Underfloor Heating Control - Efficiently

[Mike Seaman]

Hello,

 

So I came on here a few months back asking about Underfloor Heating Controls. So getting in a bit of a muddle of how best to run the Underfloor System.

I have effectively 16 zones/rooms that have a Loxone Air Temperature/Humidity Sensor in each. Each zone has a separate Manifold Valve (On/Off) installed on a common underfloor heating manifold.

The manifold can be controlled to a temperature using a 4-way mixing valve from the boiler, I utilise a 1-wire temperature sensor clamped to the copper pipe on the mixing controller underfloor outlet. This valve controls the temperature to a setpoint, using a Mixing Controller and works well, i.e. the mixing controller works perfect, thus allowing you to tell it a temperature from boiler and it will control to that.

There is a separate pump for the underfloor system after the mixing controller, solely for the underfloor heating pipework.

In the next couple of weeks I will have a 1-wire temperature sensor installed outside, so I can measure external temperature.

 

I am currently using Intelligent Room Controllers for each room, central Intelligent Temperature Controller(Mixing Demand Temperature) and Climate Controller to turn Pump & Boiler on.

At the moment the boiler seems to be running a lot as I am running a lower temperature on the underfloor system due to the Intelligent Temperature Controller running at 25-33 DegC based on room demand. I manually had it running at 45 DegC, before I started using the temperature controller demand and this seemed to heat the rooms up quicker, but I didn’t think this was most efficient.

 

Does anyone have any experience of running underfloor heating systems and the best method of control, including setup within loxone. I understand from other folk that you don’t run it like a radiator system and effectively should be constantly running a temperature in the underfloor screed, however I do not want to run it hot all the time. I just don’t know here. Its all working, but just not very efficient at all.

 

Any help is greatly appreciated.

 

Thanks

Mike

[Duncan]

maximum efficiency is going to be achieved by using cheaper fuel (gas) plus the boiler in condensing mode as much as possible plus the temperature difference between in and out being as small as possible (turn down the set temperatures and wear more clothes) rather than the heating/cooling of the building fabric - if its kept hotter than necessary all day it will consume more energy, but if its too cold in the morning shower or evening when you are watching tv then efficienty isnt always the most overriding requirement for heating.

energy loss/required is the area under the curve of temperature difference between in/out and time - the bigger the temperature difference and the longer the time the more energy is lost/required.

as youve discovered already, response time may be as or more important and at the end of the day, comfort is the sole reason for heating so having the system set up to deliver the comfort you require is the priority. either strategy can be delivered by the hardware you have in place.

there is no right or wrong regarding how ufh is run - heating control can be quite simple or complex, partly depending on strategy:

1) european ufh is often run 24/7 at a fixed temperature during the heating season and uses very low manifold temperature with no manifold thermostatic control and the manifold water temperature is controlled by the boiler or output of a buffer tank - external temperture compensation is then done at the outflow of the buffer tank or at the boiler level  - however these buildings work best if you have a secondary source of heating for the evenings such as a fire/log burner which add additional quick heat and a focus to a room.

2) uk ufh tends to use a variable room temperature more like a conventional radiator system, which then requires a quicker warm-up time and therefore a higher manifold temperature - external weather compensation (if used at all) is required to be at the manifold level and helps to keep the warm-up time short even when its colder outside, and reduces overheating when its warmer outside by reducing the manifold temperature.

the ITC block increases manifold temperature when there is more demand - this makes sense when a lot of zones are run off a manifold and/or the flow is limited, but with good hydraulic design and plenty of capacity at the manifold(s) this will result in variable heating up times and with a higher temperature manifold is more likely to get overshoot/overheating and best suits the european constant model.

loxone has a separate mixing valve controller which only adjusts its target temperature according to your chosen weather compensation at manifold level (if you bother to impelement it) but not the number of rooms demanding heat, and may give you better response and temperature control with a variable temperature setup and quicker warm-up times

on a related note, weather compensation requires not only the external temperature, but also the wind-chill effect (perceived temperature in the loxone weather service) as the resulting air changes in windy weather significantly increase heat loss.

of course being loxone you can reconfigure from one strategy to another quite easily and see which you are happiest with to live with

for either to work well you should get good hydraulic balancing of the various manifold and zone flow rates with all zones open fully then allow the control systems to modulate it.

