeline.spaces.msn.com

 

[proinwv2]

I am on the road and don't know exactly what the API leak specification is but the ISA/ANSI specs all deal with the allowable amount of leakage. It is a matter of degree. And, as previously stated, it is a factory bench test, not a real world test. Actual conditions can quickly degrade the performance due to temp or pressure, cleanliness of the fluid, strains in the piping to name a few. For critical apps you do need to carefully determine your needs and come to an agreement with the manufacturer before the order process is completed.

Paul Ostand

http://www.ostand.com

 

[JLSeagull]

ANSI B16.104 is obsolete and withdrawn. See FCI 70-2 for Control Valve Seat Leakage. This is much greater leakage than required by API STD 598.

It is common for new valves to meet the zero bubble requirement in API STD 598 when new. Sand was already mentioned. The nasty rusty hydrostatic testing fluid is often left in the lines for a prolonged period leaving scale and such in the seats. Such practices cause the valves to leak even during the initial startup.

Some attempts to fix the problem include adjusting the limit stops to get a fresh piece of the seat into service in addition to the lubrication tricks.

 

[Valvecrazy]

There are really only two kinds of valves available, those that leak and those that are going to leak. Any process that is counting on any valve to be completely leak free, will fail sooner or later. Processes should make allowances for a certain amount of leakage from any valve.

 

[JimCasey]

One trick not mentioned is to use a plug valve.
Nonlubricated valves have a (usually) PTFE sleeve for sealing. The plug has a 2 degree taper. If you catch 'em leaking, there is an adjustment to drive the plug farther into the taper, restoring the shutoff. Nonlubricated plug valves are available thru ASME class 600, Lubricated plug valves are available for higher pressures. Ball valves are hard to turn, but comparatively, plug valves are HARD to turn. Bring a long, long, cheater bar.

 

[NGiLuzzu]

I agree with many of the opinions expressed above, most of all with bcd's one.
I would add that sometimes, for very special services, it should be better to get a "trial valve" (the Manufacturer may be willing to provide it for free or at a reduced price) and test it directly on the service, if possible, before placing the order for the whole supply. A benchmark comparison between different valves should be also a advisable practice, whenever economically and practically feasible.
Laboratory or factory tests, instead, could be very time and money expensive but very hardly able to simulate the actual working conditions...


Regarding final factory pressure testing standards, it must be observed that API 598, EN 12266, ISO 5208 and similar ones are applicable to valves in general, while (if read strictly) FCI-70-2 and European equivalent IEC 60534-4 are applicable to control valves only.


For more information, maybe you'll find useful to take also a look at:

- thread408-76110 about the latest edition of the FCI / IEC standards;

- thread408-139863 and thread408-122060 about the seat leakage classes defined therein.



Hope this helps, 'NGL

 

[277Peter]

Dear pipexp

Select valve from a reputable brand, allow enough time in your project schedule and join the queue. There are good reasons why the best manufactures have full order books and normally it comes down to reliability.

 

[JLSeagull]

What reputable brand or pattern?

There are several reputable brands, valve patterns, trim details etc. Ball valves are often applied to gas pipelines and operation unless the temperature is elevated. I have seen valve engineers specify all sort of details regarding the valve internals. In my experience, most still leak after being in service for a short period.

Which are these best manufactures? They all have full order books during busy times. Do the good ones come from Italy or Texas instead of Oklahoma or China?

Do we damage the valves during construction and commissioning? Can poor construction practices damage the valves?

The valve selection needs to consider the fluids. However I am sure that many valves are damaged between the factory and the initial startup. Unless you can name the valve that is immune, this damage would likely be independent of the manufacturer.

 

[BigInch]

If they were inspected at factory, at fabrication laydown area, after installation, and use and removal, the reason for fault could be readily determined.

I've seen all of the above reasons for fault at one time or another.

JLSeagull, I DO think (and thought it was rather obvious) that there are better valve manufacturers than others, as well as better construction contractors than others, etc. You don't really think that all valves begin with the same quality when you take them out of the box do you? Did I misunderstand your implications?

http://virtualpipeline.spaces.msn.com

 

[UNSR60001]

pipexp,

I would be very happy to write an opinion should you provide the following details:

Application.
Line size.
Preassure, min., normal, max.
Temperature, min., normal, max.
Fluid.
Application.
Current Valve type and configuration.
Encountered problems with existing valve.

 

[JimCasey]

Here in the secret valve factory it is not uncommon for a customer to ask us to perform durability tests.

50,000 cycles on circulating bentonite was a pretty neat test. We also have done tests on dowtherm, water, air, and LN2 recently, and actuator tests in a freezer or in an oven.

Like UNSR60001 says: lay out all the service conditions and tell the vendor where you are having problems. Ask for HIS solution and ask him to demonstrate via test. Then buy enough valves from him to make it worth his effort. ( a lot of customers forget that last part)

 

[dcasto]

SAFE T SEAL or POWER SEAL, can't go wrong if the valve cannot leak. IMHO

 

[BigInch]

OK let's get the real story. I will just bet that its the end result of a typical performance vs price conflict. Did you happen to go with the lowest bid on that job, a new valve manufacturer, fastest delivery time, or did you chose valves with a proven history of acceptable performance on similar installations with your company?

http://virtualpipeline.spaces.msn.com

 

[Ashereng]

There are many reasons why valves fail, and many are due to reasons other than the fault of the valve.

If your valve passed the leakage test, there is a good chance the problem rests elsewhere.

Like others stated above, selecting the correct valve requires more than chosing the "name brand" from a reputable vendor. Selecting the correct valve for the particular service often requires experience. Someone mentioned to "trial" a valve. This is good if you can afford it - nothing beats personal experience in my experience.

Talking to others who have a similar application is also very useful. I have worked in a situation where people from 4 companies would share experiences on various valves, instruments, PLC apps, etc - as we were located in the same area, and our lines affect each others. Even though we worked for different companies (and some of us were direct competitors), sharing of infomation made everyone's job a bit more easy (and none was prorietary)


Warning:

Every valve will leak, it is a matter of time.

Okay, off of the soap box.


If one valve is leaking, have you tried 2 in series? For the most critical application that I have come across, we had 2 parallel trains with 2 valves in series to try to ensure the flow will be stopped. The paralle train was so maintenance can replace the valves. This is one company that is series about stopping that flow!

"Do not worry about your problems with mathematics, I assure you mine are far greater."
Albert Einstein
Have you read FAQ731-376 to make the best use of Eng-Tips Forums?

 

[Gorman3]

You may want to consider the possibility of damage occuring during you testing. If valves come to you with extra sand, dried grease, or u are using very dirty testing water you could be causing failures that won't appear until after the test.

Do some inspections on seats after the test, see if there are any scratches that were not there previously. If so the most likely culprit will be sand or some other matter, possibly welding remains,that are still in the casting before test.