Heating and cooling: Self-regulating UFH: how to control mean water temperature of flow and return

1 to 30 of 70

- CommentAuthorShevek
- CommentTimeJun 26th 2014 edited
I'm trying to design (or get my head around) our UFH control system, which is intended to be self-regulating, by controlling the mean water temperature of flow and return as described in this thread:
http://www.greenbuildingforum.co.uk/forum114/comments.php?DiscussionID=10244&page=1

But how exactly do we do this? Do we simply set the supply temperature using our boiler control (to give us the right room temperature) and the leave it? Or is control of the mean flow and return more complicated than this?

We're fitting an Atag 325ECX, which has its own pump and controls for supply temperature. Some text from installation manual below.

The boiler has a self-adjusting and self-protecting control system for the load. This involves checking the temperature difference between the supply and return water. Table 3 shows the water displacement a circulation pump can deliver for a particular installation resistance.

A325ECX (from Table 3)
Pump type = UP 15-60
Water flow rate at delta-T 20 degrees = 18 litres/min
Permissible installation resistance = 20 kPa / 200 mbar

If the installation resistance is higher than the value stated, the control system will adjust the load until a temperature difference between supply and return water is reached that is acceptable for the control system.
When the temperature difference still remains too high the boiler will switch itself off and wait until the high temperature differential between the flow and return water has decreased again.

The control system will, if an unacceptable temperature difference is detected, repeatedly try to establish a water flow. If this does not succeed, the boiler will block itself.

The boiler is equipped with an intelligent control system: CMS (Control Management System). Each boiler anticipates the heat demand of the heating installation or hot water demand. This will cause the boiler to tune its capacity to the installation. This means that the boiler will function longer and at a low level.
If an outside sensor is connected, the control system is able to function, depending on the weather. This means that the control system measures the outside temperature and the supply water temperature. On the basis of this information the control system calculates optimum supply water temperature required for the installation.

CH controls
With a demand from the heating controls, the boiler activates its 1 minute delay period. This is to prevent the heat exchanger from losing its heat too quickly in the event of a hot water demand. Then the pump starts and after 30 seconds the gradient control becomes active. The starting point of the gradient control is the currently existing supply temperature. A Delta-T control (25K) ensures a stable control according to heat request. If the supply temperature is below the T-set value of 20Â°C the boiler will immediately start.

In case of weather-dependent controls (1kOhm outside ARZ0055U sensor connected) a daily temperature is set instead of a supply temperature. Control will take place according to the heating line.

The ATAG A is equipped with boiler sensors of 10kOhm. The resistance value and corresponding temperature are shown in the accompanying table.
- CommentAuthortony
- CommentTimeJun 27th 2014
You will need more control than flow temperature, manifolds, possibly a secondary pomp and mixing circuit, room thermostats or thermostats of some kind.

The boiler thermostat will control boiler firing but this is not enough control for the heating.

How many zones? how will you balance these and balance out solar gain, occupancy, open windows, additional heating appliances?
- CommentAuthorShevek
- CommentTimeJun 27th 2014
Tony, I'm trying to do it without room thermostats and all that, as described in this thread:
http://www.greenbuildingforum.co.uk/forum114/comments.php?DiscussionID=10244&page=1
- CommentAuthorDantenz
- CommentTimeJun 27th 2014
Think all the above taken from Atag's installation manual is a long way of saying that the boiler will simply modulate the burner output and the pump speed in order to maintain flow/return temperature differential - weather compensation will automatically adjust the boiler flow temperature rather than the CH control knob on the front; pretty standard stuff these days for modern gas boilers - load compensation would also be good though which I'm sure Atag also provide.
- CommentAuthorjoe90
- CommentTimeJun 27th 2014
Not sure if its relevant but I am sure that someone on this forum stated that if you have the UFH water temp set at 1 or 2 degrees above the temp you require the house to be it will self regulate. If true this is the kind of KISS approach I like (I am sure this works best in a high thermal mass house with constant temps).
- CommentAuthorShevek
- CommentTimeJun 27th 2014
Yeah it's all in the thread I posted above.
- CommentAuthorVictorianeco
- CommentTimeJun 28th 2014
That 1deg above room temp would be great if the house was super insulated. In my house, renovated eco I think I will need to be running at 45c. Or do you mean the floor temp as opposed to flow temp?
- CommentAuthorbarney
- CommentTimeJun 28th 2014
There will be a temperature drop from the flow to return (if you want to transfer energy) - so the UFH has a mean water temerature that is passing heat to the floor slab - depending on what thickness of "slab" you have, then that will set the surface temperature of the floor (which must be less than the MWT). Then you will get a further temperature drop across any floor covering. Eventually you arrive at a surface temperature - which is then passing heat via principally radiation and convection to the room.

