Running gas pipe

[Garthzog]

Hi.

I need to get a gas pipe run from the kitchen to the joined garage for a new boiler install. The gas rate is 3.5m2/h, and looking at the calculations here it seems like 22mm should be fine:

http://www.cda.org.uk/megab2/build/pub124/sec3.htm

Including bends the length of run is going to be around 8.25 to 9.25m (depending on bends) but I've seen some people say that the calculations can be over optimistic so you're looking at around half that in reality. If so, some of the pipework will need to be in 28mm.

The pipe size from the meter is 3/4, and that will go to 22mm copper for a short run before it can go to 28mm, but how much in your experience will need to be run in 28mm for the run to be considered OK for the boiler? It would be a pain to have to rerun again if the supply is found to not be adequate when tested. The only other appliance that will be on the gas supply will be a 4 ring gas hob.

TIA

 

[sooey]

If it was me I'd come off the meter with a length of 28mm, tee the hob off that and carry on to the boiler in 22mm.
Watch me get ripped now though.

 

[gas4you]

Including bends the length of run is going to be around 8.25 to 9.25m (depending on bends)

If you were corgi reg or competent then you wouldn't have to ask

Is the above length all the way from the meter?

How many 90 bends, how long is the run from meter to T for gas hob and how many elbows?

 

[sanj.varah]

If you were corgi reg or competent then you wouldn't have to ask

I don't think that is fair, I'm a part time lecturer at college/university and I get some plumbers being referred to my fluid mechanics class!

I think the effective length calcs supplied by the copper tube manufacturers are actually pessimistic. When gas flow goes round an elbow, you do get a pressure drop but if you measure down the line you get quite a lot of recovery as there is sufficient time for a turbulent boundary layer to redevelop.

I reckon those tables are put together assuming worst case scenario - you have many fittings right next to each other, which causes much worse flow.

Its also worth noting that most fan assisted boilers actually suck the gas up the gas pipe, so those calcs kind of go out the window anyway.

Personally I would run it in 28mm, this keeps you well in.

sanj

 

[megawatt]

Its also worth noting that most fan assisted boilers actually suck the gas up the gas pipe, so those calcs kind of go out the window anyway.

Have I just been transported to some kind of alternate universe?

I'd stick to fluid mechanics if I were you and leave gases well alone

 

[Gasguru]

I think the effective length calcs supplied by the copper tube manufacturers are actually pessimistic.

I reckon those tables are put together assuming worst case scenario - you have many fittings right next to each other, which causes much worse flow.

Its also worth noting that most fan assisted boilers actually suck the gas up the gas pipe, so those calcs kind of go out the window anyway.

sanj

Surely you mean the optimistic and best case scenario

The fact remains the pressue drop should be no more than 1 mbar from meter to appliance regardless of the effect a 1:1 gas valve has.
Although most fully premix burners appear to operate fine on low working pressures other gas appliances may well be very unsafe.


9m equivelent lenght and 3.5m/hr will be marginal, adding another 0.6 m/hr for the hob no chance.

How far along the run is the hob?

For ease I would come off the meter in 22 then up to 28 as far as the hob tee (if it were say half way along the run) then drop down to 22. All a bit trial and error. Since the calculations are garbage use judgement and a little overkill.

 

[chrishutt]

When gas flow goes round an elbow, you do get a pressure drop but if you measure down the line you get quite a lot of recovery as there is sufficient time for a turbulent boundary layer to redevelop.

Could you explain that a little more, particularly the reference to a turbulent boundary layer? I don't think I've come across that before.

Its also worth noting that most fan assisted boilers actually suck the gas up the gas pipe, so those calcs kind of go out the window anyway.

Even boilers with zero gas pressure where the gas and air mix cannot really be said to "suck" gas from the gas supply pipe since the appliance inlet pressure should still be at least 18 mbar, although they will operate equally well at lower pressures (which is perhaps what you meant?).

 

[baxpoti]

Have I just been transported to some kind of alternate universe?

No.... you've just hit the 21st century

 

[megawatt]

This would be the century where the normal laws of common sense don't apply then?

 

[chilli985]

need more info really but it seems you would have to start it in 28 regardless. need distance to first tee and how many elbows

 

[Onetap]

When gas flow goes round an elbow, you do get a pressure drop but if you measure down the line you get quite a lot of recovery as there is sufficient time for a turbulent boundary layer to redevelop.

