Valve passing issue few weeks after new installed

[Yumie]

Hi, its my first post. I'm a technician working on a gas platform somewhere in SE Asia

We have a condensate system that pumps condensate from separator into the main gas line, mixed with the gas and sent to shore via pipeline
The pump itself a 3 piston reciprocating pump with capacity of 58m³/h and operates at discharge of 95-100 bar (main gas line is 90bar)
Most of the condensate is recycled (approx 23m³/h/train) to a vertical separator which operates at 20bar

The main issue is the return line LCV we have suffers passing issue very frequently, we did replaced it last year and it lasted whole 3 weeks before passing signs was reported again
LCV is a 3" Mokveld Axial control valve and the pipe is 4" DSS, Sour
We also have some H2s and mercury in our system which may also contribute to the passing rate

So I was wondering if the large pressure drop ▲70+bar contributed to the issue, and if the mass flow of 23m³/hr also hurts the valve in a way?
Would an orifice plate were added upstream of LCV, would that help drop the pressure by one stage, thus lowering ▲P at the LCV and lessen rate of wear on the valve?
Would lowering return mass flow from 23m³/hr to 5m³/hr also helps?

I have a few ideas but had no one to validate. I'd be happy to provide more information to my best ability if needed

Thank you for reading, bless you for helping

 

[1503-44]

By "passing" I guess you mean "leaking across", right?
That's a very large pressure drop. I think you are probably destroying doing a good job at destroying the internals. It is a very corrosive service too. Worse with high temperature? You could try with a special high pressure drop "whisper" anti-cavitation type trim, you may also do better with two high pressure trimmed valves in series. Advise the manufacturer of all conditions when procuring the new valves.

Reducing the recycle flow might only increase the discharge pressure, resulting in being no better off if you continue to recycle even at lesser flow.
Which brings up the question... Why are you recycling?


--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[LittleInch]

A flow sketch would be useful to make sure we all understand but from what I can figure out you have a "condensate" line at about 90 bar which either pumps into the gas line at relatively low flow (12m3 (58- 2x23)) or returns this condensate into some process units at 20 bar.

Mokveld make good valves so I would first ask them what they think, but I think you are flashing across the valve and basically eating the inside from cavitation. I think you probably need something more akin to a choke valve / plug and cage with trim suitable for sour service and flashing service.

You need to tell the control valve suppliers all the information to let them chose the right valve.

When you say "passing", what do you mean. A control valve is intended to flow, not seal or isolate. If you want isolation use an isolation valve like a ball or gate valve. If you want control use a control valve.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Yumie]

Yeah its leaking across if thats the term, fluid flows through at fully close pos. Temperature operates at 40°c, not too high and we have HiHi alarm at 50°c
We are recycling a lot due to pump high capacity (58m³/hr) compared to condensate produced from wells (12m³/hr).
I'm not too sure why they chose this pump, I've been digging old records and we never produced much condensate to start with.
One reason I could think is for during pigging activity from another platform that pushes settling condensate in pipe to us, and so that we could evacuate it to shore immediately

Please look at diagram below, this is the current operating setup. I do plan to suggest we swap the high intake from the bigger vessel slugcatcher (SC). This will also reduce the need to recycle to TPS while increase recycle to SC. Ah that LCV valve is also leaking across thus me going to site and manually open/close it when SC is high or low level

Something I drew for my study


I agree that a control valve should be used to control and never expect a seal off, I've read somewhere optimally it should be controlling around 40-80% opening of the time.
On some crappy days our separator just go up and down the High and Low Level alarm, the LCV would operate at 0-100%.


Oh I also didn't mention that we "choked" the downstream LCV's Isolation ball valve to about 15-20° opening, this practice started after the passing issue was reported, almost a year now *facepalm*. I'm the guy that adjust the ball valve every time the panel asks for it. This choked operation also gave me an idea on using orifice because in theory I'm already reducing the ­▲P after LCV with current operation?

Thanks for reply good sirs





Not an Engineer

 

[georgeverghese]

Usually, the condensate flow ought to go to the TPS before going to the pump. That way, sand and other erosive solids will settle in the TPS. The current arrangement where SC condensate goes directly to the the triplex pump exposes both pump and recycle LCV to erosive sand. In any case, the control valve trim ought to be made of hard wearing tungsten carbide trim or similar. Also install a tungsten carbide trim fixed choke downstream of the LCV to take the bulk of the 70bar dp.
With this current arrangement, suspect you must be scouring the liner on the triplex pump very quickly also, especially during feeder pipeline pigging operations.