as each ufh system is unique and has different responses, it may be necessary to tweak the control constants Ki and Kp of the mixing valve controller for each manifold to optimise the temperature rise rate without too much oscillation - the generic defaults of the mixing valve controller wont always work well -  this is something that loxone can do that most other ufh control systems cant - this is very advanced control theory stuff and requires you to get the hydraulics and flow temperatures right first and

[Simon Still]

On Tuesday, 8 January 2019 16:07:03 UTC, Duncan wrote:

on a related note, weather compensation requires not only the external temperature, but also the wind-chill effect (perceived temperature in the loxone weather service) as the resulting air changes in windy weather significantly increase heat loss.

That's an interesting point.  Presumably not relevant in an airtight house with controlled ventilation (not the case for all but a very few UK houses).

[Rob_in]

Hi,

What sort of control/monitoring have you got from Loxone to the boiler? I think you are saying you can control the water temperature (via mixing - see below), but what about the flow? If you know the leaving water temperature, plus the inlet (return) temperature plus the flow rate you can calculate the actual power being put into the floor. You can correlate this with your fuel usage and work out boiler efficiency and put some figured on your hunch things could be better.

You also say, "...the boiler seems to be running a lot as I am running a lower temperature on the underfloor system due to the Intelligent Temperature Controller running at 25-33...". This seems a reasonable temperature for a modern, well insulated house (we use 25-29 limits). We have a heat pump though (Daikin Altherma) which are designed to be most efficient (best COP) at low temperatures. As you have said your boiler has a setpoint higher than the water outlet and you are using a mixer this implies you are heating water up just to cool it down again. This doesn't sound logical or very efficient to me. Can you somehow control the water temperature directly without having to mix it down?

What room temperature differences are you using for active/away? Allegedly it's more effective to not let the temperature drop too far and then have to heat up a lot when it's in use. This is allegedly more relevant with UFH when the heating medium (your floor) is very slow to respond. I say, 'allegedly' because just read an article where they say British Gas recommend only heating rooms based on use.

Cheers,

Robin

[Rob_in]

Actually, I think it should apply even then. If you think about it then windchill is basically blasting a surface with cold air and sucking heat away faster than if it were just sat there without the wind.

Our house is airtight with a HRV system. If only we had a windspeed detector to quantify this as it's often really windy here! ;)

Cheers,

Robin

[Viktor Granbom]

The chill factor is interesting. To save energy I have set up our heat pump with outdoor compensation and compensation for the temperature 12 hours ahead but maybe I should use the perceived temperature for compensation instead...

[Simon Still]

On Wednesday, 9 January 2019 18:48:14 UTC, Rob_in wrote:

mperature, but also the wind-chill effect (perceived temperature in the loxone weather service) as the resulting air changes in windy weather significantly increase heat loss.

That's an interesting point.  Presumably not relevant in an airtight house with controlled ventilation (not the case for all but a very few UK houses).

Actually, I think it should apply even then. If you think about it then windchill is basically blasting a surface with cold air and sucking heat away faster than if it were just sat there without the wind.

I'm not sure - an external wall has unrestricted convection anyway.  I don't think it would make much difference.  Your number of internal air changes per hour wont change based on windspeed. Whereas a non-airtight house relying on passive ventilation through trickle vents and gaps in the structure will have far more on a windy day than a still day.

[Simon Still]

On Wednesday, 9 January 2019 18:41:13 UTC, Rob_in wrote:

Allegedly it's more effective to not let the temperature drop too far and then have to heat up a lot when it's in use. This is allegedly more relevant with UFH when the heating medium (your floor) is very slow to respond. I say, 'allegedly' because just read an article where they say British Gas recommend only heating rooms based on use.

I've concluded this is nonsense.  

"boiler in condensing mode as much as possible plus the temperature difference between in and out being as small as possible (turn down the set temperatures and wear more clothes) rather than the heating/cooling of the building fabric - if its kept hotter than necessary all day it will consume more energy,"

In a poorly insulated house you want a fast reacting heating system and to heat the rooms you're using when you're using them. You may overshoot the target temperature (inefficient) but it won't be for long and the room will quickly cool back to target so it's not much of an issue. 