If you can get the rate of energy input into the slab to match the rate of loss allowing for all the variables, then you have a totally self regulating system.

So locking the slab temperature by means of varying either the mass flow rate of energy into the slab (variable volume) or varying the temperature difference between the fluid and the slab (variable temperature) will give you a very simple control solution as the rate of energy released from the slab is effectively locked to the room gains and losses. (if the room has an effective temperature the same as the slab, no gains or losses are taking place between the room and the slab - as that temperature diverges, the the slab is either losing or gaining heat from the room, at an increasing rate locked to the temperature difference of the room and the slab)

It can be done even if you have high heat losses - but the thermal inertia of such a system often means that the energy flow into the slab and the requirements for comfort are significantly out of step.

Regards

Barney
- CommentAuthordjh
- CommentTimeJun 28th 2014
Posted By: ShevekI'm trying to design (or get my head around) our UFH control system, which is intended to be self-regulating, by controlling the mean water temperature of flow and return as described in this thread:
http://www.greenbuildingforum.co.uk/forum114/comments.php?DiscussionID=10244&page=1

But how exactly do we do this? Do we simply set the supply temperature using our boiler control (to give us the right room temperature) and the leave it? Or is control of the mean flow and return more complicated than this?

I think as Barney says, the key is controlling the slab temperature so the short answer is that it is more complicated than controlling the supply temperature. The whole point is that you need to vary either the supply temperature or the flowrate (or both) in order to keep the slab temperature constant.

I confess I haven't reread that 3-page thread and I haven't ever thought much about the control system as I'm not using UFH so I can't offer much constructive help. Perhaps others have given it more thought?
- CommentAuthorSteamyTea
- CommentTimeJun 28th 2014 edited
There is a third variable, time.
You can pump at the same temperature and mass flow rates, but for longer or shorter periods.
- CommentAuthorbarney
- CommentTimeJun 28th 2014
Yes ST, time is also a factor - but usually (in a domestic) comfort temps are required for occupancy time - which can be highly unpredictable - and any long lag system generally can't respond to that variable fast enough.

Night set back of the slab temperature can work well - but to a large extent that is linked to having a good thermal reserve in the room and a low rate of heat loss.

I've done it once on a slab controlled system in a school (or actually on half a school) - it didn't produce much in the way of energy saving in practice.

Over a long time cycle then I would asume that no heat is applied to the slab when not generally required - ie turn the system off in the summer

Regards

Barney
- CommentAuthorEd Davies
- CommentTimeJun 28th 2014
Is what's needed just a thermostat which can be used in place of a normal room thermostat but which operates off the mean of the flow and return temperatures from the slab (being used as a proxy for the slab temperature)?

Does a boiler like this expect a simple on/off room thermostat or does it use a temperature sensor so it can modulate its power input to the room depending on the discrepancy between the actual and set-point temperatures?
- CommentAuthorbarney
- CommentTimeJun 28th 2014
That's what the boiler does I think Ed - it just provides a constant delta T to flow and return, modulating from the return temp.

So, presuming both flow and return temps can be preset and the required delta set, then MWT is known - what that needs to be in relation to a particular set up of slab (ie the offset) shouldn't be too difficult to determine.

From there, you probably don't need any further inputs other than "off".

As the slab loses heat to the space, the return temp drops so the boiler will modulate up to increase both flow and temperature into the slab until the return reaches set point - the slab is now recharged and if it's at temperature.

I think I's still want a handful of floor sensors that are averaging and providing an input to the boiler modulation personally.

I'd wonder if a small (ish) buffer vessel with constant boiler input at optimum condensing temperatures (nice and simple) and a mixing supply to the UFH controlled via a slab sensor array mightn't be all together simpler

Regards

Barney
- CommentAuthorEd Davies
- CommentTimeJun 28th 2014 edited
Posted By: barneyThat's what the boiler does I think Ed - it just provides a constant delta T to flow and return, modulating from the return temp.
But isn't delta T something different: Î”T = (Tf - Tr) rather than mean temperature Tm = (Tf + Tr)/2?

Î”T (in combination with the flow rate and the specific heat capacity of the water) sets the power input to the house. Tm measures the current slab temperature.

I'd wonder if a small (ish) buffer vessel with constant boiler input at optimum condensing temperatures (nice and simple) and a mixing supply to the UFH controlled via a slab sensor array mightn't be all together simpler
I can't see how the buffer vessel will ever usefully be cooler than the slab so won't the slab return always be better for condensing? Is a higher temperature return to the boiler ever better? If not then adding a buffer tank seems a rather Victorian way of dealing with a simple control problem.

