Confirming full closure for ball valve 2

[valvesealing]

I am working in valve maintenance company and i find it difficult to confirm the ball is fully closed for the ball valve which has been in service for more than 10 years. There were several valves which were passing severely(in a minute the downstream pressure will be same as upstream) even with the indicator plate showing fully closed. Once we removed the valve and inspected the valve components expecting to find damage to the seat and ball. But instead the problem was with the ball not fully closing because of twisted stem and drive pin. Also there is no surety that stopper pins in the gearbox haven't been tampered with before. I would like to know whether there is any method adopted for conforming full closure of the ball.

Thanks

 

[gerhardl]

There is only one short answer to your question: yes and no!

The no part depending on that nothing expect changing the mishandeled vaalves is done.

The yes part depending on getting the 'helicopter view' and start with procedures and practising of a more updated and modern way of handling operation and instructions on how to do things.

The QA procedures and practise seems to have been way off target, but perhaps 'normal????' for your area/plant?.

Getting the way of handling things in place, the normal way of ensuring fully closed ball valves is using limit switches and/or mechanical end stops. In addition you will have to select strong enough valve mechanical components and correct selected materials to ensure correct operation with minimum sticking under normal conditions.

If sticking and mechanical overloading is 'normal' you will also have to consider other valvetypes and general pipelayout and operation and cleaning practices.

PS: Good QA systems start with getting the managing director involved and positive.

Good luck to you!

 

[zdas04]

I have to disagree with the "yes and no" answer. I don't think that there is any way to catch it. All of the indications and stops for a gear-operated ball valve are in the gear box. This is adequate for virtually all cases, but not all. A twisted stem is indistinguishable from a failed seat and I don't think you could ever tell the difference without disassembling the valve. I've never seen a twisted stem (they are massive beasts and I don't want to run into the individual who could twist one even with the mechanical advantage of the gears and handwheel) and would almost bet that it is a manufacturing defect that should have been caught in the factory leak test, but the valve was installed and in service for 10 years. I think the only options were to identify the problem, try to fix it with sealant injection, finally pull the valve and look at it on a bench.

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.

 

[valvit]

Many ball valves have mechanical stops built in to the valve to act as a backup if the stops in the gear or actuator do not work, or are not set correctly. Depending on the specific valve design, there are different ways of bypassing the integral stops (with a stepped key for example). When performing field service, it is often useful to know how to bypass any integral stops so you have a way of correctly setting the stops on an actuator or gear box. By allowing the ball to over-travel, you can drive it forward, and then back until you find the sweet spot where the valve will seal. By definition then, this will be the fully closed position.

Here is my method of finding fully closed and fully open positions:
NB. CW = Clockwise (to close), CCW = Counter Clockwise (to open)
1 Use a stepped key, or other means to bypass any integral stops on the valve.
2 Rotate the valve to the position where you expect the valve to be fully closed.
3 Check for leakage past the seat (or seats if pressurizing the body cavity).
4 Move the ball CW while monitoring seat leakage. If the leakage increases, reverse direction. If leakage decreases, continue moving CW until leakage stops or is below acceptance criteria.
5 Set the corresponding stop in the actuator/gear box.
6 Remove actuator/gear box and mark the position of the key on the mounting plate of the valve (MARK A). Do not stroke the actuator/gear box while off the valve.
7 Measure 90-deg off of MARK A (fully closed) and make a mark for fully open (MARK B).
8 Install the actuator/gear box back on the valve, and remove the top of the actuator/gear box (weather cover, position indicator, any other accessory) to expose the valve stem.
9 Make a mark on the top of the valve stem/key (MARK C) and a corresponding mark on the top of the actuator/gear box (MARK D). A continuous straight line works best for me. At this time, MARKS A, C, and D should all be in alignment, but MARK A may not be visible with the actuator/gear box installed.
10 Measure 90-deg off of MARK D (fully closed) and, on the top of the actuator/gear box, make a mark for fully open (MARK E).
11 Rotate the valve CCW and align MARK C with MARK E.
12 Remove the actuator/gear box. Again, do not stroke the actuator/gear box while off the valve.
13 Using a machinist's square or other suitable straight-edge, visually compare the alignment between MARK C on the top of the stem, and MARK B on the mounting plate. Most of the time, these two lines will not align due to lash-up of the gearing in the actuator/gear box. As accurately as possible, measure the angular difference between MARKS C and B.
14 Install the actuator/gear box back on the valve. Add or subtract your measured angle from MARK B, and on the top of the actuator/gear box, make new MARK F. Move valve CW or CCW depending on the measured angle, so MARK C now lines up with MARK F.
15 Repeat steps 12-14 until you have MARK C in alignment with MARK B. One iteration is usually enough.
16 Set the open-position stop on the actuator/gear box.
17 The open and closed positions are now set.

The above technique is used when you do not have visual access down the valve bore. If you do, then the above is not necessary.

Technically, this would still work with a twisted valve stem, but if you suspect stem damage, the only safe option is to remove the valve and repair it before continuing its use.

