How to know if a PSV has opened?

[edanyel]

I'm sorry if this is an obvious question.
Anyway it bothers me, and i'd like to hear opinions.
If the vent pipe goes to a flare, if in the test stand the PSV opened at the correct set press and reclosed at the correct blowdown, if there has been no leakage after reclosing, if the line protected has no Pressure registering instrument, could anybody say if a PSV has opened?
Thank you
Daniel

 

[robsalv]

tricky.

If there's tell tale residue on the discharge side then you could surmise it did open.

If the protected pressure vessel has a computer trend which shows that the pressure rose to the set point and the later test indicated the valve performed correctly, then yes, you could surmise it did open.

If an operator heard the valve lift, then yes.

If there are no clues, there's no way to tell.


Crosby have a tell tail device that you install on the discharge side which indicates that a PRV has lifted, for exactly this reason.




Here's a curly one for you, pressure vessel with ethylene, pressure trends went above set point but no lift was heard or observed. This is a safety incident. SV is removed and tested on the bench and it popped perfectly but had a low bubble pressure (65% of set, operating pressure was 90% of set). Pressure trend/transmitter is properly calibrated.

Why didn't the PRV lift in the field?

- - - - - - - - - - - - - - -
"Life! No one get's out of it alive."
"The trick is to grow up without growing old..."

 

[JAlton]

If the pressure transient is very rapid up and down, sometimes the PRV cannot react quickly enough to respond. Its weird, but I have seen it happen, rather not happen.

JAC

 

[robsalv]

JAlton, this wasn't a pressure spike scenario, but that's an interesting observation. Cheers mate.

- - - - - - - - - - - - - - -
"Life! No one get's out of it alive."
"The trick is to grow up without growing old..."

 

[JAlton]

Were all instrumentation calibrations verified?

JAC

 

[robsalv]

Yes they were.



- - - - - - - - - - - - - - -
"Life! No one get's out of it alive."
"The trick is to grow up without growing old..."

 

[JoeWong88]

Spring loaded type PRV is understood as "slow-acting" pressure relief device compare to bursting disc or buckling pin. It may not lift when it see only pressure spike especially in low compressibility liquid system. For instance, HP gas in tube side and cooling water in shell for a Shell-&-Tube Heat exchanger, if the overpressure scenario is credible, a bursting disc is preferred compare to Spring loaded PRV. Having said that there is still some testing / field experiences show that even bursting disc is "not" fast enough. Read more.

Pressure spike in gas system is less likely. However, extremely quick reaction could lead to pressure spike. This tell a point : Why bursting disc is common in reactor.

JoeWong
Chemical & Process Technology

 

[MikeClay]

Assuming the relief valve lifts at set point, then there needs to be a blockage somewhere in the relief path. If in the inlet piping or relief valve, it would probably have been noticed when the relief valve was removed.

I also assume that any block valve in the path for maintainance is locked/sealed in the open postion

With a simmering relief valve, ethylene will normally be flowing into the discharge piping at a low rate. Since ethylene is a monomer, it may have polymerized. Once polymerization has started, it will continue until the line is blocked.

This would be my first guess.

--Mike--

 

[edanyel]

Robsalv my question wasn't tricky I've been really asked about that matter and my answer was no, I have no way of telling if the valve did open or not.

Thanks for your comments.

I'll give you my silly guess about yours....was it partially open all the time?

Rgds
Daniel

 

[Mecha20]

very interesting...i had a scenario in a well known company where there PSVs were actually set higher than the pressure rating of the pipeline. luckily it was fixed before anyone got hurt....

improper re-seating can also be a guide to know if it had lifted...

 

[robsalv]

Mike, I've never seen ethylene polymerise in our transport piping systems in the way you've described. Ethylene gas service is typically a very clean clean service.

Ok, in my example, the pressure rise was not a spike. It occurred relatively slowly and was above the set pressure for minutes. The Operators had high pressure alarm indication in the control room and were dealing with the overpressure for minutes. One can also assume that the discharge was not blocked - I should have added this was a decent sized PRV, 8 x 6.



Daniel, I agree, once a relieving event is over, there's often no way to tell whether a PRV has lifted if no one was there to notice... a bit like that proverbial tree falling in the forest. There are two other possible indicators though to add to my list: a) If the SV fails open/fails to reseat* OR b) if the SV chatters itself into oblivion... either are pretty strong clues that an SV has lifted


(*an SV will fail to reseat if the operating/system pressure remains ABOVE the blowdown pressure. Factory blowdown settings tend to be between 7 - 12% of set pressure, so it is conceivable that a system running close to it's design pressure could actually have PRV's that wont reseat once lifted.)


Ok, I'll give my best guess as to what I believe happened with this ethylene PRV. The investigation could not identify any mechanical failure or other factor that interfered with PRV operation. The incident had me scratching my head for a while.

The clue is the online leakage. I believe that the gas auto refrigerated due to the pressure drop across the seat (down to flare header pressure). This effectively froze the disc to the nozzle or caused some other cold related interference in the moving parts.

When the PRV was being prepared for removal and then transported to the workshop, the PRV had plenty of time to warm up to atmospheric conditions removing all traces of the "evidence". The PRV operated flawlessly on the test bench - except for the leakage.

So in this case, we had clear indication of an overpressure incident and when the operators were out in the plant trying to manage the situation, they clearly noted that the PRV had not lifted/was not relieving and so raised a safety incident.



Mecha20 - how much higher are we talking?


- - - - - - - - - - - - - - -
"Life! No one get's out of it alive."
"The trick is to grow up without growing old..."

 

[insult2injury]

The failure to open could also have been due to pressure buildup above the disc.

I2I

 

[robsalv]

I2I that's very true - but there were no other pressure sources into the header at the time.

- - - - - - - - - - - - - - -
"Life! No one get's out of it alive."
"The trick is to grow up without growing old..."

 

[tickle]

There are several ways of telling if there has been flow past a PSV.

1. rupture disk in series
2. flow switch - can be a differential temperature switch/transmitter for gases
3. canvas caps over the discharge pipe (pops off if there is relief)
4.

Obviously, not all of these can be used if you are flowing into a vent / flare system