Any sensor array will not cover the slab as thoroughly as the pipes. It will have the advantage over using the water temperatures of being able to measure the temperature of the actual slab away from the pipes but, other than that, I don't see the benefit.

Actually, won't just the return temperature be a better measure of the slab temperature and therefore be the best variable to control?

Ideally something like Tr - kÎ”T = Tr - k(Tf - Tr) = ~~(1-k)Tr - kTf~~ (1+k)Tr - kTf where k is a tuning parameter ~~in the range 0 to 1~~ and is likely quite small. Ie., estimate the slab temperature by looking at the return temperature then subtracting off an estimate of the amount the return is increased over the slab by how much more the flow temperature is.

Sorry Shevek, getting a bit carried away there.

(PS: getting even further carried away, I think you'd also want to take the flow rate into account. When the flow rate is high the power being delivered for each degree drop between flow and return temperatures is greater so the difference between those temperatures and the slab temperature would also be larger.)
- CommentAuthorbarney
- CommentTimeJun 28th 2014
But if the boiler sees return and sets flow to keep a constant delta T then MWT is known

I was thinking of buffering to be able to simply control flow, return and MWT without having to consider the boiler modulation (or even the capacity of the boiler to modulate at all - say an oil boiler)

something like Tr - kÎ”T = Tr - k(Tf - Tr) = (1-k)Tr - kTf where k is a tuning parameter in the range 0 to 1

Yep - but again a direct measure of Tslab and a suitable offset to Tflow and fixed delta T would do the same thing as the offset or K value should be a constant - and to some extent easier to do by mixing than modulating

Regards

Barney
- CommentAuthorShevek
- CommentTimeJun 28th 2014
Heh, no fire away Ed. I'm trying to keep up.

Here are excerpts from the manual explaining controls and the parameters that can be set (I've only just noticed them in the manual). There are controls there to read flow and return water temp but I can't see anything in there for pre-setting them.
https://www.dropbox.com/s/wokxk0ghytgbjw2/controls.pdf
http://www.dropbox.com/s/cdltq37xqfp4b6d/parameters.pdf

Just occurred to me too that the UFH pipes in the kitchen/dining will be set in a screed (+ a 100-150 mm slab), while the pipes in the living room will simply sit on top a 150 mm slab with battened timber flooring over top. So mean flow and return might work well to control the kitchen/dining but maybe not so well for the living room? Or at least that they'll behave differently?
- CommentAuthorEd Davies
- CommentTimeJun 28th 2014
Posted By: barneyBut if the boiler sees return and sets flow to keep a constant delta T then MWT is known
Surely the MWT will vary as the return temperature changes? Perhaps I don't understand what you mean by delta T. As noted above, I'm using it to mean the temperature drop of the water flowing through the slab; that is, the flow temperature minus the return temperature. Do you mean that or something else?
- CommentAuthoralec
- CommentTimeJun 28th 2014
you absolutely must use Atags own compensation controller, the wyze or the Brain..set a maximum flow temperature and the controller does the rest...

thats because compensation controls vary the temperature of the flow according to demand. parameters in the controller allow you to specify UFH...and the relationship between flow temperature and temperature difference between set point and actual temperature...
- CommentAuthorShevek
- CommentTimeJun 28th 2014
Alec, are you saying we have to use weather compensation? And/or are you saying the Wize or the Brain will let us do what we want to do and control heat and flow via flow and return temperatures?
- CommentAuthoralec
- CommentTimeJun 28th 2014
yes, with the wyse or the brain will deliver what you want, but what determines the flow temp is the deviation from set point... the boiler runs at any temperature to maintain the set point, but with UFH you should set the maximum flow temp to 50c or there abouts.

Its more like if the ambient temp is 21c the floor surface will be about 22c..what happens in the pipes doesn't really matter..but won't exceed the max flow temp
- CommentAuthorEd Davies
- CommentTimeJun 28th 2014 edited
Posted By: alecyes, â€¦
Is that â€œyesâ€œ you have to use weather compensation or â€œyesâ€ they'd let Shevek and co do what they want to do and control via flow and return temperatures?

â€¦ what determines the flow temp is the deviation from set point
Deviation of what? A temperature, I assume. But the temperature where? The return temperature?
- CommentAuthoralec
- CommentTimeJul 3rd 2014 edited
No you don't have to use weather comp with those controls (but why wouldn't you?)

flow temp, and hence return temp is determined by the the demanded heat in the room. in other words if the set point is 20, and the room is 18c the flow temp is higher than if the actual room temp is 19.5.

Indeed at 19.5 I would expect the flow to be around 21-22c..depending on the thermal resistivity of the floor covering...

