LARGE DIAMETER BALL VALVE LEAKAGE 4

[BFZHA]

Hi,
During last years I have been involved in projects for gas pipeline and gas facilities. The requirements for soft seat ball valves was zero leakage as per API6D. All factory tests certify that all ball valves tested do not present any leakage. At this step, all things are right.
BUT, few years (2 or 3) after their installation, a small leakage are detected by operators at the seats for large diameter valves (16", 40" and 48"). We note that the valves were supplied by different manufacturer. To solve this problem, sealant was injected to eliminate leakage.

The issue I want to discuss here is that for large diameter ball valves, it seems that is impossible to "conserve" the zero leakage feature of the valve after some time of their installation or use. This is due, in my opinion, to the inertia of the ball that exerts pressure on the soft seats so that after some cycles it causes deformation causing small leak at the seat.

So, I want to know if some one has observed the same problem. If it is the case, can we observe the same problem for metal-metal seats ?

I want to know if there any study concerning this issue?

Thanks

 

[Danlap]

Interesting topic of discussion.

I believe in order to assess the problem, all of us requires more detail information.
Since this is concerning big Nominal Sizes Ball Valve, I assume you're talking about Trunnion mounted Ball Valve.
Soft seat, trunnion mounted design supposedly meaning that the seat shall be capable of withstanding in either flow direction its pressure/temperature ratings.

What is the seat design? Standard seat with soft goods (i.e. PTFE, etc.) filler, is it with Double Piston Effect or Single Piston Effect Seats?
What is the Normal working condition? Normally Open / Close? Operated in daily/weekly/monthly basis?
What is the (actual) cleanliness criteria of your medium? Is dirt expected to be present? corrosion might be built up due to void period, etc.?

Correct me if I am wrong, however specific study about this (exact type, exact manufacturer, exact material) has never been published. Since like for your example, most Plant (refinery, platform, etc.) are using different manufacturer (hence different material and design --> ultimately different lifetime).
If there is any study, it would be for private distribution and consumption of the end user. And hardly ever published to public since it is concerning ones manufacturer reputation.
Alternatively, practitioners use a more statistical approach. Defined by RBI department, Engineering and other multidisciplinary specialist. They publish for instance OREDA (Offshore Reliability Data) which is quite handy to determine Valves' expected lifetime in regards with type, design and medium.

In regards with your case, I have face similar problem in some extend. My plant are using 6" to 8" Floating Ball Valve. PTFE Seat. Medium is Benzene. Normally Close. We are sure that there are no dirt or whatsoever that could scratch the soft seat. And apparently after 3 years it is leaking spontaneously on 5 valves in the same time.
My plant design limitation --> No pressure indicator(s) in between the valves, only several PSV. We know its leaking since the PSV's were shimmering quite often, Block valve that we know for sure were tight is located on the very end of the pipeline.
Long story short --> Cause: My Plant's operator Close the valve not in fully Close position, it were offset a bit. Since it is floating ball, the ball exerts Non equal pressure towards seat (since small area of the seat is not touching the ball. The seat become oval and then leaking occurs.

There are many potential different causes for leaking problem. I suggest to built Fault Tree Analysis.
[ul]
[li]Trunnion design is expected to have very limited "Ball" pressure against seat OR not at all in comparison with Floating.[/li]
[li]Manual Operated Ball Valve without "proper" stopper / indicator in the hands of different Operator (strong / not so strong Operator) may result in small significant different of valve position--> In reality may not exactly fully close or fully open.[/li]
[li]If all of the above is not applicable, most likely it is dirt that carried over due to your sequential operation. A hairline scratch on a standard soft seat Ball (Non-DPE design) valve would definitely cause leaking to the upstream side.[/li]
[/ul]

In respect with your case, it is quite strange since its happening below than 3-5 years, occurs on different valves (simultaneously I assume). Under the assumption that the selection of the valves were correct; Valve produced under good QA/QC; the operator knows what he was doing.

 

[LittleInch]

A true bubble tight seal on ball valves of that size generally only lasts until the valve is shipped out from the factory after test so getting 2 to 3 years out of it is going well...

The issue is that any minor scratch on the ball from sand, grit, weld metal etc. can cause a leak. With gas in particular, the issue of opening (or closing) under a differential pressure is what often causes seal failure. Even if an equalising line is used or only small volumes are involved, the initial gas jet just as the ball moves off the seat can be significant in terms of seal damage. If there are high liquid pressures like the benzene example noted above then again seal damage occurs in the very first part of opening.

Metal metal seats are not capable of sealing bubble tight without a soft seal lip or insert. These suffer the same sort of damage and can in the end cause more issues as if the soft seat is damaged there is less of the seal to still seal.

Is there a study - no for the very good reasons Muktiadi explains.
If you want or need true bubble tight seals over along period of time you need a different valve than a ball valve.

Some interesting test info here showing just how small a scratch it needs to be...

http://www.sealweld.com/pdf/paper-destructive_testing.pdf

My motto: Learn something new every day

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

 

[BFZHA]

Thank you for your reply.
First some detail information :
- The valves are trunnion mounted classe 300# and 600#,
- Seat are PTFE made and Single Piston Effect.
- The valves are Normaly Closed of course this is why the leakage was "easily" detected.

