Gas leak ?

Decided to measure the gas consumption of the boiler pilot light, via the meter test dial, yesterday, within usual range, as reported on the internet. Then today I thought I'd see if there was any leak in the whole system. Turned off the boiler totally just after breakfast and then took readings every hour. Reading increased about 0.0001 cu ft in first hour and 0.0001 cu ft in the next 3 hours. Would most of that first hour just be recovering in pressure over the whole gas distribution system ? Is 0.00003 cu ft per hour leak acceptable, no gas has ever been smelt by anyone. How much influence from changes in atmospheric pressure ?

I tried googling for that info and also what the criteria for manometer testing is, ie what mm or no mm of water drop over how many minutes at what gas pressure, then volume of pipe runs etc to determine it that way. But neither found.

Reply to
N Cook
Loading thread data ...

The message from "N Cook" contains these words:

I can't recall the detail but I am sure it has been aired on this ng in the past. I usually save such postings but it doesn't seem to be the case for that information.

Could be in the gas fitting faq. Logical place for it I suppose but I haven't checked.

formatting link

Reply to
Roger

If I can't find it there or renewed effort on the net tomorrow , I will have to look out BS6891, just to find what the observing time is, and assume

0.5mm or 0.05 mbar is the resolvable limit in that time, work out the volume of 22 and 15 mm pipe , Boyle's law etc to deternmine what the soundness test means in terms of cu ft / hr as an upper bound
Reply to
N Cook

The details are in the FAQ.

Resolution is about 0.25mbar i.e less than 0.25mb drop in 2 minutes is a pass. Here is the text:

(ii) Check for soundness against leakage Original text by "The Boilerdoctor" posted to uk.d-i-y on 15/12/1998, revised 29/1/2002.

This test only applies to a U6 domestic meter; other sizes of meter require different figures for allowable pressure drop. The gas under test is Natural Gas. LPG requires a different treatment; longer test periods, higher test pressures, and no drop at all is permitted. A standard manometer capable of reading 0 to 12" WG or 0 to 30 mbar, and a length (approximately 18") of 1/4" rubber hose is required, and also a supply of leak detect solution.

During the test, DO NOT SMOKE or allow any sources of ignition near.

  1. Check that all outlets and appliance taps are off and pilot lights extinguished. If the installation includes a cooker with a drop down lid which automatically extinguishes the rings when closed make sure that this lid is raised but the taps are off - this is to ensure that as far as possible the cooker internal connections are included in the test.
  2. Turn off the gas service valve at the live side of the meter, remove the test point screw on the discharge branch of the meter and connect the manometer to the test point using the flexible hose.

Open the gas valve SLOWLY until a reading of approximately 8" WG or 20 mbar is obtained in the system, then close the valve. Allow to stand for two minutes to stabilise. The pressure should remain close to the initial pressure, though thermal effects may cause a slight rise or fall. If the pressure has risen at all, bleed a little off to ensure that the system pressure is not exceeding 8" WG or 20 mbar. If the pressure is above this level, a condition known as Governor Lock-Up may have occurred causing a misleading reading. Governor lock up occurs when the pressure downstream of the governor is at or above the set pressure, and hence the governor internal valve is closed. This traps the high pressure gas upstream and if the supply valve is then closed a quantity of high pressure gas remains trapped between the supply valve and the governor. This is a resevoir of high pressure gas and a small drop downstream of the governor will result in the governor valve passing some of this gas. A small loss will thus be masked by this replenishment of the governed side by the high pressure side. Testing at just below the set pressure means the governor valve will be open and thus the test is valid all the way back to the supply valve.

  1. If this happens, repeat the procedure ensuring the gas valve is opened very slowly.
  2. Take an initial reading of the manometer. Wait two minutes. Take a final reading of the manometer. Subtract the final reading from the initial reading and record any loss of pressure.
  3. If the system pressure has risen during the test, the most likely cause is "let by" of the gas valve, in which case the test is invalid and you must seek help from Transco to repair the fault on the Meter Control Valve (MCV). If no detectable pressure rise has occurred, bleed off the pressure until the gauge reads about half of the test pressure. Observe the pressure again to ensure that the pressure does not rise from this reduced level. This is to check that the gas valve is not letting a small amount of gas by, which just might match a loss rate at full test pressure.
  4. If all appliances are still connected to the system, a permissible pressure loss of 1.6" WG or 4 mbar is allowed provided there is no smell of gas! If the appliances are isolated from the system under test or, in other words, the test is on pipework only, then no drop is permitted. If the above conditions are not satisfied, the leak must be traced and rectified, and the system re-tested in its entirety. This is because of the possibility of more than one leak.
  5. If the test is satisfactory, remove the gauge, replace the test point screw, and restore the gas supply. Then test the test point for leakage with leak detection solution.
Reply to
Ed Sirett

My calculation, not checked by anyone else yet, for 0.25 mBar reduction in 20mB over 2 minutes

11m of 22mm pipe and 9m of 15mm pipe, no known dead runs.

volume of that pipe between meter and appliances 5,800,000 cu mm = 354 cu ins = 0.2 cu ft

By Boyle's law drop in pressure of 0.25/20 = 1/80 means an increase in volume (the leak) of 1/80 so 0.2/80 = 0.0025 cu ft in 2 minutes = 0.075 cu ft per hour well above 0.0001 cu ft / hour let alone 0.00003 cu ft per hr

Reply to
N Cook

I would be wary of using just the meter to make any kind of leak assessment. The procedure posted by Ed using a manometer (water gauge) is a far more reliable (and the correct) way to do it.

