Willis Heating System

[JohnW2]

All your sitting here saying "in my mind, in theory, it doesn't work" isn't going to change the universe's willingness to demonstrate that it does work ...

That 'universe' seems to be fairly restricted to one small island.

... so perhaps you need to pour less scorn on those who understand it trying to help you do the same

From the very beginning of this thread which I started, and for the reason I started it, I have been asking people to help me understand their view - but, unfortunately, no-one has yet succeeded.

Not all cars have water pumps.

Very true.

Some just rely on the same principles here, to circulate water in order to redistribute heat

Indeed, but that's because things are such that the conditions for convection are satisfied. If the only source of heatwas being injected into the coolant above the bulk of the (cooler) coolant (and the radiator), it wouldn't work.

As I've said, I have no problem in agreeing that the Willis system will 'work' in eaxctly the same sense as would be achieved by a 'high up' immersion in a single cylinder - but, as things stand, that's all.

 

[Harry Bloomfield]

So what do you think is the difference - the absence of a connection between the bottoms of the two cylinders, perhaps? If so, why should that make any difference to anything?

Because that will put a complete end, to water rising out of the top of the small cylinder.

Have you never studied, a none-pumped heating system? They need both a flow, and a return pipe.

 

[JohnW2]

Because that will put a complete end, to water rising out of the top of the small cylinder.

You seem to be back to talking about the Willis system, since there was no 'small cylinder' in the diagram I was asking you about.

However, as I've said repeatedly, convection will not result in "water rising out of the small cylinder (or anything else)". What it will result in is warmer water within the water already in the small cylinder (or anything else) (plus any connected pipework) re-distributing within the existing volume of water so as to be above any cooler water in that existing volume

Have you never studied, a none-pumped heating system? They need both a flow, and a return pipe.

They do - but they presumably only work IF the source of heated water is connected to the cylinder below its top - since it would otherwise be difficult to see how any of the hot water would get into the cylinder, since that would require it to 'fall' through cooler/denser water (i.e. 'defy gravity'), wouldn't it?

 

[Harry Bloomfield]

You seem to be back to talking about the Willis system, since there was no 'small cylinder' in the diagram I was asking you about.

Small cylinder, large cylinder - it makes zero difference.

What it will result in is warmer water within the water already in the small cylinder (or anything else) (plus any connected pipework) re-distributing within the existing volume of water so as to be above any cooler water in that existing volume

Only if there is a route at the base, for the displaced cooler water, to replace it.

 

[EFLImpudence]

...

 

[EFLImpudence]

 

[JohnW2]

Only if there is a route at the base, for the displaced cooler water, to replace it.

As I keep saying, convection requires no 'replacement' of any of the water. Give or take the need to address the issue of slight expansion due to heating, convection will take place in a 'sealed' cylinder/pipe/whatever if it is heated at the bottom.

 

[JohnW2]

<image>

Yes, it's getting confusing (at least for me). I recently came across a very similar diagram on-line as the one you posted, and don't really understand why water will flow through the coil if the the hot feed is at its top - so I'm thinking about that!

Thinking aloud (and remembering that it's probably over 60 years since I last lived in a house with such a system!) I would have thought that if one wants to directly heat water in a cylinder without a pump, one would have to have the 'hot water feed' connected to the bottom of the cylinder (or coil), with the return coming from the top - but your diagram (like the one I found) shows the opposite of that - so, again, I'm thinking

 

[Harry Bloomfield]

As I keep saying, convection requires no 'replacement' of any of the water.

The Willis system is entirely dependent on a convection current - a flow of heater water, out of the top, which is replaced by cold coming in the bottom.

 

[SUNRAY]

I started writing this ages ago and been out since so some may now be repeats.

John are you comfortable with the fact that hot water rises to the top of cooler water. For example it's always the top of a radiator which heats first?


Following on from that are you comfortable with a vessel filled with water and a heat source - shown red and low down - the warmed water will rise to the top?

Following on, fitting a pipe loop from top to bottom the heated water will start to rise into the pipe

and subsequent heated water will continue to rise to the top and into the pipe

in the restricted space by the heater the natural flow of water due to heat is up

and there is little chance for recirculation down past the heater

as there is a modest flow permitted from top to bottom via the pipe

that is a much easier path meaning the heated water will quickly be flowing round the pipe and entering into the bottom of the vessel

I haven't any experience of a domestic Willis system but recent google indicates a vessel of very modest size, roughly the size of of the immersion heater itself

or to put it another way less than the capacity of a kettle and one of those will boil water in a very short time.

What I do have experience of is calorifier systems with external heaters which took me some time to understand the operation but as I've just discovered is roughly the same format as Willis's. However being commercial contained multiple large 3ph heaters in individual vessels between headers and isolatable for maintenance purposes. In the same system maybe heat exchangers in small vessels added in the same way from boilers or solar arrays.

I have found very different opinions regarding the position of the cool feed to the heater battery, some say it is essential it comes from the stored water rather than the cold feed otherwise draw off can easily be directly from the heater battery and unpredictable temperature.

 

[JohnW2]

The Willis system is entirely dependent on a convection current - a flow of heater water, out of the top, which is replaced by cold coming in the bottom.

OK. As I imagine you probably realise, my problem/frustration is that you folks are all asserting what you believe (quite probably correctly) is what does happen, but without helping me to understand what is wrong with my intuitive (I would have said physics-based) argument that 'it should not happen'.

