38" triple-offset butterfly valve passing - URGENT 2

[thermmech]

Greetings!

We need urgent help

6 pcs of 38" triple-offset butterfly valves are passing (more than 50% of number installed). The valves worked fine for about 12 months. Increasing the torque from 75% to 100% didn't help.

Service: isolation valves for mercury removal filters (dry natural gas)
Class 600#
Pop = 66 barg
DP = 76 barg
Top = 25 C
DT = 82 C
Materials:
body = A351 CF8M
seat = 316SS / Graphite

Plant is 1 year old, there is a possibility of large quantity of dust / scale in the piping. In theory, these valves should be somewhat OK with solids in process fluid.

I haven't had a chance to talk to the OEM, will do that tomorrow.

Any comments on cause, rectification would be much appreciated. If you need more info please advise.

Thank you,
S.

 

[tonyliyong]

Hi gent, can you please provide the state of disc? There maybe a couple of reasons which may cause such conditions. you may exclude any doubts one by one. Anyway, you'd better check two points following:
1. Can the lever be turned? what's the angle of availability?
2. Check the disc position, open or close?

 

[gerhardl]

What is the actual flow and pressure?

When 'something' is going wrong on several valves after a certain periode of operating time, and nothing 'unregular' has happened (no pressure peaks, no un-normal operating conditions, no service (opening of pipeline, foreign objects introduced) etc., no wear or mechanical slip on actuators, no wear or slip on adjustment and mechanical transmissions on outside contraptions and no unnormal temperature variations, this will, as you suggest, be inside problems.

Can you rule out condensation on inside pipeline caused by standstill and sinking temperature, or other inside moisture problems causing particles to clump or fasten to seats?

Can you rule out chemical reactions and chemical wear?

The inner problem(s) is most likely caused the same thing for all valves.

Guesstimate is most likely wear on seats again caused by abration, and/or increased by cavitation caused by valves not beeing able to close 100% because of inital abrational wear or dust particles on seats (tiny opening increasing velocity of gas).

If this is the case you have to look at the seat construction, materials and exposure of seat sealings and sealing rings to evaluate improvements.

An idea could be to evaluate softer seat sealings.

Note: Some operators may tend to hide 'un-normal' incidents, initial wear problems could be caused started by 'false operation', giving condensation or other irregular happenings as mentioned above.

If anything else is not possible you are left with regular maintenance (which of course would be cost intensive....)

(Please spell out acronyms and give flow data......)

 

[JLSeagull]

Which triple offset butterfly valves? I have rouble with API 6D ball valves and was just about to recommend some of the triple offset butterfly valves that claim zero leakage.

 

[thermmech]

Gents, thank you on your responses and sorry for the delay.

I'll try to address all the questions:

Manufacturer: one of the top rated ones, can send info via PM
Leakage class: API 598, TSO
Flange class 600#
Operating pressure = 66 barg
Design pressure = 76 barg
Operating temperature = 25 C
Design temperature = -5 / 82 C

Correction to body material = A216-WCB
Trim type = Rotary Vane, Disc
Disc material = A351-CF8M
Seat material = A240 TP316/Graphite

Torque = 240 kNm
Flow = not specified on data sheets, i have asked operations to provide info
Stream = treated gas, asked operations to confirm presence of particulates

I have talked with the OEM; they advised that torque increase be carried out when their representative arrives (don't know when). They also asked us to confirm proper orientation of the valves as, although they are bi-directional TSO (tight shutoff) valves, they seal better from one side than from the other.

I also do not know the state of the disc and seats - cannot shut down the unit to open the valves.

There should not be any water in the line as dehy units are upstream of mercury removal units where these valves are.

I am aware that operators might not give me the complete picture, but am trying to compile the history of the valves' operation since start-up.
Currently I am trying to confirm actuators are correctly calibrated.

I will provide info as soon as I get it.

Thanks again on your assistance.

 

[thermmech]

Gents,
I got the answers on the following:

Flowrate = 1400 MMSCFD Sweet Gas
Pressure drop accross the valve when closed = 65 barg, valve not used for throttling
No particulates in the gas

Thanks and regards,
S.

 

[gerhardl]

I suspect that you will not be able to give a complete diagnosis before valves are inspected inside. This will also give you a check if no moisture and no foreign particles are 'true real live operating information'.

Are the valves truely mechanically locked against possible axial movement for stem/disc combination, displacing or disturbing centering of disc and sealing correctly against seat? To be checked when valves are opened for inspection.

Seal and seat wear is still my first guess, but this may be caused by one or several reasons in combination.

