Relief Valve lifting low in field, lifts normally on test bench

[Legoinsaniac]

Hi all,

We have an ongoing problem with a relief valve. On the discharge side of a pump (water, ambient temps), we have a relief valve that has repeatedly lifted prematurely. When removed and bench tested, this relief valve will then lift within 1% of the 235psig setpoint. Upon reinstallation in the system, it again lifts early. This indicates a problem with the system - not the relief valve.

We've thoroughly reviewed the calculations for this piping system and components, and we know from local pressure gauges that pressure does not approach 235 psig. Are there any suggestions as to what could be causing this? Could local pressure be higher than system pressure (i.e. a standing wave in the relief valve inlet piping)?

Thanks for the help!

 

[TD2K]

Does the PSV just relieve to grade? If it relieves to say the suction could there be any unbalanced piping forces on the PSV in the field?

 

[Artisi]

When does it lift, on start-up, shut-down or during continuous operation.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[Legoinsaniac]

The PSV discharges to a sump - atmospheric pressure.

It lifts during continuous operation.

Additionally: Some people here questioned the effect of nozzle loading on PSV set points. But we've checked fit-up by unbolting the flanges and the piping aligns nicely and bolts together smoothly.

Also, we've installed a different PSV (same make/model) into this location, and it had the same failure (but it too passed the bench test)!

The operating pressure of this system is around 160 psig. Local, calibrated gauges read this pressure when the PSV is lifting.

This problem really has me thinking!

 

[katmar]

The only time I have experienced a similar phenomenon was with unexpected 2-phase flow. Slugging from vapor ingression caused pressure swings that were so rapid we did not see them on the gauges, but the relief valves leaked continuously. Ensuring liquid-only flow solved it completely.

I suppose a reciprocating pump could cause a similar effect - where the average pressure is well below the set point but very short lived spikes could lift the valves.

The million dollar question is "why put a relief valve there in the first place?". What are you trying to protect against on a pump discharging to atmosphere? Perhaps this will give us a clue.

Katmar Software - Engineering & Risk Analysis Software

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[Latexman]

A sketch, P&ID and/or picture of the installation may spark some ideas.

Was the pressure gauge on the test bench verified? Probably was.

Is the PSV installation on the vertical-up, center branch run of a Tee so it sees a good static pressure?

Good luck,
Latexman

 

[BigInch]

Maybe he's protecting the discharge piping.

Type of pump, type of PSV distance between PSV and pump disch, velocity off the pump.

Give us a picture.

What would you be doing, if you knew that you could not fail?

 

[zdas04]

When I had a similar "impossible" situation, I found that an old fashioned pressure chart recorder identified the problem pretty quickly. For me, it was a short duration spike above the setpoint. Putting a pulsation damper on the pump discharge got rid of the impossible situation. I never did figure out what caused the spike, but it was enough to know that it was gone.

David

 

[Latexman]

The reseating pressure of a PSV in liquid service is *about* 70% of opening pressure. In your case that is very, very close to your normal operating pressure. 235 x 0.7 = 164.5 psig. "The operating pressure of this system is around 160 psig.

So, if any pressure spike opens the PSV, it can't close (reseat).

Look for pressure spikes. Is it a PD pump? What kind of pump?



Good luck,
Latexman

 

[Latexman]

A downstream valve that closes too fast may be the culprit.

Good luck,
Latexman

 

[Duwe6]

Yes, your PSV is showing you something that your gauge(s) aren't. Since the PSV is properly calibrated, take a hard look at the gauge(s) calibration and response time. Are/is the gauge(s) on a 'snubber' to even out pressure spikes?

 

[Legoinsaniac]

Thanks everyone for the helpful comments.

I'll be looking more closely into the possibility of a pressure spike from the pump, air entrapment, and valve closure.

The pump is centrifugal. 10" discharge pipe, 3000gpm, 12ft/s. There is a downstream check valve, and then a vertical tap off for this relief valve.

 

[BigInch]

12 fps (over 10) indicates a possibility for waterhammer spikes, although I am not saying that it is what is causing this.

What would you be doing, if you knew that you could not fail?

 

[zdas04]

Is the PSV upstream of the discharge check? I've seen slamming discharge checks spike pressure more than 100 psi upstream of the check.

David