Testing valves for leaks while operational

[mysticpage]

Does anyone know of anyway that I can test a valve (remote operated ball valve, normally closed) for leaks without removing it? Is there a way to acoustically test it (i.e. with a stethescope)? The fluid in the process is of a higher viscosity (dowtherm). Any help greatly appreciated!

 

[agthompson]

There are relatively inexpensive ultrasonic leak detectors that work well on gases or steam. I am not sure about the effectiveness on a viscous liquid--but try talking to some of the manufacturers of this equipment. They are effective even on very small leaks in noisy environments.

Also if the product is above or below ambient temperature you might be able to get an indication of leakage with an infared scan.

Finally you can hot tap a small valve to the side of the piping--and know without a doubt.

 

[mysticpage]

Thank you! I am looking into the ultrasonic equipment now. I talked to a vendor of the detector and he said that even the viscous fluid should be detectable. He said that it should have a distinct gurgle or popping noise. Have you ever used a portable ultrasonic detector before? Think it's something the plant operators can handle or do I need a technician?

 

[unclesyd]

If you want to keep the meter intact use a technician, my experience.

 

[agthompson]

mysticpage:
The leak detectors are extremely simple to use. They do require some experience in determining what to listen for though.

 

[Cockroach]

This typically depends on the design of the valve, my personal prejudice is trunnion mounted, double block and bleed.

In this case, with the valve closed, you can pull the body plug and monitor the pressure cavity with a gauge. If the seat is shot, obviously a rise in pressure occurs.

If your valve is a floater, you don't have this luxury. You might be able to locate a leak via audio means, but this may not be an indication of any problem other than bypass. (i.e. is the seat leaking or is a seat gasket undergoing bypass?)

Monitor your torque values with time. Obviously with leaks you have both seat wear and an accumulation of pressure somewhere not intended. You should see a spiked rise in torsional values with time. I typically record factory values during pressure tests (API 6A and CSA Z245.15) so that a baseline can be established for comparative strategies down the road.

Good luck with it. I think you have a difficult problem at best.

Kenneth J Hueston, PEng
Principal
Sturni-Hueston Engineering Inc
Edmonton, Alberta Canada