Heat losses in a vertical pipe

[Bob Minchin]

I have an un-pressurised thermal store that has a 15mm copper vertical vent pipe at the top that goes up into the loft and hooks over the Fill and Expansion tank. Although it is insulated with foam, I have a hunch that I'm losing some of the stored heat by convection up the pipe.

How much of the heat is likely to be conducted up the pipe itself as compared by convection in the water it contains?

I have a couple of options.

I could re-plumb using stainless steel pipe which is much less thermally conductive. Useful if the loss is conduction through the pipe.

I could instead reduce the bore of the pipe and hence the amount of water contained in the pipe if the main contribution to the loss is convection.

Someone has suggested that I put a U bend in the pipe run and that the downward section of pipe run from the store will stop the convection losses. This sounds like it might work but I'm not 100% convinced.

What does the team think please

[GB]

My gut reaction is almost nothing lost by conduction. The calculation is not that hard, though. Estimate the cross sectional area of the copper. Work out the temperature gradient (the difference in temperature between the top and the bottom inside the pipe, divided by the pipe length).

Why not use plastic pipe? As the plastic pipe is *much* less conductive than copper, you'll get most of the benefit (if there is one worth having) just by replacing a short length. The conductivity is Copper : 400 (I suspect copper pipe is less as it's not pure copper) Steel : 16 PVC : 0.2 (is plastic pipe made out of PVC? ) Water : 0.6

All W/(m K)

In any case, you'd need to compare the conduction through the pipe walls with the conduction in the water, which has a much greater cross sectional area.

What size is required for safety? If smaller is okay, why did you use

15mm in the first place? If you need 15 and replace it with 10, you'll need to run three pipe runs to get the same overflow capacity.

Convection is much harder to work out theoretically than conduction, but I expect it's tiny. Any upward flow gets obstructed by the downward flow

- all in a 15mm pipe.

Neither am I.

I wouldn't do any of this without experimenting first. Cut off the heat inflow and measure the temperature drop in the store, both with the expansion pipe connected and with it capped off by the tank. You'll need to repeat this a few times in different weather conditions.

You're overdoing this. :)

[Dave Liquorice]

I doubt there is much convection in an insulated 15 mm pipe. Uninsulated in the loft might be different. Conduction might lose you something but with insulation probably not a lot. I guess you could measure how fast the store cools with no heat input, then remove the insulation and see if you can spot a difference.

That you're over thinking the "problem". Insulation is cheap, and effective.

[Roger Hayter]

If the pipe is carrying away heat then it must get hot. If the lagged pipe does not get hot much above the tank at steady state when it is not overflowing, then it is not carrying away a significant amount of heat. Even if it does get warm, lagged pipe does not lose much heat (see tables from lagging firms etc) and so not much heat is being carried away. This is regardless of mechanism, but the length of pipe which gets hot will answer your question to an extent.

[harry]

You can put in a "clack"valve. The flap will remain closed against convection losses but open on over pressure You might need to put in an air vent immediately below the clack valve. Or drill a small hole in the clack. The clack will fall shut in the vertical position

[GB]

I can't actually find a clack valve. Would a spring loaded non-return valve do much the same. It gets in the way of any convection currents, but allows water through if there's an over-pressure. Plus, it's cheap!.

[Chris Hogg]

[Dave Liquorice]

Don't forget any inertial over presure surge when a pump starts up. This may push water past the valve that can't return. Subsequent starts shoving a bit more and bit more etc eventually leading to each start giving a burst of pump over. If there is enough pressure with the system running to open the valve this won't happen.

[Vir Campestris]

+1.

You aren't losing heat if nothing exposed is hot. Feel the pipe before you do anything complex!

Andy

[harry]

There is the possibility that air may accumulate under it. You won't find clack valves in the likes of B&Q You need a valve orientated so that the flap/jumper falls shut with gravity. But can open under pressure. This will stop any convection.

In vertical pipes, warm water rises in the centre and cooler sinks down the periphery. The bigger the pipe the worse this effect is.

The interface between rising and falling water limits the effect

Another way to limit convection is to put a short almost horizontal bit of pipe immediately above your hot water tank.

You can easily tell what is going on. If you have no convection problem, the only warm bit of pipe will be the foot or so of above the tank. (This is heat conducted through the metal pipe) More than this is a convection problem.

Insulating the pipe stops the problem too, the convection current is caused by the heat loss from the pipe. No loss=no convection current.

[newshound]

Also, the heat that has convected up to the free surface has to go somewhere to become a heat loss. It's a tiny area. There will be some convection into the air prior to the "hook" but the only way out is through the walls of the pipe (and the insulation).

+1, basically. If the copper pipe feels "warm" at the top, then go to plastic (but keep the insulation).