Another issue i am having with floater ball valves is the valve keep failing the seat leak test 1

[Mous1747]

we test the ball valves as per 6D requirements, so when it comes to the seat leak test the valve lose pressure. we tried different seat materials including 15% glass filled and 10% carbon filled, but the results were the same.
i finally realized the OD of the seats was undersized because it was not dimensionally checked by quality upon receiving.i am not sure if this a critical dimension since the seat will expand under pressure anyway.
should i make the seat thicker? but this will increase the torque. i was trying to perform some FEA simulation but i am not sure how that will be helpful.

have any one here had this issue with leaking seats on a ball valve? and how did you go about resolving the issue? any advice will be much appreciated.

 

[Mous1747]

I am wondering if the thickness of the seat is not correct and need to be thicker to provide more compression, the problem is i don't know how to determine/calculate and then optimize the compression. i have seen a range between .35 to .70 mm

 

[JoeH79]

If anything virgin PTFE is more likely to seal better than reinforced PTFE as it is slightly softer as a sealing service.

Occasionally manufacturers change the design of valves but with ball valves this would generally be around the stem packing area, but is work checking with your supplier in case there has been recent changes. Is it a new valve or one being repaired? Is the ball damaged or scored in any way?

One point that is often overlooked with floating ball valves is that while they are bi-directional they are designed to seal under pressure i.e. the upstream pressure pushes the ball into the downstream seat to create a seal. I have come across it in low pressures under 2 bar where this is the case and also where the valve has been operated at its full pressure rating and then reduced to 1-2 bar.

In such a scenario Trunnion Mounted valves are worth considering since the seats are spring loaded and will seal more effectively at low pressures

 

[Mous1747]

Hey JoeH79, thanks for looking into this. it's actually a brand new valve and we perform the seat leak test as per API 6D, so for an 8" 300# valve the seat test pressure will be 814 psi for the duration of 5 minutes as per API 6D clause 11.4.2

it always leak or lose pressure at about 800 psi which is very close to the passing pressure. i think it's a dimensional issue, since the OD of the seats are undersized and egg shaped, but i am told by the seat supplier it's normal for PTFE to distort.

could it be the thickness of the seat that needs to increase? i would like to know how to calculate or determine the compression per seat in mm

 

[Mous1747]

MVPs of valve engineering . can i have your inputs on this issue ?

 

[Danlap]

Hi Mous1747,

Your story is not detail enough. Drawing and additional information will help others to provide insight for your problem.
For instance, please help providing answer after my question mark below

Is this a new design from the manufacturer?
(Virgin) PTFE prone for distortion due to high pressure closure testing. That is why high closure seat testing as per ISO 5208 is optional, since it may damage the seat prior actual usage in the field.
PS: API 6D is closely related with ISO 5208; and also most field with 300 psi pipeline rating is not normally operated at 814 psi as your testing pressure.
The way some of end user interpret this statement is that: the prototype testing (valve number 1) subject for high pressure closure testing, and the production test (valves number 2 and so on) in regards with HP seat test might be waived.

What is your testing medium? Gas or water
Is it possible for you to test the valve without blind flange installed on the measured side? If yes, is it possible for inspector to see where is the leakage located. Is it from the seat side or from the back seat side (secondary sealing)?
What is the design look like? Not to generalize the concept of ball valve, however most of medium to big size floaters with soft sealing looks like below

Behind the seat usually designed with two extra sealing with either o-ring and or PTFE.
To provide effective sealing, the grove dimension (including the OD of the seat) of the extra sealing behind the seat should be appropriate.

Did you perform any cycle (open to close and the other way around for x times) prior the seat testing? Or just 1
8 and 10 inch ball valve for some reason indeed quite challenging for some manufacturer to perform prototype testing statistically spoken.

If this is without any cycle test, without any secondary sealing design, my suspicion would be indeed there is a seat design issue including dimension and tolerance. Not necessarily only the OD.

Kind regards,
MR

https://nosuchvalve.com

All valves will last for years, except the ones that were poorly manufactured; are still wrongly operated and or were wrongly selected

http://www.eng-tips.com/faqs.cfm?&rat1=2&fid=376

 

[Mous1747]

Thanks Danlap for your detailed answer. unfortunately in my case we don't have the luxury not to test the seats, we test the seats 100% to min 814 psi for #300 for the min duration as per 6D. as you see here in the drawing, this design is very simple just the seat in the seat pocket. (no secondary sealing design)

the seat comes in already distorted from the supplier the OD is undersized and egg shaped. i believe the residual stress in the material from machining cause the distortion. therefore we looking into the option of stress relieving of semi finished seat then do final machining after the heat treatment. or using a mandrill to hold the ID.

i think the main issue here is the thickness of the seat and the corresponding pre-compression, i am not sure how to achieve this optimum pre-compression to avoid the leaking of the seats, i thought about making the seats thicker then keep skimming of the bottom of the seats few thou at a time until reaching the targeted thickness.

we test with water and we don't perform the QSL cycle. is there a way i can put the seat on the ball and hold it with hand the fill the seat with water and see where it's leaking from? im not sure if this would be a useful test?

 

[Danlap]

Hi Mous1747,

First of all, I have to say I am quite impressed that this design (without Seat holder), this size and assuming this is the manufacturer prototype capable to pass the test up to 800 Psi.
Did you ever thought to change the material with the one that is less prone for distortion e.g. PEEK or others?

If your egg shaped judgement prior used is correct, then yes you may limit the risk of machining faults by the method you've mention above. This is quite experimental since I am not sure how does the pressure and the weight of the ball (assuming the valve tested vertical) resulting the downstream's (supposed to seal) Seat expand against the seat pocket accordingly.
Plus this is also a very thick PTFE Seat

If the valve tested on horizontal position, there is also a likelihood that the force towards the seat is uneven due to the gravity. The force will be slightly bigger on the bottom side (crushed due to the weight of the ball). therefore bottom side will seal better in comparison with the top side

Prior testing, yes you can sit the Seat on top of the ball and pour water or oil in it. Hearing your stories (capable of seat tight up to 800 Psi), I doubt that this method will show you something since this is water with some viscosity and at atmospheric pressure.
Or during the test you can do this also in a safe manner, unblind the downstream side of the valve and put camera with point of view capturing the whole ball area. Dye the water with corrosion inhibitor or something else.
If it start leaking on the top side at valve horizontally tested, then maybe the ball weight theory is a considerable factor.
Seat holder??

Kind regards,
MR

https://nosuchvalve.com

All valves will last for years, except the ones that were poorly manufactured; are still wrongly operated and or were wrongly selected

http://www.eng-tips.com/faqs.cfm?&rat1=2&fid=376

 

[Mous1747]

Hi Danlap, the seat pressure test is done vertically on a test stand. so it's not the weight of the ball. i think because the gap between the seat and the seat pocket will create a leak pass from behind the seat ? so if i have a snug fit between the seat OD and the pocket ID will eliminate this leak pass ?? is there a calc sheet out there for floater ball valves? i would like to perform seat calculations. i would like to perform a FEA simulation as well maybe i will have to re-design the seats. is there anyone have experience on seat re-design or simulation?

 

[Mous1747]

issue resolved and lesson learned

i just increased the height of the seat and increased the (pre-compression/squeeze/crush ) from .45 mm per seat to 1.45 mm