Thermostat fr underfloor heating with high thermal inertia

[bernardgreen]

[No engineer has ever been know to spend any time, or money (on computing, for example) to find a way to make things lighter and stronger at the same time, i.e. to try and do more with less...

Space craft, air ships and the Boeing dreamliner might suggest otherwise.

 

[ban-all-sheds]

Did the not register, Bernard?

 

[bernardgreen]

yes it did,

 

[JohnW2]

[No engineer has ever been know to spend any time, or money (on computing, for example) to find a way to make things lighter and stronger at the same time, i.e. to try and do more with less...

Space craft, air ships and the Boeing dreamliner might suggest otherwise.

I don't know about 'suggest otherwise'. I think that you and BAS are actually suggesting exactly the same thing, the difference being that you do so in straightforward language, whereas one has to do a bit of decoding of the irony to work out BAS's meaning

Kind Regards, John.

 

[ericmark]

I will admit with the heat treatment stopping the sine wave was hard which was one of the reasons for trying to use a PLC so reaction speed could be varied. However with the underfloor heating the time delay will be constant so once set it would work.

But how far does one go. Using SCARDA there is really no limitation on how many curves and corrections can be written in. However weather forecast seems to have eluded us for years so will never be perfect.

Using SCARDA one could of course alter things while still at work. The internet cloud is the limit. But this would not be to save money it would be a toy. Impress the neighbours yes save money no.

The heat put into the underfloor system can't be controlled once it is in the floor it needs controlling before that point. So the floor it's self can't be used as a store. Using liquid likely best idea and an insulated tank which holds the hot liquid heated during the night and circulates it only when heat is required. Theory one could use batteries but practice not really an option.

So step one is rip up all the electric heating wires and lay pipes instead. I am sure all can see the problem.

So down to earth you want warm floors. As you say simple simmerstat will adjust power into floor to keep them warm. You also want fast react stored heat system. Myson radiators and Megaflow or other hot liquid storage would seem way to go. With Myson there is no exposed radiator parts so no real need to limit the liquid temperature one could use oil at 200 degrees if you wanted. (I am not saying do that). The higher the temperature the smaller the reservoir required. Mysons because like a car heater are fan controlled are far smaller than normal radiators and easier to control. They have built in fan speed and thermostats and can be linked easy to a time switch.

Ups I said down to earth and Mysons also cost. And this is the problem what is cost effective? We can all dream up ways to do it but at what cost? Only jonnyhifi knows what his budget is and what with his house is cost effective yet still keep his feet warm.

 

[JohnW2]

However with the underfloor heating the time delay will be constant so once set it would work.

As I wrote earlier this evening, I don't think that's necessarily true - and I think that's one of the problems. The rate of heat flow from heating element through the floor (hence time delays) will presumably vary with changes in the temperature differential bewteen element and room - and that differential will itself be changed by the rate of heat flow. It's therefore a control loop with negative feedback and varying delays.

As for the down-to-earth pragmatic practicalities, I seriously doubt that any of the 'cleverness' being discussed here is appropropriate for the vast majority of domestic situations. Probably far more practical to simply regard the UFH as a low-level 'background'/'comfort' heating (supplemented by 'proper heating' of the room itself), and therefore to size the UFH installation so that, when left on 'continuously' (i.e. with no thermostat or control system, other than 'overheat protection') it achieves a reasonable fairly low level of heating, primarily just of the floor itself.

Kind Regards, John.

 

[ericmark]

I would expect the element temperature if just switched on (not pulsed) would be fairly constant, depending on type of course, Raychem type likely to be very constant and the difference between 0 - 20 degrees for room will alter the rate of heating agreed but only when it's approaching set temp say 18 degrees when set to 20 would the difference really be important so yes it will change but not enough to worry.

Using mark/space to adjust heat input would change the delay which is why I was talking about it so the delay would more match what is required.

But what would really mess it up is mark/space with Raychem type heater. This would mean the two systems would fight each other. Switching off and then back on would mean the power used when switch back on would increase.

Just can't see it being worth all the effort. Underfloor heating is only really any good when either it's left on 24/7 so only a gradually change in heat input is required or when it's not the main source of heat but just an extra.

Even with a standard water in radiators system one gets over shoot. The Myson does fair better in some cases, but that also has problems in that the thing cools down too quick. Using built in thermostat works OK but where there is also a wall thermostat rather than decrease the sine wave it increases it. As the standard wall thermostat is too slow.

The point is the money needed to make underfloor heating work would likely also pay for a proper system anyway. So why bother?

 

[ban-all-sheds]

So step one is rip up all the electric heating wires and lay pipes instead.

Wouldn't help much - whatever is used to do the heating there will still be the problem of a floor with a large thermal mass. OK - you might be able to heat it up faster with hot liquid in pipes, but it wouldn't cool any faster when you turned it off. Slower, I expect, because you'd have the added thermal store of the hot water.

