How can I cure my gravity circulation problem?

[gmartine]

The earlier copper cylinders were 100% copper and the new cylinder is 100% stainless steel. If you view conductivity on a log scale between rubber and metal, glass is somewhere in the middle and all metals are very close together at the end. Any difference between copper and steel conductivity is negligible in this application.

If the rate of heat transfer is critical to how a thermosyphon works and yours appears particularly sensitive, irrespective of the scale compared to others if you then used a material that was subsequently 1/20 less efficient at doing so it isn't a negligible factor.

 

[Neville Hillyer]

Above posts are interesting but while copper should give a faster warm up than stainless for the same coil area, using stainless should not cause stalling.

Is there any way that the coil has been installed somewhat like the attached even though the flow/return connections are shown at 600mm apart in your drawing.

Also, if you don't me asking, what do you intend to do to resolve your problem as you have lived with it for a year or so?.

I saw a report of copper being faster but this was for identical metal thicknesses. Stainless cylinders are normally designed to withstand specified pressures which results in lighter and thinner cylinders and coils. In my case I am not concerned with speed. I have ample storage for my preferred relatively low primary temperatures.

As far as I could tell my coil approximates to a normal spiral between the connections. It is however supported by stainless wire and thin straps which, combined with the very flexible coil, caused it to 'sing' when moved empty just like moving a box of very light springs.

As a retired design engineer I am determined to do all I can to understand what is happening. Next I may try isolating the coil, disconnecting the primary from the boiler and power flushing each pipe in turn from roof to kitchen.

The attached schematic has been updated for: pipe colours and sizes, cylinder and pump isolation valves, non-return valve and pressure relief valve.
[automerge]1589733757[/automerge]

If the rate of heat transfer is critical to how a thermosyphon works and yours appears particularly sensitive, irrespective of the scale compared to others if you then used a material that was subsequently 1/20 less efficient at doing so it isn't a negligible factor.

I am afraid I misunderstood. I thought your remarks were about energy efficiency rather than speed. You could have a point but I suspect this is not the cause of my problem.

 

[John.g]

Is the ~ 0.5M horizontal section of pipe hot right up to the coil inlet?.
You might also consider isolating the coil and install a 600 mm (spool) piece of pipe between the flow&return and see if you get thermosyphon circulation.

At the end of the day, if all else fails would you consider a pumped solution?.

 

[Brambles]

Neville,

I think that you have hit the nail on the head by looking at the calculations from first principles - I have never needed to calculate the pressure differential required to drive the thermosyphon effect before. It is a bit of an iterative process, but with water at 60 degrees C ( for the density) and a required ( my estimate) velocity of 0.5m/second the pressure differential to drive the flow 1meter is:

22mm inside diameter 1.00 inch head
28mm inside diameter 0.15 inch head

Thereafter for a 7m coil multiply the above differential by 7.

If you reduce the required velocity to 0.25m/second the equivalent pressure differential needed drops by a factor of 4.
At 0.125m/second the differential pressure required drops by a factor of 16.

There are rough estimates in the above for the friction factors and Reynolds Number. However, I was surprised by the significant influence that internal diameter has, and presumably in the longer term corrosion and silting of the flow pipework.

 

[SJB060685]

Neville,

I think that you have hit the nail on the head by looking at the calculations from first principles - I have never needed to calculate the pressure differential required to drive the thermosyphon effect before. It is a bit of an iterative process, but with water at 60 degrees C ( for the density) and a required ( my estimate) velocity of 0.5m/second the pressure differential to drive the flow 1meter is:

22mm inside diameter 1.00 inch head
28mm inside diameter 0.15 inch head

Thereafter for a 7m coil multiply the above differential by 7.

If you reduce the required velocity to 0.25m/second the equivalent pressure differential needed drops by a factor of 4.
At 0.125m/second the differential pressure required drops by a factor of 16.

There are rough estimates in the above for the friction factors and Reynolds Number. However, I was surprised by the significant influence that internal diameter has, and presumably in the longer term corrosion and silting of the flow pipework.