In a very well insulated house you overshooting target is more of issue so lower temperature emitters and a different approach make sense. "If its kept hotter than necessary all day it will consume more energy," still applies but the 'set back' temp will be very close to the occupied temp. The savings to be had are comparatively small as well - I think i worked out that my space heating costs are only about 20% of my total energy bills.

[Rob_in]

Hi,

On a related note... is the IRC supposed to 'learn' how long a room takes to heat up and cool down, watch the current temperature trend in the room and adjust demand accordingly?

I notice that IRC demand always appears to be zero until a minimum room temperature is reached. With UFH this is already at least an hour, probably 2 hours too late given how long UFH takes to heat up.

I have already implemented weather forecasts and linked that to forecast the internal house temperature so know a few hours in advance when heating will be required. It's quite frustrating then how it always seems to be a tad too cold most recent mornings.

I could connect the IRC room temperature to the predicted temperature 2 hours away which will get the heating to come on early, but then am still considering how one should modulate/mix the predicted temperature with the real one to get the heating to go off at the right time. The IRC already generally overshoots the room target temperature because it doesn't understand after putting several KWh of heat into the floor, that heat will be released into the room over the following few hours.

I get this overshoot is supposed to be catered for with weather dependant heating curves and if your curve is super accurate the system should always be putting the same amount of heat into the floor as will escape from the fabric of the building given the outside temperature. One still needs some kind of room thermostat though as other activities such as cooking, taking a shower or even working out can easily add too much heat in a well insulated building.

It also doesn't help that our boiler will produce water at 25C minimum. This doesn't sound like much, but if the floor was at 25C constantly we would cook.

Maybe I should use a heat curve which in chilly, but not super cold weather imagine would predict temp in the floor should be 22-23C. Then use some kind of PWM on the boiler to try and keep the floor there?

Sorry, just thinking out loud here ;)

Then there is the fact that if a sunny day happens to be forecast it's pointless turning the heating on at 8am for example, because when the sun comes up we get a boat load of solar gain that gets the place toasty much quicker than the heating would. I have a version of this already working and it's pretty good, but mention it as is clearly a factor.

I'm curious as to if anyone had any joy with trying to predict what their UFH should be doing like this or if we are just supposed to trust the IRC.

Cheers,

Robin

[Simon Still]

If you have a well insulated house then look at Duncan's "european" approach above.  That's how my heating operates (using Viessmann's own weather compensation system).  

Ours runs about 4am to 10pm. I tweaked the heating curve through the first winter but haven't touched it for the last two.  From memory it runs at minumum 21C and switches off completely if it's above 13C outside.  On the concrete floors  the flow termperature out (at the boiler, so not the floor temp) gets to  29C when it's 0C outside. 

There are no internal thermostats at all - just the external.  Since the floor temperature is usually only a few degrees above room temperature you simply don't get any overheating problems (remember heat transfer falls as the delta between floor and room gets smaller).  

We dont have any extra heat sources either -  the termperature is constant and comfortable.  Sometimes  I might put on a thin cardigan in the evening.

[Simon Still]

On Tuesday, 15 January 2019 15:00:25 UTC, Simon Still wrote:

If you have a well insulated house then look at Duncan's "european" approach above.  That's how my heating operates (using Viessmann's own weather compensation system).  

I'd add that Duncan doesn't seem to think this strategy works but I'm very happy with it. 

"however these buildings work best if you have a secondary source of heating for the evenings such as a fire/log burner which add additional quick heat and a focus to a room."| 

I would also heavily advise against taking this advice.  I've spent time in 3 other modern insulated houses with UFH which have, a closeable fireplace (sliding door) in Switzerland, a traditional open fireplace and a modern freestanding log burner (second two in the UK).  Using the secondary heat source in all 3 resulted in massive overheating. The uncle with the open fire insists on using it (and we sit there in t-shirts sweating and fighting dehydration).  The other two never get used.

I think the European log burner thing is a hangover from when insulation and airtightness standards were worse.  You might conceivably want to run the house at 19C or 20C during the day and a few degrees warmer in the evening but any form of wood burner will simply put out too much heat and isn't controllable.  If you did want to do this you want a source that's very quick to heat up and cool down - an elec fan heater would be ideal.  

 

[Rob_in]

On Tuesday, 15 January 2019 16:13:30 UTC+1, Simon Still wrote:

On Tuesday, 15 January 2019 15:00:25 UTC, Simon Still wrote:

If you have a well insulated house then look at Duncan's "european" approach above.  That's how my heating operates (using Viessmann's own weather compensation system).  

