Advanced weather compensation internal temperature monitoring

[SBMS]

I was reading heatgeek.com and they mentioned that many heat pumps with advanced weather monitoring don’t need an internal thermostat or temperature probe to work out the internal temperature for modulating the heat pump:

 

“Advanced weather compensation can obtain accurate room temperatures without any internal reference once you've dialled your weather compensation curve in. 

This means it continually runs rather than being flicked on and off by a thermostat, which minimises flow temperatures and cycling which of course maximises efficiency. “

 

Can anyone explain how it does this? Does it look at return temperature heat loss or something to do this?

[ReedRichards]

There's some good sense mixed with a fair bit of tosh on Heatgeek.com.  What they call "advanced weather compensation" is Weather Compensation combined with Load Compensation.  Load Compensation modulates the heat output according to the difference between the set temperature on the room thermostat and the actual temperature in the room.  But this requires that your boiler or heat pump knows what the room temperature is.  You could never infer the internal temperature from the return water temperature unless you got rid of all heating zones, TRVs (or the UFH equivalent), never cooked, never lit a fire, never opened a window and never left the building.   

[Beau]

Our GSHP has this. You set it up manually and adjust the heat curve until you obtain steady internal temps. How it technically does it I couldn't say for sure but it can see the outgoing and return temps on the UFH and knows the external temp. Ours will keep the house within 1 degree C unless I run the wood burner 

[SBMS]

2 hours ago, Beau said:

Our GSHP has this. You set it up manually and adjust the heat curve until you obtain steady internal temps. How it technically does it I couldn't say for sure but it can see the outgoing and return temps on the UFH and knows the external temp. Ours will keep the house within 1 degree C unless I run the wood burner 

So you don’t have any internal

room stats?

[Tiny]

Hi,

We have a Viessman W200 system boiler and 300Lt UVC, with weather compensation, and priority hot water, with UFH on the ground floor, UFH in the basement, towel rails upstairs in bathrooms. No thermostats in the building. The boiler controls the pumps and flow temps on the three circuits/zones. It keeps each zone to within one degree of the set temp. No big swings in temp up or down. We removed the pumps and mixers from the UFH manifolds. The boiler has touchscreen controls and an App for remote access so long as you have WiFi in the house it can access. It works really well.

[Beau]

1 hour ago, SBMS said:

So you don’t have any internal

room stats?

There is a thermometer fitted but it has no influence over the GSHPs running. You can alter how much influence it has but I was told the system would be most efficient with it run 100% from the outside thermometer and the weather compensation so I had it set that way at the start and have never changed it in 10years

Edited April 10, 2022 by Beau

[ReedRichards]

There is some "magical thinking" going on here.  If you want the same temperature inside your house 24/7 then in principle you should be able to tweak your heat pump/boiler to supply water at just the right temperature to achieve your desired temperature when it is, for example, 0 C outside and then with accurate Weather Compensation the heat pump/boiler can adjust the water flow temperature to maintain your desired indoor temperature as the outside temperature varies.  If you light the log burner then I guess in principle the extra internal heat source will cause the return water temperature to increase "unexpectedly" and the boiler/heat pump could infer that something else is heating the inside of the house and reduce its output to compensate.  I'm not aware of any boiler/heat pump that actually does that but I guess in principle it is possible and perhaps some do.

 

The above only works with a lot of initial tweaking.  If you can tell your controller that you are too hot or too cold then it may be capable of doing this.  If you tell your controller a desired internal temperature then there must be an internal temperature sensor, even if you are not aware of it.

 

You also have to sacrifice the use of controls like TRVs and you must want your dwelling to be at the same temperature 24/7 or, if you are lucky, the controller might offer you a set-back temperature option.

 

Since I got my first programmable room thermostat in 1998 I have found that what works for me is a nighttime set-back and a daytime temperature that increases as the day goes by.  I tend to become more sedentary towards the evening and need the house warmer than in the morning.  So the type of control I have described above does not work for me.  A better option is Load Compensation where the boiler/heat pump modulates its output according to the difference between the desired room temperature and the actual room temperature.  That gives you much more direct control over what is going on.  Many gas boilers and a few heat pumps have this capability.             

[TerryE]

13 hours ago, SBMS said:

“Advanced weather compensation can obtain accurate room temperatures without any internal reference once you've dialled your weather compensation curve in ..."  Can anyone explain how it does this?

 

Sounds like a load of marketing bollocks to me.  Overall any house acts as a system that can be characterised and controlled given enough measurement points and data.  However, I doubt that any acts as a linear system, and do it is still extremely difficult to realise as characterised control system.  Some example:

• Most houses leak like sieves, to prevailing wind speed and temperature has a tremendous impact on air-exchange related losses
• Solar gain and direct radiant losses are critically dependent on the area and orientation of fenestration.

We've got a top grade passive-class house and our heating system uses two main inputs: (i) the forecast daily average temperature, and (ii) how much the average house (actually central hall) temperature has deviated from the target set point.  This give us a pretty stable 24×7 house-wide temperature with a ~1 °C ripple (all of our heating is done overnight).  IMO, anything like this would be impossible to achieve with a minimum BReg compliant house.

[ReedRichards]

7 hours ago, TerryE said:

 

...This give us a pretty stable 24×7 house-wide temperature with a ~1 °C ripple (all of our heating is done overnight).  IMO, anything like this would be impossible to achieve with a minimum BReg compliant house.

And I would not want to.  I don't want to live in a house that's the same temperature 24x7 (although admittedly I have never tried it).

 

But I forgot to make the points about the influence of wind and sun in my earlier post.  

[TerryE]

10 hours ago, ReedRichards said:

I don't want to live in a house that's the same temperature 24x7

 

My wife and I are both retired.  We like wandering around in comfort and in bare feet etc.  When we visit our kids' homes, we really notice how cold it can get when you wander from room to room, and the heating is on say during a daytime window, but the temperature drops rapidly when the heating goes off.  Slippers and thick jumpers are pretty much mandatory in the winter.  In our house, the temperature drops by about 1° C per day if the heating is off in the winter.

 

 

[ReedRichards]

Whereas I don't sleep so well if the room is too warm - but when I get out of bed in the morning I like to wander round in a dressing gown and slippers so the house needs to have warmed up a bit by then.

[SteamyTea]

21 hours ago, TerryE said:

However, I doubt that any acts as a linear system, and do it is still extremely difficult to realise as characterised control system

Power delivery will be almost linear, the power needed will not i.e a smokey woodburner and a window open.

But I think if you divide a heating curve by an equation of a strait line, where C is the offset, you get close enough to a straight line. 

Probably more problematic in a small house than a large one.

 

Seems odd for the price of a sensor not to have at least 1 roomstat, then a fully automatic system can easily self learn.