There is one thing about which we might perhaps have more agreement .... that, as far as I can see, the fact that the 'Willis heater' is external to the main cylinder is pretty irrelevant - since it seems to me that having it 'external' achieves little, if anything, more than would an immersion at the same 'height' in the main cylinder. If you agree with that, it might explain why the idea has not caught on very much outside the island of Ireland?

 

[JohnW2]

John are you comfortable with the fact that hot water rises to the top of cooler water. For example it's always the top of a radiator which heats first?

Yes.

Following on from that are you comfortable with a vessel filled with water and a heat source - shown red and low down - the warmed water will rise to the top?

Yes.

Following on, fitting a pipe loop from top to bottom the heated water will start to rise into the pipe .
..... and subsequent heated water will continue to rise to the top and into the pipe

Yep & Yep.

... and subsequent heated water will continue to rise to the top and into the pipe .... in the restricted space by the heater the natural flow of water due to heat is up .... and there is little chance for recirculation down past the heater .... as there is a modest flow permitted from top to bottom via the pipe that is a much easier path meaning the heated water will quickly be flowing round the pipe and entering into the bottom of the vessel

Aha! This, I think, has been my problem. Everyone has been talking about convection, so my mind has been focussed on that, but this is not convection - rather it is 'brute force'!
The hotter (less dense) water is not moving down through the cooler (denser) water below (which would be contrary to the Laws of Physics) but, rather, is remaining above the cooler water, but (by 'brute force') is pushing the interface between hotter and colder water further down the pipe.

So, thanks to you, I think I've now got my thinking straight about this, and do now understand 'how it works'. However, I'm still inclined to stick with what I most recently wrote - namely that I don't really see what advantage this external heater offers over having an immersion within the main tank at the same height (above base of cylinder) as the external one - do you?

Thanks again.

 

[SUNRAY]

Well it is convection in as much as hot water rises but being a restricted space there is no space to circulate.

I can only reference commercial installations - one installation had about ½ dozen HW vessels dotted around the site, I'll estimateat about 1000gall/5000L - 2.4m*1.5m dia or so - they each had wet solar panels, heat exchangers from gas boilers and up to ~6 3ph immersion heaters. The thing about all of that was very limited specialist kit and easy maintenance with each heater and HX behing isolating valves.

One advantage is modest instant HW is available directly off the heater vessel without having to wait for vast amounts to be heated first. However I don't know if this is a real feature.

 

[JohnW2]

Well it is convection in as much as hot water rises but being a restricted space there is no space to circulate.

OK - but, at least in terms of what I was taught (albeit an awfully long time ago), 'convection' is not just a matter of "hot water (or whatever) rising", but actually requires circulation.

..... One advantage is modest instant HW is available directly off the heater vessel without having to wait for vast amounts to be heated first. However I don't know if this is a real feature.

Yes, I've been saying that all along but, as I have said, there are other ways of achieving that.

.... However, I'm still inclined to stick with what I most recently wrote - namely that I don't really see what advantage this external heater offers over having an immersion within the main tank at the same height (above base of cylinder) as the external one - do you?

Actually, on further reflection, the Willis system might possibly be 'worse' than a 'high immersion' (with 'high thermostat') in the main tank. With the latter, the immersion will switch off the element once the top bit of water, down to the immersion/thermostat is 'up to temp'. However, with the Willis system, I think that it all depends upon where it's thermostat is sensing temp (does anyone know?) - in the 'worst case' I think that (unless it were manually 'turned off') the process you described could go on until the entire contents of the main cylinder had been heated - just as presumably would be the case with a 'high immersion' controlled by a 'low thermostat' in the main tank?

 

[robinbanks]

So you understand A, but not B

I'm not really sure what do with that

They're both immersion heaters. You seem to accept that water heated by the element, near the element, floats away from being near the element. The thing that takes its place (for there must be something) is cooler water, not yet heated by the element. When it has been heated, it too will move away

When you partition the heater off from the tank you contrain the fluid movement. This means you can stratify the tank more easily. When a bartender make a drink with a cream layer on top of something else, it's poured over a spoon to spread it out and slow its entry into the other liquid ensuring it floats from the get go. Adding hot water to the top of a tank via a small pipe, slowly (convection isn't a fast pump) doesn't stir it in with the tank water. An 18 inch long horizontal immersion in the tank is very different, and stirs the water more. The longer you run a Willis the more of the top portion gets heated. The benefit therefore is touted that you can choose some amount of water from 0% to 100% of the tank to be very hot just by the length of time you've left it to run

without helping me to understand what is wrong with my intuitive (I would have said physics-based) argument that 'it should not happen'.

Not sure why you assert that convection shouldn't happen. Convection is an observable thing. The reason why we put radiators under windows is the cooling falling air next to the window mixes with the heating air next to the radiator and pushes out in a mix into the room. If we put the radiator on the wall opposite the window the occupants of the room experience worse drafts and feel colder because there are heating/cooling areas doubly driving a convection cycle

If it were a pump, pushing water round a circuit, water leaves the output and goes into the input, and the cycle runs. You surely must accept that pumps exist and work

Heating water and using its resulting buoyancy to move it is just pumping it, but really slowly compared to mechanical pumps. In a Willis let's say no water molecule will pass the heater more than once in a heating cycle; we aren't circulating this water quickly by any stretch

In other posts you've queried why you don't just get warm water out of a Willis because 50% of the water is cold and goes through the immersion, and 50% hot comes out of the tank. It's not like that; it's hard to get water through a Willis heater quickly, they don't flow water at the same rate and you'll often see recommendations not to use them in hard water areas because they block up. If it's 100 times easier for water to flow straight through the tank than the heater then that tiny amount of water that does flow through the heater makes negligible different to the output temp