Willis Water Heating System

[JohnW2]

Can I suggest a little experiment with some plastic pipe as per EFL's diagram? Fill the pipe up with water, but add some coloured dye at one end, join the two ends, formed in a vertical loop, then warm up one side, where the dye is. The dye will rise up over the top of the loop, and fall at the cooler side.

It would not 'fall on the cooler side', since buoyancy considerations dictate that it has to remain above the cooler water on that side. It could only 'fall on the cooler side' if it became cooler than the water below it - and that's not going to happen.

However, your experiment does not really reflect the actual situation, in which the amount of coloured dye (heated water) would be progressively increasing. In that situation, increasing amounts of warmer (coloured) water would rise to (and accumulate at) the top of both arms of the loop - but since one of those arms is the cylinder, that means that progressively more water would appear in the upper parts of the cylinder.

Kind Regards, John

 

[JohnW2]

Phew It is certainly convection, convection currents in water.

As I've said, it's not really 'convection' in the sense we normally think of it ('rising and falling', creating 'circulation'), but really just a question of 'warmer water rising' due to buoyancy.

Next problem - when hot water is drawn, what stops it drawing hot water from both the cylinder and cooler water through the Willis?

That's presumably not a problem so long as the amount of hot water one wants is not more than about double or so the volume of the Willis, since the taps will then get a 'mixture of hot and hot'. A potential problem arises if one wants a lot more than that (such as the 54L for a bath I mentioned, which Willis instructions say requires about an hour of heating), since one will then start drawing cold water through the Willis after it's hot content has all been drawn off.

The only thing I can think of is that flow from Willis is fairly restricted, so that a much higher proportion of water (to taps) comes from the cylinder than from the Willis. Piccies I've seen show 15mm pipes from the Willis, whereas a feed to a bath tap will probably be 22mm - but I'm not sure that would be enough to avoid some degree of a problem.

Maybe there is some 'explicit restriction or, perhaps more likely, maybe it is just 'accepted' that when one wants a large volume of hot water with a Willis, that the temp of supplied water will decrease after a while (after its content of heated water has been exhausted)?

Kind Regards, John

 

[JohnW2]

Mr Willis has a lot to answer for......

He sure does

Kind Regards, John

 

[EFLImpudence]

Ah, I took your diagram too literally (no heat source) particularly because I took (probably incorrectly) your mention of 'bubble' to mean an 'air bubble'.

I did say "if it continuously replenished".

However, I think this may well be a "Eureka! Moment"!

I think that we (certainly I) have probably totally confused ourselves with all the talk about 'convection' and, in particular, "circulation around convection loops", because I now don't think that has really got anything to do with it.

I did say "If the pipe were small enough to prevent convection within it" that is changing place with the cooler water above.

Referring to your diagram, if one continues to heat water at the position of your 'lower density white part', then all the newly heated water would rise up that arm of the 'loop', progressively increasing the amount of hot water at the top of the loop (in both arms). Although the heated water in 'the other arm' of the loop would still be 'at the top' of the water in that arm (and could not fall 'by convection', since the water below it was cooler), the bottom of that region of hot water in that 'other arm' would gradually be 'forced down' by the increasing amount of hot water - despite the fact that convection/buoyancy was keeping uit above the cooler water below.

...but you said the newly heated water would rise up; therefore cold water must be drawn in at the bottom

Given that, in the actual situation we are discussing, that 'other arm of the loop' is the cylinder then, yes, this appears to be (at long last) an explanation of how the cylinder gradually fills with heated water from top downwards. QED!

Yes, but not only by conduction; it will be moving.

Put another way, this is absolutely nothing to do with convection or 'circulation in loops'

I disagree. Are you saying a gravity system, which this is, does not circulate the water?

but is merely a consequence of the fact that as progressively more hot water is created and 'rises' to the top of the 'loop', this increased amount of hot water has risen to the top of both 'arms' of the loop, one of which 'top arms of the loop' is the upper part of the cylinder. The important thing (per all the confusions) is that nothing is 'circulating' - the hot water is simply staying at the top of the loop (both 'arms') just as it should - and all the talk about 'water rising and causing more to be drawn in' has really just confused the issue even more!

It cannot be rising if not circulating.