 

[NGiLuzzu]

thermmech,
if I may start with a stupid question: how do you notice that the valves are passing without shutting the unit down? Is there a pressure increase downstream, for example? Are you able to assess the entity of the seat leakage, somehow?

In my opinion, a stainless steel (CF8M) disc within a carbon steel body (CF8M) is not the best choiche, but as the temperatures range is not very wide, this should not be the main problem... More probably, it is the seating element in 316 and graphite to become critical in precesence of dust, scale, solids and/or hard particles.

Is that "seat" on the body or on the disc?

And how about the other mating element?

Based on my experience, for a similar service I would have suggested a "solid" (i.e. full metal) seal ring on the disc (eventually hardened with HVOF or other treatments) against a seat surface on the body, hardened with a ground Stellite weld overlay, for example.


Hope this helps,
'NGL

___________________________________

 

[thermmech]

gerhardl,
Thanks on your responses. There is a possibility some of the valves have not been properly oriented, but not all of them. Means we will have to shut the unit down to inspect the valves and determine the root cause.

Anegri,
These valves serve to isolate mercury removal filters (2 in parallel). There is a pair of butterfly valves upstream and downstream of each filter vessel. Also, there is a bleed connection between each two butterfly valves.
Upon closing of all 4 valves, we still detect pressure between each pair of valves.
As 1 filter has reached capacity, but cannot be isolated for cartridge replacement, now we operate on the other spare filter, but valves are passing on that one too.

Regards,
S.

 

[thermmech]

Gentlemen,

We had a visit by the OEM, it seems that these butterfly valves were installed "in reverse" orientation downstream from filter vessels. The valves have been installed with the arrow on the body matching the flow direction while correct orientation would in the opposite direction as the arrow is a "pressure arrow" pointing from higher to lower pressure" and not a "flow direction arrow".

So, in spite of nominally being bi-directional TSO valves - they are not as they seal better from one side than from the other.

We will open the valves soon I think and see if we can repair them by replacing the seats as by now due to extra torquing the seats might have been damaged.

Thanks for your help!

 

[tr1ntx]

You mean the flow doesn't flow from the high pressure to the low? ???? ???????

Sounds like CYA by a manufacturere who screwed up.

 

[obv23]

All though that most of manufacturers claim that butterfly valves are bi-directional- they are wrong - you should take care of the mark ''main pressure direction'' because the opposite side is 30% weaker in most case.

 

[Sbelt]

thermmech,

I understand the OEM's recommendation regarding the valves being installed in the reverse (incorrect) orientation downstream from the filter vessels. Same issue exist for valves with a preferred high pressure side on pump discharge isolation. When the pump is down for maintenance the discharge isolation valve is required to seal against pressure in the discharge piping system. If this valve were marked with an arrow (not all triple offset butterfly valves are) the correct installed orientation would result in the arrow pointing in the opposite direction of flow.

But, in previous post you indicated that the upstream valves also leaked and I'm assuming these valves are installed in the preferred orientation. If so, then this (installed orientation) would not explain why these (upstream) valves are also leaking.

Also in previous post you indicated that the valves worked fine for about 12 months. As indicated by another in their reply this seems to me to also warrant further investigation.

My experience with triple offset butterfly valves regarding seat leak performance is that although when tested in the preferred pressure side the seal is typically better the difference versus the non-preferred side is not significantly different. For example the triple offset butterfly valves I've seat leak tested (most of the top rated ones) when requested to test to API 598 in both directions most always pass in the non-preferred direction as well. Although there is one possible exception and that is when the valve is actuated with something other than the OEM's supplied gear operator (i.e.pneumatic,electric actuator).

The operating torque required for a given triple offset butterfly valve at a given differential pressure is not the same in either direction (preferred/non-preferred). The operating torque required is actually lower in the preferred high pressure side direction (shaft upstream). Because most triple offset butterfly valves are torque seated, the operating torque value utilized for actuator sizing and selection is critical to the valves seat leak performance.

I have witnessed before failed seat leak test in an actuated triple offset butterfly when API 598 seat leak tested in the non-preferred direction (shaft downstream). In this case the actuator was a pneumatic spring return type in a fail clockwise (closed) configuration. The actuator was correctly sized for this valve based on operating torques in the preferred high pressure side (shaft upstream) direction. When tested in the non-preferred direction the actuator's spring end output torque was not sufficient enough to maintain a tight seal and the valve failed the seat leak test.