 

[1503-44]

I think that's probably 6m3/h going to shore.

Which valve on the diagram are we talking about?
So that pipeline only shows condensate. Where is the gas flow?
Why are you separating only to reinject into a combined gas/condensate pipeline?
Are you separating then metering gas and condensate then reinjecting those back into a combined flow pipeline? Are you removing sand? What are you separating at the tilted plates? Nothing? Sand?
I think the slug catcher should go to the tilt plates, then the pump. George seems to agree.


Not sure, but it seems like you should run the pump 6 min at 58m3/h every hour, then shut it off. That way you can avoid recycling. Can your slug catcher hold 6m3?



--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[LittleInch]

Appreciate you're not the engineer here, but what you're trying to do is treat the symptoms not the disease.

So options to consider

1) Buy a second smaller pump and use that most of the time then if the level in the SC rises turn the big one one
2) Can you retrofit a Variable frequency unit to your big pump? This is a PD pump so speed is proportional to flow.
3) Yes you can install some orifices or even multi orifice plates but I would put them downstream of the CV as I think you are flashing the condensate in this pressure drop and its better to do that d/s the valve
4) Buy a choke valve / plug and cage control valve not the axial flow version
5) Ask someone on your process design dept to look at this and come up with options as there are many things which don't make sense.
6) Replace that ball valve on the route back into the SC with a control valve or globe valve at the very least. All you're doing is destroying your ball valve and at low percent opening they are very difficult to control, plus it is a fixed flow.
7) Look at Mr 44's suggestion and use your SC as a holding vessel and operate the pump on and off every 15 or 20 minutes?
8) What are those separators separating? Is there more flow coming in as otherwise your flow numbers don't add up.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Yumie]

In the old days they used to flow the condensate into TPS before the pump, but today operates bypasses that. I'm not too sure why, I've only been here for 2 years.
I like the idea of a choke after the LCV, I may suggest to construction to modify this if its an option


1.We have 3 recycle valve in total, they're all passing just fyi
LCV 20x0 returns to TPS, one per train
LCV 2010A/B (didn't draw) return to Slugcatcher

2. Sorry didn't think it was relevant to show main gas line, the condensate merge with gas line right before our launcher barrel, sent to shore with the same pipe.
Our train gas header from wells are 20bar of pressure, thus they went into 2 stage turbine compressor to ramp up to 50bar and then 90 bar pressure before being exported. Gas from SC also shares the same path.
For condensate it is 12m³/h net to shore, 6m³ is being produced by our well, and another 6m³ came from slugcatcher from another platform

3&4. Yes we have separate gas and condensate meter, most likely to calculate the flow separately

5. Nope. The name of vessel is indeed "Tilted plate separator", but PNID shows it has plate pack coalescer in it. I think its two different things? I haven't seen much. We don't have sand management system nor we have much trouble of it, yet.

6. TPS functions to separate water/condensate, the water goes to its own package before being discharge to sea

I liked the idea of slugcatcher going to TPS. In original PNID and years before I came, there exist (still does) a freeflow line which goes to TPS first but today it went straight to the pump. I'm not too sure of the reasoning behind this, some say that line is now fully clogged due to years not in service and inoperable

7. I can't figure out the good way to run/stop pump to maintain level, it may work but since the pump intake is from 3 different vessel simultaneously, with 2 of them are smaller. The problem with on/off is what if one separator is 20% and other is 60%, there is no automated valve to control intake% thus the pump needed to continuously run.


Might as well share my full idea here since I'm pitching this tomorrow when I fly to work

1/ The first is to eliminate premature lifespan of the LCV return to TPS, if the LCV works fine I wouldn't need to intervene as often. Thus the idea to reduce ­▲P across the LCV with orifice, or fixed choked as George said.

2/ The second is to change flow management by swapping high intake from TPS, into taking more liquid from SC instead. This will reduce mass flow return to TPS, also able to control the level of separator above it ¿better?, (small mass out, small mass in). The separator (not shown in diagram, only has about 6m³ between Low and High alarm so it's quite hard to maintain PV at SP, (the water level in TPS also directly affect level in separator (gas/liquid) above, especially if the TPS PV went flying or drop). This will also improve TPS separation since retention time is longer? Please tell me