Phew! Unless I've got this wrong, it's fairly satisfying to have at last 'got there' (since we all know that, because it 'works', there had to be an explanation!). I may award myself a drink!

Tea.

 

[JohnW2]

• I did say "If the pipe were small enough to prevent convection within it" that is changing place with the cooler water above.
  [*]but you said the newly heated water would rise up; therefore cold water must be drawn in at the bottom
  [*]I disagree. Are you saying a gravity system, which this is, does not circulate the water?
  [*]It cannot be rising if not circulating.

My point was that there cannot be circulation around the entire 'loop' (Willis+Cylinder) by convection, since that would require that water 'sank'/'fell' from a hotter to a cooler part within the cylinder. Any movement of water between the bottoms of the two arms would be due to displacement (by more hot water above), not convection, and I wouldn't personally regard that as 'circulation' (which, to me, implies a process involving 'continuous flow', not just 'local adjustments').

If (as I do) you had a pair of water buts connected by a pipe near their bottoms, if you poured water into one, some would flow through the interconnecting pipe to equalise the water levels in the two - but I wouldn't call that (which is analogous to what we're talking about) 'circulation', would you?

Despite your comments about 'pipes too small to allow convection', I suspect that there could well be 'convection' (up and down) within the pipe coming from the Willis. If you did experiments with pipes of different sizes, I think you might be surprised by how small a pipe 'allows convection' (fill pipe with water, heat the bottom and see if hot water appears at the top).

Don't forget that, contrary to diagrams, the 'up and down' paths of a 'convection loop' to not have to be physically separate - warmer molecules (vibrating more) can 'rise through' water at the same time as cooler molecules (vibrating less) are 'falling through' the same bit of a water (passing one another within that water).

Yes, but not only by conduction; it will be moving.

I've never suggested that the Willis system works by conduction. On the contrary, I've said that conduction could not possibly explain how it works.

Tea.

Not quite what I had in mind

Kind Regards, John

 

[EFLImpudence]

My point was that there cannot be circulation around the entire 'loop' (Willis+Cylinder) by convection, since that would require that water 'sank'/'fell' from a hotter to a cooler part within the cylinder. Any movement of water between the bottoms of the two arms would be due to displacement (by more hot water above), not convection, and I wouldn't personally regard that as 'circulation' (which, to me, implies a process involving 'continuous flow', not just 'local adjustments').

I think that is what I have been saying, isn't it?

If (as I do) you had a pair of water buts connected by a pipe near their bottoms, if you poured water into one, some would flow through the interconnecting pipe to equalise the water levels in the two - but I wouldn't call that (which is analogous to what we're talking about) 'circulation', would you?

That is not analogous to what we are talking about.

Despite your comments about 'pipes too small to allow convection', I suspect that there could well be 'convection' (up and down) within the pipe coming from the Willis. If you did experiments with pipes of different sizes, I think you might be surprised by how small a pipe 'allows convection' (fill pipe with water, heat the bottom and see if hot water appears at the top).

Ok, but again, you are not using correct analogies.
Substitute the heater in the Willis with a small pump/fan for a true analogy.

Don't forget that, contrary to diagrams, the 'up and down' paths of a 'convection loop' to not have to be physically separate - warmer molecules (vibrating more) can 'rise through' water at the same time as cooler molecules (vibrating less) are 'falling through' the same bit of a water (passing one another within that water).

Ok. does the speed of transfer of the cooler molecules exceed the speed of the rising water?

I've never suggested that the Willis system works by conduction. On the contrary, I've said that conduction could not possibly explain how it works.

Neither has anyone else. It is gravity.

If the loop were not water but air heated by the Willis element, do you not think that would result in circulation of the air?

 

[Harry Bloomfield]

As I've said, it's not really 'convection' in the sense we normally think of it ('rising and falling', creating 'circulation'), but really just a question of 'warmer water rising' due to buoyancy.

It certainly is! This is the definition - https://en.wikipedia.org/wiki/Convection

 

[JohnW2]

I think that is what I have been saying, isn't it?

I didn't think so - you have repeatedly talked about 'circulation' - which, to me, implies a constant flow around a loop/circuit - something which does not happen with the Willis system. What do you understand 'circulation' to mean?

