16 replies to this topic

[nabeelsp1036]

Hi,

 

I'm in a process of sizing a PSV located on condensate header of a gas compression plant. 

 

Refer to the attachment to get a clear picture of the process.

 

 

Questions:

 

1) What scenario is applicable to the PSV? Is it gas blow-by case or LCV failure case?

 

2) If it is gas blow-by case then do we need to consider the inlet separator as high pressure source? Operating pressure = 400psig. Design pressure = 1200psig. Condensate header PSV set pressure = 655psig (defined by client).

 

what is the definition of high pressure source? Does it corresponds to the 'design pressure'? or is it the 'operating pressure'?

 

3) Do we need to size PSV for liquid relief and gas relief separately? As per my understanding, in case of gas blow-by the liquid shall exit through PSV first and then the gas. So we need to size PSV for both and then select the worst one.

 

Experienced thoughts and guidance is required.

Attached Files

•  Xerox Scan_22062023134208.pdf   635.76KB   64 downloads

[Pilesar]

It looks like you are trying to protect piping from overpressure. Why? What is the source of overpressure? When you discover that source of overpressure, then you will find the high pressure source.

[snickster]

I would think you would have to look at both cases and whatever cases may be present.  In any case there will be all liquid, all gas or maybe liquid and gas.

 

I assume blowby means that the LCV is still operating properly but LL shutoff level controls on the scrubber fails that so gas eventually exits the scrubber and into the piping.  So in this case the LCV is still partially closed as when in the latest control mode, or is the LCV now uncontolled and has the possibility of allowing full liquid flow of the liquid still in the piping, then followed by gas?  In any case I think in this scenario, if the LCV did not fail and is still partially closed then this would be the scenario with the lowest flow due to the pressure drop across the control valve in the partiall closed position.

 

If the LCV fails fully open then you could also have liquid first and then gas if the SDV fails too at same time ( or LL shutdown fails to signal the SDV).  This would appear to be the worst case since the LCV may be fully open, and SDV fully open too.  In this case there will be first liquid and then followed by gas.

 

This is the way I see it but all posible scenarios must be checked and possibly a HAZOP performed.  For the gas flow in the higher pressure sections of the drain piping the flow will reach sonic choked flow in the orifice of the LCV's, which is the basis for determining the flow.

Edited by snickster, 23 June 2023 - 07:41 PM.

[nabeelsp1036]

It looks like you are trying to protect piping from overpressure. Why? What is the source of overpressure? When you discover that source of overpressure, then you will find the high pressure source.

@Pilesar

 

Sources of over pressure are shown in the attachment. There are multiple ones. But of course PSV shall be sized for the maximum flow which needs to be checked from every high pressure source.

And yes, PSV on condensate header is for overpressure protection. I'm sizing it for Gas Blow-by case. 

[nabeelsp1036]

I would think you would have to look at both cases and whatever cases may be present.  In any case there will be all liquid, all gas or maybe liquid and gas.

 

I assume blowby means that the LCV is still operating properly but LL shutoff level controls on the scrubber fails that so gas eventually exits the scrubber and into the piping.  So in this case the LCV is still partially closed as when in the latest control mode, or is the LCV now uncontolled and has the possibility of allowing full liquid flow of the liquid still in the piping, then followed by gas?  In any case I think in this scenario, if the LCV did not fail and is still partially closed then this would be the scenario with the lowest flow due to the pressure drop across the control valve in the partiall closed position.

 

If the LCV fails fully open then you could also have liquid first and then gas if the SDV fails too at same time ( or LL shutdown fails to signal the SDV).  This would appear to be the worst case since the LCV may be fully open, and SDV fully open too.  In this case there will be first liquid and then followed by gas.

 

This is the way I see it but all posible scenarios must be checked and possibly a HAZOP performed.  For the gas flow in the higher pressure sections of the drain piping the flow will reach sonic choked flow in the orifice of the LCV's, which is the basis for determining the flow.

I believe PSVs are the ultimate protections. I mean we assume that all the electronic protection are failed (SDV or LITs etc) and then we examine the overpressure scenarios for PSV sizing.

Therefore, for this case we shall not take credit of any auto protections. And the case shall be simple that our LCV are failed in 100% open position and now we need to size our PSV for condensate header. For worst case, i am using the pressure difference as 1200psig-655psig = 545psig to calculate the flow of gas that shall pass through the LCV at 100% open condition. PSV shall be sized for this flow. This example case is for Inlet Separator LCV. 

Am i good in my understanding?

[Pilesar]

By your description, the inlet separator operates at 400 psig. LCV failed open would not overpressure the 650 psig header. I don't see the source of overpressure.

[Bobby Strain]

And there is the condensate from both compressor discharge separators. For level control valve fail open, it will first pass liquid, then liquid and gas.

