Heat Pump: First Space-Heating Results

Friends, the unseasonably warm autumn has meant that I have had to wait until 23rd October for our 5 kW Vaillant Arotherm plus system to switch itself on.

I only have a couple of day’s data, but the results are interesting and promising. First I will show the data for each day and then discuss what it means at the end.

23rd October 2021

Click image for a larger version. The graph shows electrical power consumed by the heat pump and the thermal power delivered by the pump during the 24 hours of 23rd October 2021.

From midnight until 6 a.m., there is no space heating, but the heat pump operates for 1 hour to heat the domestic hot water (DHW) tank. Performance for hot water heating is described in these articles (1, 2)

From 6 a.m. until midnight the system is responding to the thermostat set at 19.5 °C and heated water is circulated around the radiators.

The heat pump operates 12 times up until 6 p.m. after which no additional heating was required.

As shown on the graph, 12.1 kWh of heat was delivered using only 2.9 kWh of electrical power. which over 12 hours amounts to an average heating power of around 1 kW using only 250 W of electrical power.

The ratio of thermal to electrical power is known as the coefficient of performance (COP) and this is summarised in the graph below

Click image for a larger version. The graph shows Coefficient of Performance (COP) of the heat pump during the 24 hours of 23rd October 2021.

The COP seems to increase slowly through the day peaking for a few minutes in each cycle at values as high as 6, but the space heating average is 4.2.

Click image for a larger version. The graph shows various temperatures relevant to heat pump operation during the 24 hours of 23rd October 2021. The external temperature; the internal temperature; the flow temperature in the radiators; the temperature of water in the DHW cylinder.

The initial flow temperature is ~34 °C which falls through the day to ~28 °C. This reduction in flow temperature is in response to the increase in external temperature from ~12 °C to ~15 °C.

This so-called ‘weather compensation’ allows the use of lower flow temperatures, which enables the system  to operate with the highest possible COP.

I was surprised that even with my unaltered radiators, flow temperatures of 35 °C were sufficient to warm the house.

24th October 2021

Here is the equivalent data to that for the 23rd – the graphs are similar and I show them only to show that the system seems to be behaving reproducibly.

Click image for a larger version. The graph shows electrical power consumed by the heat pump and the thermal power delivered by the pump during the 24 hours of 24th October 2021.

Click image for a larger version. The graph shows Coefficient of Performance (COP) of the heat pump during the 24 hours of 24th October 2021.

Click image for a larger version. The graph shows various temperatures relevant to heat pump operation during the 24 hours of 24th October 2021. The external temperature; the internal temperature; the flow temperature in the radiators; the temperature of water in the DHW cylinder.


My first conclusion is that the system works. This is quite a relief!

My second conclusion is that the system works slightly better than I had been hoping for.

Being able to heat the house with these low radiator temperatures means that over winter the average COP could be higher than my prior estimate of 3. This means I will use less energy, emit less CO2, and spend less than I had estimated.

My third conclusion is that average heating power was about 1 kW during both days. But the system has plenty of margin to deliver more power either by changing the duty cycle – the pump could stay on for longer – or by increasing the flow temperature.

As soon as the weather settles down to being reliably cold I will begin to carry out experiments to optimise the weather compensation.

My fourth conclusion will be the basis of another article – but I can see already that as far as heating is concerned – I am going to have a very cheap winter.

Keep warm.

Tags: ,

11 Responses to “Heat Pump: First Space-Heating Results”

  1. 171indianroad Says:

    Glad you are happy.

    The COP numbers are consistent with what we get here in Canada down to around -10C.

  2. protonsforbreakfast Says:

    Wow. I would really like to have a few days at -10 °C just to see how the system worked at those temperatures.

    But with this bloody global warming I will be lucky if we even get a week below 0 °C!

  3. Peter Grebenik Says:

    I have a question. How noisy is the heat pump?

    • protonsforbreakfast Says:


      Not noisy at all. I have now put it on my list to make some measurements, but here is an anecdote and an observation.

      The installers asked if a couple could visit our installation to listen to the noise. So a very nice couple came around and listened. However there was no heating demand and I had trouble getting it to turn on – it’s mostly still and completely silent in summer. Eventually I ran off hot water and asked for a hot water ‘boost’.

      We all stood around the machine waiting for something to happen. Eventually one of the visitors said his legs were getting really cold! The machine had started and none of us had heard it! Eventually it began to speed up and we could hear it. But in the garden a few meters away it was lost amongst the suburban noises of birdsong and swaying trees. The couple were very happy.

      As we stood around, I notice that the noises from our neighbours gas boiler were louder.

      In the summer when you are likely to be sharing garden time with your heat pump – the heat pump is generally not working. Ours comes on once each night to heat hot water using cheap rate electricity.
      In the winter when the heat pump is operating you are unlikely to spending much time in the garden.