 

[valvesealing]

Thanks for your suggestion. The valves which i am dealing with, were in service for more than 10 years in gas lines(basically crude) and in course of times, the valve operation gets harder. The operators use all their energy on these valves to cycle(even the actuator operated valves will be operated manually, once the torque becomes insufficient). Once these valves are beyond the energy level of the operators or in some cases, the valves which havent even operated once in 5 years(isolation valves) gets stuck and they call us. I wrote this to give a background of these valves that i am talking about. For these valves, even the use of superior sealants didn't work because the ball was not fully closed and the sealant went to the line through the gap.
I would like to know, is this case an isolated one (since for us we are facing a lot like this and zdas04, i had a chance to see even a "broken to shreds" valve stem. This one was operated lastly by four people using pipe for leverage. The stem broke in the neck portion where the packing) and gerhardl please give insight into what "start with procedures and practicing of a more updated and modern way of handling operation and instructions on how to do things." means(would seriously like to improve the procedures if there is any).

 

[zdas04]

Valvit,
That technique would work well for a trunnion ball valve where you can check seat "leakage" with the body bleed. For floating ball valves it can be nearly impossible to identify the "sweet spot", and the sweet spot may be different depending on the flow direction (i.e., small deviation from parallel seats).

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.

 

[al1979]

Yes, this could be an isolated case.

1) I hope gearbox details i.e. Mechanical Advantage is available G-box nameplate
2) Normally gearbox / actuator selection is 1.5~2 times of valve torque.
3) With pipe leverage, if you are exceeding the gearbox torque, overhaul the valve immediately.
4) Stem twisting normally occurs with SS316 stem material. From valve nameplate, you can take extra precaution while applying the input torque.

 

[zdas04]

I've seen the "4 gorillas with 36 inch pipe wrenches" scenario too (once a guy put a chain on backhoe bucket to try to unstick a valve, didn't end well), but every time I've seen it the shear pin on the handwheel broke before the valve stem (once the gears stripped before the shear pin went, but usually the gorillas bust the shear pin). Shredding a valve stem sounds like a materials issue to me.

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.

 

[gerhardl]

Sorry all, still yes and no!

Anyway valvit deserves the star for a detailed procedure example! Regarding all answers above: yes, it is possible to twist a ballvalvalve stem, depending of size and solidity of stem and torque applied (I have seen it myself. Worst exaple heard of: lever operated by truck!).

Yes, it is possible to overcome unpreciseness and disfiguration of a a gearbox, again depending on solidity of gear and type, compared to max torque. As an example: correctly dimensioned slider cranck mecanism gearbox could be a good choic.

My comment: as an engineer you can take the purely mechanical view, which seems to be the way everyone above have responded. But: correct engineeribg will only take you halfway if the surrounding universe does not arrest anyone breaking the standard mechanical rules.

Not operating and control a valve for five years? Possible but not recommended!

Applying a torque above max limit? Max limit unknown? Possible but not recommended!

You have to know the failures 'normally' occuring and the limitations applied to avoid mishappenings.

This alone should illustrate some basic points caretaken in normal QA procedures for operating and maintenance, along with the control of procedures actually been followed.

 

[zdas04]

QA, writing procedures, enforcing the following of procedures, and maintenance all have a place in prudent operations, and will frequently prevent/identify damage to valves. That wasn't the question. The question was "is there any way to confirm if a valve is "shut""? The answer for a floating ball valve is "no". For a trunnion ball valve valvit's procedure would have found that the valve wasn't going shut (not necessarily why it wasn't going shut, but the fact of the misalignment would have been evident).



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.

 

[IRstuff]

If one were inclined, one might imagine embedding a magnet within the ball valve and using a Hall sensor to detect whether the ball valve was in the correct position. However, this would not necessarily be applicable to an off-the-shelf valve.

Another option would be to monitor the downstream flow.

TTFN
faq731-376

 

[bcd]

In the end, one or more of three things happened. Bad design (ball valve stem should always fail externally). Or loss of strength due to corrosion (should be able to see the loss of material from the surface). Or the stem material was defective.

My hunch is that the valve design is not suitable for this application. Why does the torque increase so much over time? Is media or corrosion getting behind the seats? If this is a metal seated ball valve, are the hard coatings on the ball and seat failing?

Problem solution involves two parts. Learning why the torque keeps increasing over time. And secondly, ensure the ball valve stem design meets API requirements in that it must fail external to the packing so it can be seen. Do not rely solely on a shear pin in the gear to protect from excessive application of torque. Because once the shear pin breaks, then the truck gets attached to a lever and torque is applied directly at the stem by the gorillas.

 

[zdas04]

IRStuff,
It is cheaper to always use trunnion ball valves above about 6-inch. Then you can override the stops, open the body bleed and try to find a sweet spot, if it is a stem problem there will always be one. I find floating ball valves a false economy in larger sizes.

bcd,
If I was betting my own money, I'd say that Supply Chain Management specified a "generic" ball valve and took low bidder. There are companies in Shanghai who will sell you a valve with a code stamp that doesn't meet any of the code specs (I got one once that was 3-inches shorter than API 6D specified for that size and since it was a replacement the hole was already there, but the valve had a 6D designation on the nameplate). I find non-mainstream valves to be way more trouble than the cost savings would support. There are some good suppliers in China, but there are also some true bottom feeders that give the whole country a horrible name in this field. Careat Emptor applies, but SCM is a bit short on the whole Caveat part.

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.