 

Bobby

[snickster]

I agree with your post and not count on controls to mitigate issues with LCV failure.  I would imagine that there could be situations where failuire of LL level controls could happen simultaneously with failure of LCV.  I woiuld design for LCV failure fully open.

[snickster]

Also you have to consider that any alarm or shutdown controls might have already failed before the LCV fails open but nobody knows.

[breizh]

Hi,

As suggested earlier, consider performing an Hazop study and/or reviewing the existing one.

Good luck,

Breizh 

[nabeelsp1036]

By your description, the inlet separator operates at 400 psig. LCV failed open would not overpressure the 650 psig header. I don't see the source of overpressure.

So you mean sources of overpressure are the ones with "operating pressure" higher than the d/s system PSV set pressure? I don't think so. 

 

Looking at the design pressure of Inlet Separator (1200psig) is much higher than the PSV set pressure (655psig). Therefore, i believe it to be a source for overpressure.

 

Sorry for my typo. I'm checking for LCV failure scenario. Not for Gas blow-by scenario. I believe these two are different scenarios. 

 

LCV Failure: u/s system design pressure - d/s system design pressure

Gas Blow-by: u/s system operating pressure (max) - d/s system design pressure

Edited by nabeelsp1036, 26 June 2023 - 04:32 AM.

[nabeelsp1036]

And there is the condensate from both compressor discharge separators. For level control valve fail open, it will first pass liquid, then liquid and gas.

 

Bobby

This was what i was looking for. But why followed by liquid+gas? Shouldn't it be gas only?

 

Actually, I am to provide relieving flow to vendor for sizing. How should i give the mixture flow (liquid+gas)? I will be providing liquid relieving flow and gas relieving flow separately to vendor.

 

Also please guide, what pressure difference conditions of LCV failure should i consider to calculate relieving flow? Should it be

1) u/s system design pressure - d/s PSV set pressure ?

or

2) u/s system PAH - d/s PSV set pressure (or PSV relieving pressure)?

 

our company standardizes it as option-1. But i don't agree with this, since i believe this as very much conservative approach. I mean, we are assuming LCV fail open with u/s system reaching at design pressure at the same time.

 

What's your thoughts?

[Bobby Strain]

Sound like you need to engage the services of one who is experienced with this challenge.

 

Bobby

[Pilesar]

The design pressure of the inlet separator is not the source pressure and should not affect the condensate header overpressure scenario. The inlet separator pressure is controlled to 400 psig. If the LCV of the inlet separator failed open, the condensate header would see 400 psig pressure. If the inlet separator pressure control failed, the inlet separator would see higher pressure but only to the maximum of the source pressure to the inlet separator and the LCV would still be in control so the condensate header would not overpressure. Your scenario involves several independent failures simultaneously: Fail open of the inlet separator pressure control valve, fail open of the inlet separator level control valve, isolating the condensate header from downstream equipment. The condensate header does not need a PSV in my opinion. What is the rated pressure of the equipment downstream of the condensate header? Is your new 655 psig PSV going to protect that equipment also or will you add another PSV to that equipment sized for your scenario? 

[Bobby Strain]

Pilesar, how about the other higher pressure sources? You seem to have followed the red herring.

 

Bobby

[Pilesar]

There are four pressure vessels shown feeding the condensate header with liquid through level control valves. If one of the level control valves fail open, the condensate header might be subject to the vessel operating pressure. The first and second vessel operate at 400 psig. The fourth vessel operates at 15 psig. The third vessel is an interstage knockout drum for a reciprocating compressor with no operating pressure given. That third drum is the only reasonable high pressure source for the condensate header. Assume that operating pressure of the third drum is 800 psig based on where it is placed in the drawing. So if third drum LCV fails open, at worst there might be pressure source of 800 psig to piping that is rated at 655 psig. B31.3 piping test pressure is 1.5 times design pressure (i.e. 892 psig.) The condensate header normally operates less than 15 psig (since it gets liquid from the 15 psig fourth drum.) Each of these source connections to the header show a SDV in addition to the LCV. Is that SDV part of an instrumented protective system triggered by low liquid level? The only way to build pressure in the condensate header is to close the header SDV downstream on the export! The condensate header SDV is normally open. What code or standard requires this condensate header piping to be protected from overpressure? The equipment downstream of the header should be analyzed for overpressure sources and protected as appropriate. The MAWP of the upstream vessels have no bearing on protecting downstream equipment. My analysis is based solely on the original sketch (very nicely done!) and should be confirmed using real P&IDs. There might be a case for a small relief valve for liquid expansion if the condensate header were blocked in, but that is not proven by the information supplied.

[Bobby Strain]

I must be guilty of following the wrong piping. But it seems strange that the vessel shown to operate at 15 psig should drain into a high pressure system. Unless 15 psig is really 1500 psig.

 

Bobby