      Best wishes


  4. David Dye Says:

    I’m confused. Question: in your Temp. vs. time graphs for the two days above, after 1800 you are maintaining a rock-solid 19.5 internal temperature with no heating input and a varying external temperature. But, between 0600 and 1800 you have to run your CH system once an hour to achieve that, even at higher external temperatures, e.g. at 1300. Why?

    I can think of several possibilities;
    (1) The internal temperature monitoring sensor is somewhere in the middle of the building and when the heating system is running there are some colder rooms somewhere getting heated that aren’t then heated after 1800.
    (2) when your CH system is running, it will have some heat output even if every rad’s TRV was closed – eg heat loss from the buffer (4kWh/d if continuous?? said one report I read). But probably this is too small to give you 10 and 12kWh heat supplied.

    Is that right? In any case, how much of the heat supplied is needed to keep that internal monitoring sensor at 19.5C? And where is the rest going?


    • protonsforbreakfast Says:

      David ,

      Well spotted. The behaviour arises (I think) for the same reason that the heat hasn’t come on until late October even though the external temperature has been lower than the set temperature for a month or more.

      The reason is that there are other heat sources in the house. Specifically:

      1. Cooking with gas – typically 1.5 kWh/day
      2. Electrical Power – all electrical items (dishwashers, computers etc.) are fundamentally heaters – typically 10 kWh/day
      3. Human – the house typically has either one or two humans – each one outputs ~ 2.4 kWh/day

      So there is typically 1.5 + 10 + 3.6 =~ 15 kWh/day of heating all the time. This explains why a graph of [heating output] versus [heating demand] does not go through the origin.

      I looked at this issue thoroughly (in my opinion!) in this article


      I still had a gas boiler then but the issue was the same.

      Does that make sense?

      Best wishes


      • David dye Says:

        Yes but no. I similarly have a flat that has neighbours below and on either side, lots of wall insulation and so my heating only comes on in late Oct – but I keep my room stat on 20C 24/7/365. So sure, with limited heat loss then when I run a regression on my heating energy consumption, my intercept is at an average daily temp of about 14C (external temperature data from degree-days.net). And similarly, my ‘hotel load’ is about 200W or 5kWh/d, and the humans are worth another 5ish kWh/d.

        But that still doesn’t explain (to me) why you need to add heat to maintain temperature from 0600-1800, but don’t need to do so from 1800-2400 on the two days you plot above?

  5. David Dye Says:

    On another note, the Vaillant Arothem plus 5kW does both modulate and weather compensate, right? So that is, your ontime per hr looks like 10-20min? And your DT on flow temp between cycles is 4C? Do you know the fill volume in your system including buffer? e.g. if it were 100L then 4200 * 100 * 4 / 3.6e6 / 5 * 60 = 5.6 mins at full throttle to reheat your fill volume? But at the moment you look like you’re running at about 60%, so its about 10 mins per hour duty cycle? And you need that to be ‘a reasonable number’ (5 mins) because the COP takes a few minutes to come up on start-up? Plus, avoiding cycling / wear?

    • protonsforbreakfast Says:

      David – good questions all.

      Regarding your previous e-mail, I genuinely don’t know why the ASHP didn’t come on in the evening. My guess was that my wife and I must have been doing something extra in the evening cooking or watching TV.

      Modulate and Compensate. Mmmm. The Arotherm has weather compensation – there is a graph showing a number of curves that it follows. Each ‘curve’ has a different profile of flow temperature versus external temperature (i.e. demand). One picks the curve which best matches the required flow temperature at maximum demand. I have wondered whether I am on the wrong curve and so am experimenting with this setting for a few days in the home of reducing the on/off cycling.

      I am not sure what you mean by modulation. Is this the ability to change the flow rate as well s the flow temperature? I am embarrassed to say that I don’t know if it has this feature.

      Anyway – I just need a few days of definitively cold weather to experiment!

      Best wishes: Michael

      • David Dye Says:

        Modulation; I mean operate at less than 100% power. i..e an inverter driven pump can operate at fractional power. So e.g. at 60% its quite a bit quieter; I think the manual implies that it can? So a modern 26kW gas combi boiler might run at 3-5kW for heating and 26kW for the shower, which is obviously just kinder on the system from a cycling perspective. Example https://www.heatgeek.com/viessmann-vitodens-200-boiler-review/. My question here essentially is how much system volume do I really need; the age-old question of whether buffers are really required, especially if you have upgraded to giant rads.

      • protonsforbreakfast Says:

        David – I am pretty sure the Arotherm plus can modulate its power – I had imagined that this was how it achieved the lower flow temperatures.

        Regarding radiators, I have kept my old radiators, but a year before installing the heat pump I added triple-glazing and external wall insulation which reduced heating demand by about 50%. My plan was to see how the system performed and maybe upgrade some of the oldest radiators if those rooms were too cold.

        I really can’t comment on the system volume. Is your thought that if the system volume is ‘too low’ that this might be causing the cycling behaviour?

        BTW: You seem very knowledgeable – have you written down your insights anywhere?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: