Gas system test pressure

[Avocet]

Would anyone have a rough idea what sort of pressure a gas system should hold and for how long? There's a test point on my cooker and I'm sure I've got enough stuff to cobble up a manometer. I'm just wondering how many inches of water (or whatever unit) would be reasonable to apply and for how long it would be reasonable to expect the system to hold it?

 

[jdc]

Pressure will be around 30 millibars. P = rho . g . h, so 30 mb means 30 cm of water.

 

[krbp32]

18 millibars for 5 mins

 

[bluerm166]

A test pressure would normally be higher than the pressure in normal use (c.30mbar) but I don't know what the factor would be.

 

[Easticks28]

Why not fit a 'bubbler' http://www.alde.co.uk/itemdetails.php?itemId=53

I have one fitted in the gas locker immediately downstream of the regulator. Each and every time the bottle is turned on you press down on the red top of the detector, and no bubbles mean absolutely no leaks ( provided of course that no appliance is switched on )

 

[VicS]

Would anyone have a rough idea what sort of pressure a gas system should hold and for how long? There's a test point on my cooker and I'm sure I've got enough stuff to cobble up a manometer. I'm just wondering how many inches of water (or whatever unit) would be reasonable to apply and for how long it would be reasonable to expect the system to hold it?

There are two tests that can/should be done.

One, using air, on the pipework and appliances before the gas cylinder and regulator are connected.;

Pressurise to 70mb with air using a suitable pump.
Allow to stabilize for 5 mins
Then monitor for a further 5 minutes. There should be no fall in pressure.



The other using gas after connecting the regulator and cylinder:

Open the cylinder valve to fully pressurize the system. Then close it again.

Burn off a little gas until the pressure has fallen to 30mb for propane or 20mb for butane.
Close the appliance valve and wait 5 minutes.

Note the exact manometer reading. Leave for 5 minutes and read the manometer again.
There should be no discernible pressure drop


All this came originally from a Corgi Handbook ... it may now have been altered.

 

[Avocet]

Thanks Vic, I'll give those a go.

Yes, I've got a bubble detector, but have always wondered how big a leak you'd need to get a bubble to form in the time I was watching it.

 

[Steve C Cronin]

There are two tests that can/should be done.

One, using air, on the pipework and appliances before the gas cylinder and regulator are connected.;

Pressurise to 70mb with air using a suitable pump.
Allow to stabilize for 5 mins
Then monitor for a further 5 minutes. There should be no fall in pressure.



The other using gas after connecting the regulator and cylinder:

Open the cylinder valve to fully pressurize the system. Then close it again.

Burn off a little gas until the pressure has fallen to 30mb for propane or 20mb for butane.
Close the appliance valve and wait 5 minutes.

Note the exact manometer reading. Leave for 5 minutes and read the manometer again.
There should be no discernible pressure drop


All this came originally from a Corgi Handbook ... it may now have been altered.

If the pressure is holding on air, why do you need the second test with gas?

 

[VicS]

If the pressure is holding on air, why do you need the second test with gas?

I dont know the answer to that except that the first test does not include the regulator and the connection to it whereas the second one does.

 

[RogerMayne]

I dont know the answer to that except that the first test does not include the regulator and the connection to it whereas the second one does.

Because LPG is more "searching" than air, ie it can be airtight but still leak LPG.

 

[Steve C Cronin]

Because LPG is more "searching" than air, ie it can be airtight but still leak LPG.

You would need to keep the test going for a very, very long time to see any difference that could be put down to molecular structure!

 

[Steve C Cronin]

I dont know the answer to that except that the first test does not include the regulator and the connection to it whereas the second one does.

A test of the whole system, including the connections can still be performed with air, so long as you have the correct adaptors. I would always fit a permanently in-line guage like "Gas-Safe" on a boat, anyway. Not to mention turning off at the bottle!

 

[NormanS]

Years ago, with a big boat, with quite a length of exposed copper gas piping, I had a BSS surveyor doing a manometer test. "Sorry", he said, "the pressure is dropping, I suggest you check x, y, and z."
I spent the rest of that evening checking for leaks, even the cooker control valves didn't escape my attention. The surveyor had left his manometer for me, so the next morning, I set it up again, and found that, not only was the system holding its pressure, the level in the manometer was actually rising!
All that had been wrong was that the initial test had been in the evening, when the air temperature was falling. In the test the following morning, the temperature, and therefore the pressure, was rising. Boyle's Law. Lesson learned.