3/ The third involve a long time problem which is high vibration on multiple points on the whole damn platform (cause of the pump). I'm proposing to run the pump at 2 different modes;
-The SC has huge volume of 90m³ between 20% to 40% level. I will take full advantage of this.
-The SC will become the buffer vessel to fill up condensate produced from our wells and another platform.
-With unload mode(as I called it), I would 100% recycle all condensate into SC, it will takes about 7hour~ to fill up from 20% to 40%, this also effectively reduce discharge pressure to push fluid into SC maybe about 30-40 bar discharge
-At SC 40%, I'll load the pump by fully shut recycle to SC (80% flow to export, 20% flow to maintain separator level) for about 1 or 2 hours. Closing the Recycle meant the discharge pressure will now build into MainGasLine pressure of 90bar
-This will effectively reduce vibration uptime (but not amplitude) by 70-80%

Ultimately, of course there's a better way to handle this situation, like changing the pump into a smaller capacity, perhaps something thats not piston like, vibration inducing, maybe a screw pump capable of generating 20m³/hr with 70-80bar DP. But that's gonna cost a lot, and nobody want to hear this especially coming from me


Not an Engineer

 

[Yumie]

Thanks for reply. I was writing the reply above when you replied. I will read now and edit this post after.

Edited:
1.

Ultimately, of course there's a better way to handle this situation, like changing the pump into a smaller capacity, perhaps something thats not piston like, vibration inducing, maybe a screw pump capable of generating 20m³/hr with 70-80bar DP. But that's gonna cost a lot, and nobody want to hear this especially coming from me

You are true, this has always been an option but somehow its the very last thing they'll do...

2. I'm not sure what is a Variable frequency unit is, I'll do some research and come back to it

3. I see, an orifice d/s of CV make sense, I have been contemplating between an orifice u/s or d/s of CV but I couldn't brain it. But I would adjust my suggestion accordingly.

4. I'll put that in my study list

5. They probably did, but maybe it wasn't the highest priority thing in their list.

6. You are correct, George suggested above to install fixed choke instead of valve down there. I have placed it in my study

7. Yes I will suggest to use SC as a holding vessel like my point 3/ above. I think it's brilliant. But the pump can't operate on ON/OFF due to reason like my point 7. above

The problem with on/off is what if one separator is 20% and other is 60%, there is no automated valve to control intake% thus the pump needed to continuously run.

8. There is a 2 phase, vertical separator on top of TPS. It takes gas from the well to the turbine compressor, while the liquid byproduct settles down and work its way into TPS below.
The TPS itself is a 2 phase. horizontal separator that separates water and condensate

 

[1503-44]

A new diagram would help.

I think the seperator flows have a solution, but not sure without the PID and size details.
Probably No valve is needed. You don't have one now. When pump is stopped, condensate fill rate is faster. When any one seperator is at high level start the pump and pump from all seperators until one reaches low level, then stop pump. That is usually possible, manual with high and low level switch & alarms, adding automation only requires auto start/stop of the pump by the level switch signals.

There are 4 main options I see

1) Replace valves every 6 months, see if two valves might handle the pd better. Or try some orifice plate combined with one valve. Its still a lot of pd for an orifice plate. Whatever you do there will probably be maintenance intensive. A constant 70- 100 bar drop is just a bit difficult no matter what you do. Time doing that should be limited, like in a startup/shutdown activity. Running the pump start/stop could minimize that time and extend valve plate life.
2) Run on/off Many seperators can run in that mode, but depends on the vessel size and LL control.
3) Run existing pump with a variable speed device a VSD, or also called a VFD. May require new motor and lots of wiring. Not great.
4) Install a new smaller capacity pump that does not need to recycle. Constant speed probably OK.



--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[georgeverghese]

The Canadian choke valve company Masterflo make good heavy duty manual-fixed and auto operated chokes for erosive service - ExxonMobil buy a lot of these.

Looks like your company isnt too fussed about exporting dry fluids with the intent of minimising corrosion in export pipeline - there is no gas dehydration and you also export some condensate direct from SC to pipeline without dewatering at TPC / TPS. Your company must be spending a lot on pipeline corrosion inhibitor injection chemical.

 

[1503-44]

I think it's just a well head flowline on a remote production platform, so it only does basic gas/liquid separation for well metering and liquid knock out purposes before compression. I never see dehy equipment offshore and before a collection of well flow lines arrive at a central GT/NGL facility, just before going into a dry gas transmission pipeline. They might add a small corrosion inhibitor injection package, but there are many that do not even do that.

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."

 

[pierreick]

Hi,
You may want to talk to these guys:

https://kentintrol.com/severe-service-valves/erosive/

Pierre