I guess you could have motorised valves which would allow you to circulate cold water to cool it down, but that seems spectacularly loony.

 

[jonnyhifi]

the conversation continues in a fascinating way


It did strike me that one of the things that it would be prudent to do, would be to actually get some data points:

In particular, from stone cold, turn the heating on to maximum nuke: and log the temperature of the air in my lounge (the large area with the underfloor heating), and also log the outside air temperature (in case there were large "wobbles" outside) and geta sense for the... "inverse half life" as I guess I'm going to have a logarithmically increasing temperature, till it asymptotically reaches maximum. From this I could get a sense of wha tthe slope of hte temp increase graph is.

I really have no idea at the moment: historically I'd leave it on 24/7,when it started feeling al ittle overly hot, turn some of the circuits off (it was wired up with multiple circuits in teh screed originally).

I am aware this is rather akin to regulating the temperature in a cetnrally heated room, to opening the window if it gets a bit hot (as I witnessed in college rooms in my first year, where the heating system seemed to have no mechanism for room occupants to turn the rads down !?!?!).

Safe in teh knowledge a theoretical understandign of the problem exists, I may of course never actually install any form of solution -- but that's what comes of being a mathematicican I suppose

More seriously though: to log temperature: that again appears to be a non trivial task. Perusing ebay: it appears that £30 or so gets you a battery operated stand alone temperature logger, that can download by USB. so to do outside and inside is already going to cost be £60.. jsut to get a handle on the problem! god help me if I wanted to do more.
(e.g. the voltcraft dl100-T), and something similar at £34 is the cheapest i can find on fleebay at the moment...

Any cheap off the peg solutions out there for temp logging
(with ultiple sensors) anyone knows of... ?

I lovedthe link to the discussion of the amazing shower btw

 

[bernardgreen]

The logical way to improve matters would be accept the room has to be

[1] kept at the required temperature 24/7
[2] kept at a lower temperature and then bought up to the required temperature before the room is to be used. This requires advanced notice long enough for the heating system to respond and heat the room

Using internal and external temperatures and heat loss factors it is possible to calculate the amount of heat input needed to achieve and then maintain the internal temperature. Heat input woukd controlled by mark space control of the heating mat(s).

It would work but at the expense of a lot of measurements of heat loss for the range of internal and external temperatures it would be expected to cope with and the response time. Then there would inevitably be in service trimming of parameters.

 

[ban-all-sheds]

I've never looked to see if such a thing exists, but can you not get basic USB connected thermometers which you could plug in and write your own simple logging stuff - just a script to record the temperatures every minute (or whatever) into a flat file?

 

[JohnW2]

I would expect the element temperature if just switched on (not pulsed) would be fairly constant,

I've no idea what sort of temperature the elements run at (anyone got any idea?), so I can't do any sums, but the concept of what I said is true (although perhaps insignificant - that's why I'd need some numbers). With constant electrical power being supplied to the element, its temperature would obviously depend upon the rate of flow of heat away from it - which, is a function of the temperature gradient between element and room, through the floor - which would change as to room heated up.

Just can't see it being worth all the effort. Underfloor heating is only really any good when either it's left on 24/7 so only a gradually change in heat input is required or when it's not the main source of heat but just an extra.

Exactly - as I said in my previous post - at least in a domestic environment.

Kind Regards, John.

 

[jonnyhifi]

In fact I've just stumbled across a particulalry swanky looking "clever" thermostat http://www.warmup.co.uk/uk/3ie-programmable-thermostat.phtml

which appears to have 3 sensors: an in floor, in air, and outside stat: so it's truly applying some algorithm to improve on a simplistic thermostat.

I'm certainly going to get in touch and see if I can find out more....

I tracked down the thermostat I referred to earlier designed by the ipod designer tony fadell, who isinvolved in a startup company to produce it called nest:

http://www.nest.com/

a little of the story behind it:
http://www.dailymail.co.uk/sciencet...Tony-Fadell-unveils-new-Nest-Labs-gadget.html

both are very expensive devices: but then so are the likes of Honeywell controls (though not quite as eye watering...), or nearly as expensive as some kind of home brew device would be.

 

[jonnyhifi]

Having trawled throught the Nest website, I sense a similar presnetation triumph over substance as with the i"product"s that came before it.

When you get down to it, it seems to be little more than a programmable thermostat, that simply learns from when you put it on and off, combined with a movement sensor. Well: big deal ! To be fair it looks beautiful, and the programming may be easier, but to give helpful tips like "don't heat rooms when you are not in" and "save energy by turning your heating off when you leave the house for a long period" is hardly rocket science... !


The one thing you'd really want it to do, i.e. think about thermal inertia: I think (though I may of course be wrong), it doesn't adress... "as yet...".

So: it's NOT a smart thermostat at all