Thermal PSV - Lifting/Popping

[Nova N]

A Thermal PSV on a 4" steam condensate line lifts/pops when the condensate pump starts up. Then PSV reseats. The pump is 400kPa centrifugal pump with 40m head press max.
The PSV specs:
1. Conventional type
2. Set pressure is 1035kPa
3. Operating temp is 100 C and relieving temp is 171 C

According to the plant operator the PSV was good for many years and started lifting recently.

We removed the old PSV and replaced it with a new identical PSV. But the lifting/popping continues.

The steam condensate temperature is between 90 - 95 C. Also the condensate discharges into a cylindrical tank vented to atmosphere.

Anyone had a similar problem or can offer possible reasons for the problem.

Note: There has been no changes to this process recently.

Thanks
Nova

 

[mk3223]

If not the issue of the new certified PSV valve, IMO, it could be the overpressure issue related to the condensate system, such as low/no flow condition by the blocked outlet.
Was the actual system field verified for no process or operation changed?

 

[Nova N]

mk3223, yes field verification of P&ID was done by OPS and I also personally walked the length of the 4" line.

 

[Latexman]

Has the pump's starter characteristics changed recently?

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[Nova N]

Latexman, I did a few checks on the pump in the field and trends and its still pumping according to pump curve, so there's been no changes.

 

[don1980]

It's pretty clear what is happening. When the pump is started, and the discharge line is liquid-full (no vapor pockets), you get a pressure surge. And that surge force can be enough to lift the PSV. So why did this not happen in the past? A likely explanation is that there used to be a vapor pocket somewhere in the outlet system, and now it's gone (an incompressible system becomes compressible when there's a vapor pocket).

There are multiple ways to solve this problem. One is to dampen the pressure surge by installing a pulsation dampener on the pipe, close to the PSV. Another solution is to slow the staring of the pump. That can be done with a VFD, and I think there's other electrical gear that'll do this too.

 

[Nova N]

Thanks for the detailed response. We did discuss this as a possible solution, but could not explain the change. I am hopeful this will be the right solution. Thanks again and much appreciated.

Novan

 

[mk3223]

One more to check is the process source of the pump suction to verify for any process change.

 

[Nova N]

Thank you. I have already verified the temperature of the condensate which is between 90-95 deg C. I will take a closer look to see if there are any other changes.

 

[georgeverghese]

Can we see the P&I D for the source, pump and discharge vessels and related process controls?
Thermal PSV? How could that be when you have condensate at 90-100degC ? A thermal expansion PSV accounts for volumetric expansion due to ambient heating.

 

[Nova N]

Georgeverhese, although I cannot share the P&ID due to policy, I will try to describe the system better.
1. A condensate receiver tank (1m dia x 2.5m high), followed by 4" hand valve, followed by 400kPa centrifugal pump.
2. 4" check valve, followed by hand valve 4"
3. a 2" flow nozzle (and Flowmeter), followed by 4" ball valve & then 3/4 PSV (conventional).
4. Another 4" ball valve, then into the 170m3 cylindrical tank (6m dia). Entry to tank is at 4" inlet on the side (
close to the top).

The PSV is for the blocked-in condition and the lines are all heat traced (part electric, and part steam traced).

Let me know if you need more details.

 

[mk3223]

It looks that this is a simple condensate pumping system with a single suction source piping to the discharge and then the final storage tank. The issue is that the system is over pressurized and operated above the design condition.
Here are my guessing (one last time?):
- Check the discharge pressure gauge to see what's the actual line pressure while the PSV popped
- if there is other condensate line connected to this discharge line downstream the PSV?
- Are the discharge flows normal as indicated at the flow meter during normal condition and as PSV popped?
- Was the heat tracing set correctly? (What's it for?)
- 3/4" PSV is typical designed for the thermal relief. Can the set pressure can be safely increased to the max pressure of the piping class, say PN20?
- Are the temperature and liquid level at condensate storage tank normal?

Understand that the operation and you might go through it many times trying to solve this mystery. Hope to help a little here. Or, may consider to hire a outside person with a fresh eyes to check for you.

Good luck, and share with us what's the root cause.

 

[Nova N]

Thanks Again, I will start afresh and do another process audit. Will post any further findings and deviations.

1. The heat tracing is for the -40 degrees temp that can occur in winter and piping runs outside.
2. The line is designed for 1035kPa max, which is same as PSV setting.
3. Discharge tank level and temp is normal, but will confirm source tank and flowmeter.

Thanks Nova

 

[georgeverghese]

Tell us a bit more:
a)What is the (max?) op pressure in this receiver at the time of startup, when this PSV lifts ?
b)Static head between source vessel and pump?
c)Dead head differential of this pump?

How could steam based heat tracing work reliably for ambient temp at -40degC? Suspect one or more of these steam based tracing loops is not working as it should, resulting in partial freezing in this exit line and pump running dead head.

 

[Nova N]

The max operating pressure in the receiver is 1 atm - vented to atm.
The static head is 12.25kPa between source vessel and pump.
The dead head differential of the pump is 380kPa.

Thanks for the tip on the steam tracing. We will check the loops out to make sure all is working.

 

[LittleInch]

Difficult to diagnose remotely but also consider that at no flow the trace or steam heating is creating a small vapour pocket which collapses when the pump starts. The pressure shock wave from the liquid columns hitting each other can be 2 to 4 times the normal operating pressure.


Or the nrv is passing and allowing back flow which creates a small vacuum pocket at a high point.

Try changing the nrv or try isolating the pump using the hand valves immediately after pump stop.


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

 

[Nova N]

These scenarios sound promising towards resolving the issue. My first step will be to get Maintenance to check the NRV, as it could be passing. The steam/heat tracing also could be a contributing factor to the pressure shock.

Thanks Again.

 

[georgeverghese]

So the max developed discharge pressure is approx 420kpag, which is miles away from PSV SP of 1035kpag.

Would then suggest that what @don1980 and @LI suspect may be possible. One or more of the heat tracing loops may be overheating, resulting in liquid hammer as the local vapor pocket implodes when flow resumes. It is not difficult to set up a over temperature trip loop on an electric heat tracing loop, but I wouldnt know how to enable a simple high temp trip for a steam tracing loop.

 

[Nova N]

Thanks for the recommendations. I will be done a final process audit of the entire line and thereafter will action the suggested solutions.

Thanks Again to All.