 

[Deleted member 36384]

At work I pressure test to high and low values for a single test procedure (with water and nitrogen). The reason for this is quite simple, at high pressures seals can be caused to seal due to the expansion of the pipe/tube/fitting. However, at the lower pressures the seal will not hold. This is quite common where the fitting has not been made up correctly. It also works the other way about, seal at low pressure, no seal at high pressure.

 

[jdc]

...Boyle's Law...

Some physics at last! Bring it on.

The 2005 BSS tests are described here https://pro.boatsafetyscheme.com/media/152725/appendices c.pdf.

No mention of the air test which is, I think, pretty useless in an already commissioned system (but is one you might want to do as an 'outgoing' QC on a boat leaving the factory). That test might as well be done at a higher pressure than any gas to be used subsequently, so I suppose 70cm on a manometer is a reasonable but essentially arbitrary amount (one doesn't want too big a manometer). It's about a factor of 100 less than any likely burst pressure of any component!

The logic I believe to be as follows (just thinking about it without assuming prior knowledge):

C1: Have a look at it
C3: Check the tap at the bottle actually closes off! So put a tiny bit of pressure in the system with all connected and the bottle closed: does the pressure creep up?
C4: Check if the pipework leaks. This has to be done at a lower pressure than that coming from the regulator, hence with the old system of 37mb for Propane and 28 for Butane, +/- an allowance for tolerance, the pressures were selected to be 30mb and 20mb respectively. Nowadays with both gases using 30mb one might select somewhere 20 - 25mb. So pressurise, turn off the gas at the bottle, let some out so it's clear there's no HP gas upstream of the regulator which might otherwise compensate for a small leak downstream of the regulator, and see if the pressure holds.

They selected 5 mins for the test. I suppose this is rather an arbitrary time chosen to be (a) short enough to be efficient, (b) short enough so that thermal effects (Boyle's Law are probably negligible, and (c) long enough that one would nonetheless detect a leak.


I think ideas such as how 'slipperry' the gas is (whatever that means) don't come into it. Gas leaks are a question of statistical mechanics; one has to imagine a particle of incredibly tiny size, and leakage rate is the probability, on its random zooming around, of going through a hole in a certain time.

 

[Easticks28]

Thanks Vic, I'll give those a go.

Yes, I've got a bubble detector, but have always wondered how big a leak you'd need to get a bubble to form in the time I was watching it.

A very, very small one. The BSS inspector found one via my bubble detector which was producing a bubble every ten seconds or so. It was corrected by a very small tightening of a compression joint. The inspector said that the leak could have been due to the cold weather that day.

 

[Robg71]

The molecular structure of the gas does come in to play. Probably not noticeable on the level your playing with but it most certainly has an affect.
Try 300bar pressure testing on air/nitrogen, no problem..
10 bar of helium..... Ahaha, like a sieve.....

 

[jdc]

The mass of the molecule counts as you say since the energy per particle is constant (~kT) and so light molecules like Helium are inevitably moving faster.

But the suggestion I was doubting is that the different pressures being specified by the Corgi test (20mb for Butane (C4H10), 30mb for Propane (C3H8) and 70 mb for air (largely N2)) is to somehow equalise their propensity to leak when it would actually accentuate differences.

PS: my leak testing experience was the inverse problem of testing high vacuum systems. We used a little bottle of Helium 3 and a mass spec installed in the vacuum side. It all seems a long time ago now...

 

[saltylegs]

There are two tests that can/should be done.

One, using air, on the pipework and appliances before the gas cylinder and regulator are connected.;

Pressurise to 70mb with air using a suitable pump.
Allow to stabilize for 5 mins
Then monitor for a further 5 minutes. There should be no fall in pressure.



The other using gas after connecting the regulator and cylinder:

Open the cylinder valve to fully pressurize the system. Then close it again.

Burn off a little gas until the pressure has fallen to 30mb for propane or 20mb for butane.
Close the appliance valve and wait 5 minutes.

Note the exact manometer reading. Leave for 5 minutes and read the manometer again.
There should be no discernible pressure drop


All this came originally from a Corgi Handbook ... it may now have been altered.

There should be no determinable drop OR RISE in the pressure as this is also checking the cylinder valve for "letby"

 

[Steve C Cronin]

At work I pressure test to high and low values for a single test procedure (with water and nitrogen). The reason for this is quite simple, at high pressures seals can be caused to seal due to the expansion of the pipe/tube/fitting. However, at the lower pressures the seal will not hold. This is quite common where the fitting has not been made up correctly. It also works the other way about, seal at low pressure, no seal at high pressure.

Very real problem.