I'd add that Duncan doesn't seem to think this strategy works but I'm very happy with it.

I can tell you that as described that wouldn't work for us because when it's sunny we get a massive amount of heat from solar gain and don't need the heating at all.

And to a lesser extent, if we have friends over for dinner the added number of bodies also has an impact.

For those reasons we have to rely on room thermostats to turn down/off the UFH flow when the rooms are close to the required temperature.

Cheers,

Robin

[Duncan]

unfortunately any heating system is going to struggle with the combination of large thermal mass (ufh in concrete), high insulation levels, low air leak and a large secondary heat source such as solar gain where you have little control or predictability

to avoid overheating due to the sun, the control system needs access to a system to remove or redistribute the extra heat - unfortunately air conditioning isnt usually a feature of uk homes.

have you tried to use a custom mode set to lowered temperature (say 1-2deg lower than comfort) and triggered by sunshine/brightness flag to turn down your thermostats from the normal comfort temperature ? it  might give you better overall control with a degree of automation.

[Rob_in]

On Wednesday, 16 January 2019 16:20:34 UTC+1, Duncan wrote:

have you tried to use a custom mode set to lowered temperature (say 1-2deg lower than comfort) and triggered by sunshine flag to turn down your thermostats from the normal comfort temperature ? it  might give you better overall control with a degree of automation.

Yes. In fact that is pretty much exactly what we do. There is an 'eco' temperature that is active if sunshine is expected (uses the weather forecast). It works well as in winter I don't really care about overheat and just want to prevent the heating coming on needlessly in the time between us getting up to the sun actually heating us up for free.

Just circulating the UFH water (with no boiler heat input of course) does an OK-ish job of spreading heat around. I will think about also turning the ventilation up a bit to see if that can help.

Cheers,

Robin

[Simon Still]

On Wednesday, 16 January 2019 14:48:54 UTC, Rob_in wrote:

I'd add that Duncan doesn't seem to think this strategy works but I'm very happy with it.

I can tell you that as described that wouldn't work for us because when it's sunny we get a massive amount of heat from solar gain and don't need the heating at all.

And to a lesser extent, if we have friends over for dinner the added number of bodies also has an impact.

And I'm telling you that it works in practice and it will likely work much better for you, with less overshoot, than running the floor at a higher temperature on a room stat and trying to predict when you're going to get solar gain. 

Our kitchen has full width, full height glass and a partially glazed roof - we get solar gain too.  it simply doesn't matter if you're running the floor close to room temperature - you don't need to the floor off because it's adding very little heat to the room once the delta between floor and room gets small (in fact, if the room overheats it's potentially cooling it - exactly as you mention in your post below).   At 5C my flow temperature is probably only 24C, the floor a few degrees less(?).   

You say you're running your floor at 25-29 limits - I think that's too high. Is there no way of reducing/mixing down your flow temperature further?  I'm probably only hitting a flow temperature of 25C at external temperatures below 5C. 

If you are running your floor at, say, 28, when it's switched on it will take many hours to cool down after the call for heat ends when you have solar gain or additional heat sources.  

Of course the big benefit of Loxone is you can completely change your heating strategy in software so you could try a low and long approach and then switch back. 

[Duncan]

some light reading to give some ideas regarding this:

https://arrow.dit.ie/cgi/viewcontent.cgi?referer=https://www.google.co.uk/&httpsredir=1&article=1021&context=sdar

[Simon Still]

Thanks. I’d be interested to know the airtightness and insulation standards for the buildings they were looking at - certainly wind chill impact is going to depend to quite a large degree on airtightness. Also, the heat emitters used and the range of flow temperatures..

This weekend was a fairly good test of my house. Yesterday was cold (around 4C but damp, dull and wet), today is 2.6C outside but clear and sunny.

My internal target temp is 20-21C - that’s comfortable in a house with no drafts wearing a t-shirt under a light shirt. This morning the kitchen is at 20.6. Last night we had visitors and I cooked using a couple of open pans on the hob and had the oven on (so extra body heat and a lot of extra heat from cooking - made risotto so a large open pan on a fairly high heat for best part of an hour). Recirculating hood so no extract of heat to outside. The room peaked at 21.5 when I finished cooking.