Brambles any chance you could provide a link to this calculation method?

 

[Brambles]

SJB

You don’t normally need it for general plumbing - the results are normally tabulated into charts such as the ones that give optimum flow rates for different diameter pipes.

If you want to calculate the values from scratch use the Darcey Weisbach equation:

P= f * v2 * p/2*L/D
P = pressure
f = friction factor
v2= velocity of the fluid squared
p = density of the fluid
L = pipe length
D= internal pipe diameter

f is determined (estimated) by calculating the Reynolds Number (Re)

Re = V * D *(p/vi)

vi is the viscosity of the fluid.
V is velocity of the fluid.

Published tables then give you the friction factor ( f )for a specific Reynolds Number (Re) to input into the first equation.

It is an iterative process so you need to do it a couple of times to be sure that the answer starts to be consistent.

I am unclear as to what the minimum velocity is for a gravity hot water system to be effective - but I would guess that the minimum flow rate that allows the boiler to operate properly is a key design criteria for the tank to match.

Does this make sense?

 

[SJB060685]

Yes makes perfect sense.
Like you said its not really needed to go into such detail but I'm the sort of person who doesn't like not knowing something exactly. If something like this pops up I will go and have a read through and mock up some rough figures just to get a feel for it

 

[Neville Hillyer]

Is the ~ 0.5M horizontal section of pipe hot right up to the coil inlet?.
You might also consider isolating the coil and install a 600 mm (spool) piece of pipe between the flow&return and see if you get thermosyphon circulation.

At the end of the day, if all else fails would you consider a pumped solution?.

The flow is about 40cm and the return is about 50cm to the tank connections.

Not as easy as it sounds but is is possible. However, I already have excellent thermosyphon circulation through the bathroom radiator. An easier trial would be to temporarily join the two pipes at the small header tank.

As indicated I would be very reluctant to abandon gravity circulation especially as it has worked well for 50 years.

 

[John.g]

I don't think joining the pipes at the header tank is a good idea as there will be virtually no static head at this point but as you stated there is excellent circulation through the rad and one might expect the flow pipe to be hot almost right up to the coil connection due to thermosyphon circulation within the pipe itself.

 

[SJB060685]

@Brambles can you say to me which units you were using during your rough calculations for the above required pressure.

 

[Brambles]

For the Darcey Weisbach

V m/second
p kg/m3
L m
D m

For the Reynolds Number:

V m/second
D in mm
For density / viscosity at 60 degrees C for water use 3.1m

 

[SJB060685]

Thanks. Without that there was no way I could knock up some rough numbers 👍

 

[Brambles]

P is calculated in N/m2

1 inch of head (water) is around 250 N/m2

 

[SJB060685]

Yes 1 N/m2 equates to 1 Pa. 1 bar equals 100 N/m2 or 100,000 Pa

 

[Brambles]

No, 1 bar = 100,000N/m2

 

[John.g]

I get the gist of what is being discussed above but it must be remembered that the gravity driving force available is only determined by the flow/return temps and the head available.
If one realistically accepts a max head of 2.5M and a flow/return temp 75/15C with a cold cylinder and 75/60C when hot, then the max circulating force is 0.061M/2.42ins and the min is 0.021M/0.84ins. A pumped system with a differential head of 3M will give a flow rate factor of X12 in the first case and X7 in the second case and that is one and the main reason that pumped systems are now almost universal. I do have a vague memory (50 years) of some gravity driven heat exchanger with a huge finned coil.

 

[Brambles]

John G,

I fully agree. The calculation may reveal the cause, but it contributes very little to delivering hot water !

 

[John.g]

I don't think the calculation reveals the cause, gravity fed systems have been delivering hot water for hundreds of years?. Pumping the above system may and probably will give hot water but a gravity system was specified.

 

[SJB060685]

Yes Brambles, that was a typo error.

 

[Brambles]

It’s not a typo - the calculation is showing that the coil inside diameter needs to be around 28mm for the gravity and temperature generated pressure differential to be greater than the Installed pipe (and coil) resistance.