You indicated that these valves are actuated and listed a torque value of 240 kNm. You also mentioned increasing the torque from 75% to 100% and I was curious what that was in reference to. I will also mention that most triple offset butterfly valve manufacturers will provide on request a maximum allowable stem torque value for each valve size and stem material (including brand that starts with A). As long as you did not exceed this value I do not believe you should expect to find seat damage due to extra torquing as mentioned.

 

[thermmech]

@tr1ntx
The valves downstream from the depressurized vessel will see the high pressure from the downstream side. Thus, normal flow direction doesn't match hi-lo pressure direction when valve is closed.

@obv23
Thanks on your comment. OEM stated the same thing even though the data sheets cite TSO-bidirectional.

@Sbelt
Thanks on sharing your experience.
The 1 out of 2 upstream valves is passing, however it turned out it is minor.
As to a duration of valves in service: valves are closed every 2-4 months depending on need to isolate vessels. It is not obvious whether they had a problem 6 months ago since maintenance people could have used different methods to manage to isolate the vessel in spite of minor valve passing - and not report it to us.

I feel that the orientation is the main cause of the problem, while there could be contributing factors we are not aware of at the moment. We will know more when we are forced to shut down and open the valves for repair.

I just wondered - would a slab-gate valve be a better fit for this service?

 

[TOVmania]

This is simply a German company's design. Oviously, this might have been a problem of this type of TOD about a decade ago. I am involved in this type of TOD that I know such product will have a preferrable flow direction. There are some manufacturers of Triple Offest Design that actually improved such design and they provide BTSO in Bi-D. Look for them during a trade show.

 

[HitokiriAoshi]

Be aware that many valve manufactures advertise "zero leakage" to API 598 and API 6D, and also debates on the definition of resilient seated valves.

API 598 has allowable leakage for metal seated valves, which some manufactures use to test laminated seats. Also a thing to note is, manufactures can intentionally use excessive graphite in the lamination and extrude them slightly to achieve zero leakage when new, this graphite will wash off after usage and the valves will eventually develop leakage excessive of what it would if the graphite was trimmed properly at the onset.

I would ask your OEM to take a picture of the close up of their lamination ring to see what the construction is.


-Sniper out

 

[NGiLuzzu]

"Zero leakage" does not exist in nature, that's just a conventional expression and therefore must be defined in a standard (in terms of leak detection method, time duration, temperature, pressure, etc...).

Besides this, the mentioned standards regulate (production) testing of new valves, while evolution along valve service life may be lead to different results... (and the originally reported failure occured after 12 months service, for example).


Best Regards,
'NGL
_________________________

 

[JLSeagull]

So use the term zero leakage as a conventional expression and used the definition used within a standard. In this case, wWithin API STD 598 Table 5 the numbers listed for all valve sizes from less than NPS 2 through greater than NPS 14 reflect the number 0 in the column for All Resilient Seated Valves and a reference to note b. Note b:
"There shall be no leakage for the minimum specified test duration (see Table 4). For liquid test, 0 drops means no visible leakage per minimum specified test duration. For gas test, 0 bubbles means less than 1 bubble per minimum specified test duration."

And everything that anegri suggests about new valves vs. results after commissioning is correct. Regardless, zero leakage per API STD 598 is better than FCI 70-2 Class V or VI and the same issue after shipment applies. In any case if someone suggests that you could enter a vessel with just the isolation closed and no blind would be nuts. And yes I had a process engineering manager make that claim.

 

[TOVmania]

Laminated seats will create PROBLEMS ONLY when used in DISC SEAT TYPE of TOV.

In DISC SEAT TYPE of TOV, it is true that the seat is exposed to the medium during operation which accelerates wear of graphite, therefore, shortens the life of seat inevitably. Where as in body seat type, such rule does not apply. The wear is mostly caused by direct contact with the flowing medium not by itself. Disc seat type leader effectively used its shortcoming to promote solid seat as an advantage in the market. In fact it was only there to solve its seat problem of laminated metal seat in the disc seat type TOV.

I have seen with my own eyes of such "zero.00 leakage performance" of TOV even in cryogenic(-196C) condition with cycling tests procedures done simultaneously. Also other 3rd party LNG/LPG people were there go witness. Really...look outside of the box and search for new manufacturers in the market.

 

[NGiLuzzu]

TOVmania,
could You please define what do You mean with "DISC SEAT TYPE of TOV", precisely?

And be so kind as to tell us which unit of measure (and what accuracy) we must understand after the "zero.00" You mention?
;-)


Many THANKS in advance and Best Regards,
'NGL
_________________________________________________