That is not analogous to what we are talking about.

I have to disagree. The only (small) movement between water at the bottoms of the two 'arms' of the loop (cylinder and Willis+pipes) will be that needed to equalise the heights of the interface between hot and cold water in the two arms - just like equalising the water levels in the two water butts.

Ok, but again, you are not using correct analogies. Substitute the heater in the Willis with a small pump/fan for a true analogy.

That's not a 'true analogy' -0 it's a total moving of the goalposts. With a pump (anywhere) in the circuit/loop there will obviously be a constant flow 'circulating' around the circuit/loop. With the actual Will situation, al that happens is that heated water rises to to top of the circuit, whereafter 'small adjustments' have to be made to equalise the heights of hot/cold interfaces in both 'arms' of the loop. Nothing 'circulates' (at least, in the sense that I understand and use the word).

Ok. does the speed of transfer of the cooler molecules exceed the speed of the rising water?

Obviously not (at least, 'on average') - it's a 'finite situation', so what 'goes up' in a given time period has got to be balanced by what 'goes down' in the same period - otherwise either 'holes' or impossible attempts of two molecules to occupy the same place would arise.

Neither has anyone else. It is gravity.

I don't really know what you mean by that. It is when the 'upthrust' created by something of lower density being within a liquid or gas of higher density is greater than the force of gravity pulling that 'something' down that the 'something' rises (due to the 'net upward force') - or, conversely that if gravity is greater than the 'upthrust' that something will fall (due to the 'net downward force') - is that what you mean?

If the loop were not water but air heated by the Willis element, do you not think that would result in circulation of the air?

Why should the principle differ according to the nature of the liquid or gas we were considering?

Kind Regards, John

 

[JohnW2]

It certainly is! This is the definition - https://en.wikipedia.org/wiki/Convection

I'm not going to argue about definitions. My point is not so much about the word 'convection' as 'circulation' - which, as I've said (in terms of the way I understand and use that word) is not something that happens to any appreciable extent in a Willis system, not the least because it does not usually involve the storing of heated water for appreciable periods.

With conventional ('stored') water heating systems it is very different, since there is constant 'circulation' (by convection) of the stored water as it loses heat and falls to be re-heated, whereupon it rises again. Water heated by a Willis system is (should be!) used before there is time for much of that to happen - and, if it does happen, that is a 'complicating aspect' which we have not considered at all.

Kind Regards, John

 

[EFLImpudence]

I didn't think so - you have repeatedly talked about 'circulation' - which, to me, implies a constant flow around a loop/circuit - something which does not happen with the Willis system. What do you understand 'circulation' to mean?

I disagree - however slowly it happens; it must be circulation.

I have to disagree. The only (small) movement between water at the bottoms of the two 'arms' of the loop (cylinder and Willis+pipes) will be that needed to equalise the heights of the interface between hot and cold water in the two arms -

Even if correct, don't you think that when the heater is on that will keep happening?

just like equalising the water levels in the two water butts.

Again I disagree; the water butts are not a loop.

That's not a 'true analogy' -0 it's a total moving of the goalposts. With a pump (anywhere) in the circuit/loop there will obviously be a constant flow 'circulating' around the circuit/loop. With the actual Will situation, al that happens is that heated water rises to to top of the circuit,

You keep saying the heated water will rise but seem to think it can rise without going anywhere.

whereafter 'small adjustments' have to be made to equalise the heights of hot/cold interfaces in both 'arms' of the loop. Nothing 'circulates' (at least, in the sense that I understand and use the word).

...but again - if the water rises then so does the water in the bottom of the Willis. Why else is there a cold pipe at the bottom of the Willis?

Obviously not (at least, 'on average') - it's a 'finite situation', so what 'goes up' in a given time period has got to be balanced by what 'goes down' in the same period - otherwise either 'holes' or impossible attempts of two molecules to occupy the same place would arise.

Yes - up on the Willis side; down on the other.

I don't really know what you mean by that. It is when the 'upthrust' created by something of lower density being within a liquid or gas of higher density is greater than the force of gravity pulling that 'something' down that the 'something' rises (due to the 'net upward force') - or, conversely that if gravity is greater than the 'upthrust' that something will fall (due to the 'net downward force') - is that what you mean?