I am not sure why any one would want to control anything on flow and return temps, its room temperature that determines comfort...
- CommentAuthordjh
- CommentTimeJul 3rd 2014
Posted By: alecI am not sure why any one would want to control anything on flow and return temps, its room temperature that determines comfort...

It sounds like you haven't read/accepted/bought into the whole basis of the control mechanism being discussed. Until/unless you do, I think people will be talking past one another instead of having a useful discussion.
- CommentAuthoralec
- CommentTimeJul 4th 2014 edited
Not sure what you mean by that, I understand and install UFH systems that are self regulating all the time, taking the air temperature as the reference point for the boiler.

What people want to do is limited by whats on the market, the research has been done the products are proven as are there energy saving credentials...now its just for us in the UK to accept these things, and stop trying to reinvent the wheel!
- CommentAuthoralec
- CommentTimeJul 4th 2014 edited
As said installing floors to be self regulating is part of my job, you need to vary the heat in put to the floor relative to target room temperature. The rest will occur as a result of physics and the controls!

If you want a low flow temp avoid floor coverings with high thermal resistivity... and let the brain or wyse do the work...

Its simple really!
- CommentAuthoralec
- CommentTimeJul 4th 2014 edited
for multiple floor coverings you need variable electronic mixing valves for the lower temp circuits..
- CommentAuthorEd Davies
- CommentTimeJul 4th 2014 edited
Posted By: alecNot sure what you mean by that, I understand and install UFH systems that are self regulating all the time, taking the air temperature as the reference point for the boiler.
No, that is not â€œself regulatingâ€ as the term is being used in this thread where it means controlling the slab temperature directly so that the heat flow from the slab in-turn is regulated by the actual demand of the room.

No you don't have to use weather comp with those controls (but why wouldn't you?)
Because in a well-insulated high-thermal mass house the external weather is a secondary concern compared with the current temperature(s) of the thermal mass and incidental gains. Of course it'll have a long-term effect on the heat input required but in the short term the other factors are more important and adjusting the heat input to the slab depending on weather is likely to have confusing and detrimental effects.

I wonder if the best â€œweather compensationâ€ would be to look at the forecast temperatures over, say, the next 48 hours and use them to tweak the target slab temperature slightly. Or just leave it to the user to tweak it on a seasonal basis.

I am not sure why any one would want to control anything on flow and return temps, its room temperature that determines comfort...
Radiant temperature is about as important as air temperature in determining comfort. More significantly, though, with a high thermal mass and varying incidental gains looking at the air temperature directly is likely to result in wild and wasteful swings in the heat input to the slab with little or no contribution to comfort.
- CommentAuthorEd Davies
- CommentTimeJul 4th 2014
Cartoon version of Shevek and co's house: 5 m Ã— 9 m Ã— 7.2 m high, U=0.15, with a 275 mm slab across whole of the ground floor, density 2500 kg/mÂ³, specific heat capacity 1 kJ/(kgÂ·K):

Area = 5*9*2 + (5+9)*2*7.2 = 291.6 ~= 300 mÂ².
Heat loss = 300 * 0.15 = 45 W/K.

Slab volume = 5 * 9 * 0.275 ~= 12.5 mÂ³
Slab heat capacity = 12.5 * 2500 * 1000 = 31.25 MJ/K

If the slab's 1 Â°C cooler than you want it and you want to be on track to fix that within two hours (7200 seconds) then the extra heat input required will be = 31.25e6/7200 ~= 4.3 kW.

If the outside is freezing rather than at 10 Â°C then the extra heating required is 45 * 10 = 0.450 kW.

Ie, bringing the slab back up from a minor set back (say overnight) is close to 10 times more significant for boiler control than typical/non-extreme English winter temperature swings.
- CommentAuthoralec
- CommentTimeJul 4th 2014
well boilers do need some sort of control or they would get very hot!

The floor is self regulating because it only needs to be 1c above target rome temp to "do its stuff" its a question of integrating necessary electronic controls withe the slab, and variable flow temps is probable the only way to do it.

the thermal time lag of the building means weather comp does not have to be predictive.. the best weather comoensation controls average the outside temp over time.. anything up to 48 hours I think..viessmanns by default is 23.7 hours...

also the higher the level of insulation, the less relevance the out door temperature...
- CommentAuthorShevek
- CommentTimeJul 11th 2014 edited
Do you think if I went to the likes of John Guest they would understand this method of control if I went to them for a quote/design/parts lists?

1 to 30 of 70



	Site Map \| Home \| View Cart \| Pressroom \| Business \| Links	Contact Us
Logout
© Green Building Press

Green Building Forum

Categories

Feeds

Heating and cooling: Self-regulating UFH: how to control mean water temperature of flow and return