Reply to
John Rumm

The other interpretation is that the "soundness test" is a very lax test. For a pilot light consumption of 0.005 cu ft / hour you could have 15 such pilot lights all blown out and still not show a leak via the manometer test. Hopefully you could smell the combination of 15 blown out pilots-worth of gas though.

Reply to
N Cook

20mb gauge is about 1020mb absolute. No redo your calcs.
Reply to
Ed Sirett

Anyone who has done the test knows that a pilot light will produce a descending gauge that drops in seconds. Your calcs confused gauge with absolute pressures.

Reply to
Ed Sirett

Interesting point but surely making the pressure in the pipe 1020 mBar and the air outside at 1000 mBar makes no difference to the calculation

Reply to
N Cook

surely the calculation is the same as considering an absolute pressure of

20mB in the pipe and a vacuum outside.
Reply to
N Cook

Quite correct the calculation is independent of the external pressure. It is not however independent of the absolute pressure of the gas in the pipe.

The mistake you made was the bit where you applied Boyle's law to the gauge pressure of the gas in the pipe.

So 0.25mb is a change of 0.25/1020 (or thereabouts) not a change of

0.25/20 which is about 1/4000 no 1/80 which reduces the maximum leakage rate by a factor of about 50. Although you've also neglected to take the very substantial volume of the gas meter into account.

Anyway the acceptable limits for normal domestic natural gas + U6 meter installations are: Pipework max 0.25mbar (difference in gauge pressure) in 2 minutes. Or max 4 mbar difference in gauge pressure in 2 minutes provided all of * Existing installation. * Appliances connected. * No small of gas reported by user.

Reply to
Ed Sirett

Would the rate of loss of gas from the 1020mB situation to 1000mB outside be any different to 20mB absolute in the pipe escaping to a vacuum through the same hole? Yess I'd not thought of the volume in the meter, I assume that is on the consumer side of the vane or whatever momnitors the throughput.

Reply to
N Cook

I have gone back to your original post. One rev of the pointer on a meter with a pointer is 1 cubic ft. I don't see how you managed to measure

0.0001 cu. ft. or whatever.

Anyway the exact chapter and verse for gas tightness testing is given in IGE/UP/1B which is a snip at a mere £77

A summary of this is widely known and reproduced in most gas fitting as stated above.

we are concerned that the amount of gas leakage at about atmospheric pressure is below a small amount. A loss of less than 0.25mb in about

1000mbar in 2mins is < 1/8000 the volume of the installation per minute.

A typical domestic installation including the meter might be about 8 litres of gas. So we are looking to leak less than 1 cubic cm. of gas per minute.

If you are not smelling gas you need do nothing. If you think you have a leak then if you have the equipment and skills perform a tightness test in line with the method stated, or call for pro help.

Reply to
Ed Sirett

red digits with cu ft written under this "milometer" section and then a dial scaled 0 to 1. So the red ones are not 1/10 and 1/100 of a cu ft leaving the last one as

1/100 of a cu ft for one rev which you can read to 1/10000 ?
Reply to
N Cook

Usually HUNDRED cubic feet?

Reply to
Andy Burns

As the last watch dial pointer one is marked 0 to 1 , would int not the 4 charged numbers be thousands to 1 "gas unit" ?

Reply to
N Cook

I'll get there in the end. I dug out a gas bill and although it doesn't say what a gas unit is. It uses a conversion factor of 2.83 which is the conversion of 100s of cu ft to cu m. So I'm out in the above by a factor of 100.

1 gas unit = 100 cubic feet Still don't know what internal gas meter vol is but manometer test discrimination is > 7.5 cu ft per hour and the "leak" I measured on the watch dial pointer means something like 0.01 to 0.003 cu ft per hour.

I will repeat the calc some time by assuming the complete exterior vol of the gas meter is the vol to be included with the pipe run as I cannot find what that vol is

Reply to
N Cook

Assuming this is normal domestic gas meter the black didits are m^3 IIRC The red ones are 0.1 m^3 and 0.01 m^3 (i.e about 3.5 and 0.35 cu ft.)

The dial reads 1 cu ft per rev. Due to the way the meter works the dial is not linear with respect to volumes smaller than 1 rev.

1 rpm of the dial is 18kw
Reply to
Ed Sirett

Sorry you have quite an old gas meter which has black digits reading in

00s of cu ft.

Anyway the dial is still 1 cu.ft. per rev.

Any cryptic comments like 0.071 cu .ft /rev refer to the internal mechanism of the meter not the testing dial.

Reply to
Ed Sirett

HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.