One side of the loop is denser than the other. Gravity will balance the sides.

Why should the principle differ according to the nature of the liquid or gas we were considering?

I don't think it does differ.
I think it would happen with a liquid or a gas; you think happens with neither.


I re-refer the honourably gentleman to the gravity hot water system.

 

[JohnW2]

I disagree - however slowly it happens; it must be circulation.

OK - we must use the word(s) differently. To me, 'circulation'/'circulating' implies 'continuously going around and around' (a loop or circuit) - like the water in a CH system or car's cooling system, the coolant in a fridge or, indeed, the blood in "the circulation" of your body. With a Willis system used 'as intended', the total movement will only be a (usually small, unless one has a bath!) fraction of 'one circuit'. I wouldn't say that water was 'circulating' in my CH system if it moved, say, only 10% of one 'circuit' in a day!

Even if correct, don't you think that when the heater is on that will keep happening?

It will but, as above, only to the extent of the water move 'a small fraction of one circuit' in total.

Again I disagree; the water butts are not a loop.

Exactly my point. Water butts are not a loop and (in terms of the way I use the word) there is no 'circulation' between them (just movement of a small fraction of their content between them). As above, I regard the Willis system similarly - not a loop with 'circulation' but rather a connection through which a small proportion of the content will flow from one side to the other.

You keep saying the heated water will rise but seem to think it can rise without going anywhere.

I would probably say that it's you who is the one who is thinking/talking about the 'rising' of some 'object'. 'Rise' is perhaps a potentially misleading word, since what we are talking about is simply redistribution of parts of a single volume of water. Nothing 'goes anywhere', since there is exactly the same volume of water after the 'rising' than before. Take a tall glass full of water and somehow heat the water at the bottom (drop a bit of hot coal into it?!). That heated water will 'rise' to the top, but no water will 'go anywhere' and the total volume of water in the glass will remain unchanged.

One side of the loop is denser than the other. Gravity will balance the sides.

Yes, I think that's a reasonable way of looking at it - it's certainly true that different densities would have no effect in the absence of gravity. I also certainly think that's a much more reasonable way of looking at it than is calling it 'circulation'.

However, we're now really just quibbling about words. The important thing, at least for me, is that (after a long period of getting there!) I now understand how water gets from the Willis heater into the top of the cylinder, no matter what language is used to describe the process.

Mind you, going back to where I started when I created this thread, I still don't think that, in practice, this system offers a lot of advantage over a cylinder with a couple of immersions - if the latter is also used in the same way (which would 'drive me mad'!) as a Willis system!

Kind Regards, John

 

[Harry Bloomfield]

Exactly my point. Water butts are not a loop and (in terms of the way I use the word) there is no 'circulation' between them (just movement of a small fraction of their content between them). As above, I regard the Willis system similarly - not a loop with 'circulation' but rather a connection through which a small proportion of the content will flow from one side to the other.

In the Willis system and for a full cylinder of HW, the Willis will exchange it's water volume several times over.

 

[Harry Bloomfield]

That heated water will 'rise' to the top, but no water will 'go anywhere' and the total volume of water iin the glass will remain unchanged.

That is 'going' somewhere in my book. I might take a walk into the local high street then return home, nothing much has changed, but I have gone/ been somewhere.

 

[Harry Bloomfield]

Mind you, go back to where I started, I still don't think that, in practice, this system offers a lot of advantage over a cylinder with a couple of immersions - if the latter is also used in the same way (which would 'drive me mad'!) as a Willis system!

That I agree with! It offers the ability to heat a much smaller amount of hot water than either an open vented gas boiler, or an immersion heater, but to make it more economical it's use needs to be continually planned in advance, or via an extensive program.

 

[JohnW2]

In the Willis system and for a full cylinder of HW, the Willis will exchange it's water volume several times over.

Yes, as I said, if one tries to heat a large volume of water with a Willis system, then one has the same issue as in a conventional system, in that some of the heated water will cool before it is used, and therefore will need to be re-heated - hence the 'traditional' convection loops. However, that is not the situation we have been considering, and I'm far from convinced that it would necessarily be sensible to use a Willis heater to heat a whole cylinder of water.

Kind Regards, John