Could you explain that a little more, particularly the reference to a turbulent boundary layer? I don't think I've come across that before.

He is right about the local pressure loss and the recovery downstream. It's Bernouilli's equation, conservation of energy; h = z + (rho)g + v^2/2g, if I remember it. There are eddies and turbulence immediately downstream of the fitting and the kinetic energy in the local high velocites causes a corresponding drop in the pressure. Further on the eddies subside and the pressure reverts to normal. Turbulent boundary layer, dunno, but sounds plausible.

This is important with flow measuring devices , orifice meters, with which you measure water flow by measuring the pressure loss ( with a mercury manometer usually) across a calibrated orifice plate. The pressure loss across the orifice is measured, but pressure loss caused by the orifice meter is less than the measured pressure loss, due to the downstream pressure recovery.

I think he's wrong is in assuming that the manufacturer's pressure loss figures don't allow for the pressure recovery; they do, or should do.

 

[kevindgas]

For ease I would come off the meter in 22 then up to 28 as far as the hob tee (if it were say half way along the run) then drop down to 22. All a bit trial and error. Since the calculations are garbage use judgement and a little overkill.

eh! were did you learn that? if you are running 28mm pipe then you need to come off the meter in 28 not 22 as you are restricting the flow and effectively only giving 22mms worth of gas to the 28 pipe.

 

[Agile]

Regardless of what the fluid dynamics expert thinks the plain fact is that people who regularly measure the pressure at a gas appliance find that its always less than the calculated value. Usually 20-70% less.

This is not explainable by theoretical aspects but simply that the workmanship and design of pipe and soldered fitting can seriously reduce gas flow.

Firstly the pipe may not be so smooth due to dirt and corrosion, secondly most copper pipe is not straight cut but cut with a pipecutter which reduces the diameter. Another problem is solder runs into the pipe and jointing compound at screwed joints.

One problem with theoreticists is that they consider all aspects regardless of the significance of the problem. Fan flued boilers mostly create a depression of 1-2 mB in the combustion chamber and with a supply pressure of 20 mb obviously this is only going to be a very small effect on the gas pressure at the gas valve inlet.

Back to the OPs question, to have your boiler correctly notified to CORGI and thence to Building Control, so they can quote the right info for the HIP when you sell the property, a CORGI has to fit all the gas pipe. Certainly some dodgey CORGIs will probably "sign it off" but since its so costly getting registered, few would risk their registration in that way and many will disapprove of any gas work by anyone who is not fully competent.

Tony

 

[chrishutt]

For ease I would come off the meter in 22 then up to 28 as far as the hob tee (if it were say half way along the run) then drop down to 22. All a bit trial and error. Since the calculations are garbage use judgement and a little overkill.

eh! were did you learn that? if you are running 28mm pipe then you need to come off the meter in 28 not 22 as you are restricting the flow and effectively only giving 22mms worth of gas to the 28 pipe.

Sorry kevingas, but it's you who's wrong. It doesn't matter where in the run you have 28mm, 22mm or 15mm for that matter, providing the overall pressure drop is within 1 mbar. For example a metre length of 22mm pipe immediately downstream of the meter has exactly the same resistance as a metre length immediately before the first branch off the main run.

 

[sanj.varah]

turbulent boundary layers occur on the walls of the copper pipe. the molecules of gas rub against the side of the wall and friction slows the gas flow down. This is why you get a pressure drop along a length of pipe.

Very close to the wall of the pipe, the gas velocity is zero and this is called a boundary layer (its basically a static layer of gas). two types of layer exist, laminar and turbulent, at a given point (called a reynolds number), the flow goes turbulent.

When you go round a bend, the problem is the boundary layer breaks apart and you get all manner of things happening which are (as Agile pointed out) very difficult to model.

Just to clarify something else. Bernoulli's equation does not apply to gas, it only applies to incompressible fluids which are also inviscid (have no viscosity/friction). Although having said that, it is readily applied to water calcs. The governing equations for gas flow are Navier Stokes.

Although what they teach in college or whatever states that it makes no difference where the bigger bore section of pipe is, in reality due to the low pressure of gas, its always better to have 28mm at the source and then getting smaller with branches. This is because you never ever get 100% recovery.

I'd just also point out, that the calcs the OP hyperlinked allow for a pressure drop of 0.1mbar. Agile stated that pressure drop is 1-2mbar in a combustion chamber, that more than negates that effect.