Liquid volume loss during pressure test

[MikeJ]

Hi, I'm doing a pressure test with a liquid medium. I have a fixed volume and I know the pressure drop, and temperature change - how do I calculate volume of liquid lost in a leak?

 

[sophiecentaur]

Hi, I'm doing a pressure test with a liquid medium. I have a fixed volume and I know the pressure drop, and temperature change - how do I calculate volume of liquid lost in a leak?

I assume that the leakage is expected to b e small.
Weigh it before and after? Or you could feed the vessel through a narrow sight tube (capillary) , using a different fluid in the pump. The interface would be in a different place after the pressure is relaxed and the volume change through the capillary tube could be calculated. Difficult to do at 300Bar, though. But there is a work around. The sight tube could be in a stronger chamber that could balance the pressure. Observing before and after would tell you the loss. This would need to be done at constant temperature.
Pressure testing of diving bottles is often done inside a separate chamber and the water displaced from that outer chamber is measured to find how much the bottle has expanded under pressure. Before and after volume would give you what you want - unless your vessel under test has expanded permanently.
You'd have to do a broad search for 'pressure testing' and home in on any specialist hit you get.

 

[MikeJ]

Thanks, yes the expected leakage will be small. Weighing is not really an option; the unit is 450kg with 0.5L oil so we are unlikely to see any difference.
Also, test pressure is 200barg so a capillary tube is also tricky.

I was hoping there was a definable relationship between pressure, volume and temperature (similar to Gay-Lussac's gas law) where I can find the change liquid volume (i.e. how much has leaked) through the drop in pressure as the vessel volume is constant. I would also assume there is no measurable air/gas in the system.

Regards
Mike

 

[sophiecentaur]

pressure, volume and temperature

I have a feeling that it could be equally difficult to measure those quantities accurately enough to reveal a leak.You could have the same problem distinguishing between liquid loss and creep that you could encounter in my volume change suggestion. The capillary tube problem could be solved if it is mounted in a suitable protective cell during pressurisation. as the capillary tube would not experience any excess internal pressure.
Bubbles could be a problem if the system is not bled well - although, with time, any dissolved gas would re-emerge with no net change in volume.

 

[sophiecentaur]

P.S. How long would you have, to run the test? Given time, you could actually measure the rate that the pressurising pump needs to run to maintain, say 24hours of pressure.

 

[MikeJ]

The tests are run over night; about 16 hours. The other reason I'm looking to try and establish a volume to that the likely leakage is via some Solenoid Operated Valves (SOVs) in a sampling system. So I'd also like to establish, given the pressure drop, how much is leaking into the sample chamber and hence how much contamination to expect with a given leak rate / pressure drop.

It is not practicable to measure how much fluid to add via the pressure pump to return pressure to starting point.

Also, the capillary is not practicable as we fill the test loop (about 1/2 litre) using the pressure pump so it creates a hole hose of problems to add one in and use different colour fluids (or add an interface item and get is lined up in the capillary).

 

[sophiecentaur]

If the liquid loss is enough to cause a pressure drop then you can measure the pressure /volume characteristic as you pump up the vessel from zero to working pressure, plotting a graph and using a curve of best fit. As the pressure drops, you can deduce the amount of liquid lost by reading off the graph. I realize that would involve measuring the volume injected in some way but you must expect a few difficulties along the way.