The fluid is a treated gas in the three cases. Generaly these valves are installed in lines downstream filters batteries and pipe are lined (internally coated). We can assume that the seat is damaged by "debris" detached from linning (erosion effect) or "jet effect".

So, if you agree with me, in all cases the leakage is inevitable for ball valves after some years of use.

The issue now, is the warranty with end user since the requirement is zero leakage. We can reach this requirement at the factory during tests only. But as we discussed this feature can not be conserved after 3 or 5 years of use.
Can we can talk about an ideal characteristic that tends to get lost in time. Then can we, tolearte leakage or downgrade the class of leaks based on years of use?

I think it is an interesting topic of discussion also.

Regards.

 

[gerhardl]

First: the answers above both explains detailed what is 'normal and expected' and often common reasons for leakage. Some comments in addition: A lasting bubble thight seal is only obtainable under ideal conditions, eg. with a clean fluid. Since you are assuming existing debris from the pipeline, the fluid will not be clean. Any particle large enough to scratch, abrase or assemble in amount as a layer on a surface will give a 'non ideal' fluid.

For the rest of your questions the answer will (as stated aboce) depend on a number of constructional details, regarding fluid, pipeline layot, including placement and type of valves, orientation av valves, any filters, types, placement etc etc.

Is the pipeline requrement pigable lines? Double or trippel excentric BFL valves could be an alternative, even with metallic sealings.

PS. If trunnion mounted BVs are not mounted with horizintal stem, this could in itself cause leakage over time for some makes.

 

[Danlap]

Bfzha,

Re: your issue,
Warranty --> the manufacturer(s) will easily pin you down, thus any legal action is a hard battle to win (for either contractor or end user). Why? a. you're using different Ball valve (despite the selection reason behind it on the first place); manufacturer(s) were not (deeply) involved for the valve selection. A good manufacturer (not all) would assess the process condition and provide the best valve material and or type. However, normally this would add on your delivery lead time and also the outcome will be reflected by higher cost as well (due to more expensive and presumably "stronger" material). b. Cleanliness of the medium cannot be guaranteed --> (Virgin) PTFE and standard Seat are susceptible for leakage due the smallest contaminant (i.e. debris, corrosion, etc.)
Normally legal action (warranty claim) can only be raised due to visible poor workmanship on valve(s), or other proven condition that had cause critical impacts (i.e. fatality, huge to environmental leakage).

Agree with LittleInch and gerhardl, Ball valve is not suitable for a zero tight leakage (after several period) expectancy. On top of that, Trunnion mounted Ball valve has the most spare parts compare to other valves. Thus, eventually (on a very fluctuated process condition P/T/phase), other problems are deemed to occurs (i.e. Springs (loaded seat) get jammed, wear on one parts would contribute to others failure, etc.)
Ball valves that I personally see have No failure report for at least 20 years (note: with allowable leakage due to metal to metal) are HIPPS, Subsea Ball Valve and small numbers of Engineered Ball valves (involves R&D from manufacturer, and end user). Other than that it is always 5 years or less (generally known as commodity ball valves) for end user to expect some initial failure to occurs (leakage, jerky operation, or even jammed, etc.). On a good company, normally this trends are documented in some RBI reports.

It is not an easy attempts on Brownfield to impose the usage of other valves type (triple eccentric BFV, Globe) shall Tight Shut Off is desired however 'like to like' Ball valve replacement is also preferred.
Jargon of 5 to 20 years life expectancy by End user is driven by emerging Material and Design R&D for valves in global, thus each manufacturer (big or small) are desired to catch up.
As for the 3 to 5 years period warranty by vendors, conservatively only applied under ideal condition (defined by vendors).

Regards,

 

[BFZHA]

Thanks for all of you for your reply.

when I mentioned the warranty period, I wanted to mention the financial impact of interventions in order to solve the leakage problem. the client always believes that since the valve was delivered with zero leakage it must retain this characteristic.

My question is, since we know from experience that ball valves are not 100% tight so why in API6D zero leakage is specified. That is why I mentioned ideal characteristic?

Regards.

 

[LittleInch]

I think you'll find API 6D simply allows for a set criteria for different leakage rates so that supplier and purchaser have something to deal with. Which class of leakage you want is up to you, hence it is not "specified" in ^D, simply gives you the option.

I've never quite understood why anyone specifies bubble tight or zero leakage as all it does is delay the testing and everyone knows that as soon as the valve is operated in anger, it pretty soon won't be zero leakage.

But I guess it sounds good to someone who pays the bill....

My motto: Learn something new every day

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

 

[BFZHA]

LittleInch, I agree totaly wih you.
However, the choice of type of valves was made by client during FEED studies. Unfortunately, during preliminary studies, it is not necessary that those who are close to field problems who specify equipements.
So studies or statistical approache will convince client to change his choices.
I think I need to spend some time to cover this issue.

Thanks for all of you and best regards.

 

[kang007]

We manufacture Metal to metal type ball valves for critical applications like as CCR, RDS, Coal gasification, poly-silicon with powder.

If you are interested on metal to metal, please see the following movies.

http://youtu.be/usMP0Vr95hg

http://youtu.be/aKSd3LPMOEI

We had some leakage test with helium gas 99.9% for 6"X600# valve, poly-silicon plant. the leakage rate is 2.2X10-6std cm3/sec.
Best regards,
Kevin. kang007@gmail.com