Weather Compensation: Experimental Tweaking

Friends, as I mentioned in my previous article, I have no real idea how to actually operate my 5 kW Vaillant Arotherm plus heat pump – or to check how well it is operating. That’s because there is no readable manual for the controller and the App does not do what it says it does.

But since I have an independent monitoring system, I have begun a series of experiments to tweak the heat pump weather compensation settings, and see what happens!

If ‘Reading the Manual is like taking a course in theoretical heat pumps, then this is more like a course in experimental heat pumps.

This is quite a technical article, and it is nearly 1500 words long. So if you are not really interested in heat pump arcana I would recommend giving this one a miss. On the plus side, it does have some nice graphs :-).

Weather Compensation

Weather Compensation is the idea that when the weather is mild, one can heat water in radiators or under-floor heating to a low temperature – perhaps just 25 °C. But when the weather is colder, and the heating demand is greater, one can increase the temperature of the hot water to perhaps 40 °C or 50 °C to meet the heating demand.

Using weather compensation to match the output of a heat pump to the heating demand contrasts with using a thermostat for the same purpose.

Click on image for larger version. The heating supplied to a dwelling can be changed to match demand in two ways. In a traditional thermostat-based system, the radiator flow temperature is fixed and switches on and off to maintain a constant indoor temperature. In contrast, using weather compensation the radiator flow temperature is adjusted.

In a thermostat-based heating system, the flow temperature to which water is heated is pre-set: in boilers it is often as high as 70 °C, and for heat pumps it might be 50 °C. And then to match heating to demand, the thermostat switches the heating source on and off intermittently to maintain the desired temperature.

Weather compensation is particularly valuable when using heat pumps because the coefficient of performance (COP) of the heat pump varies with both flow temperature and environmental temperature. But it can be tricky to adjust the settings for any heat pump, but especially one with no decent manual!

Weather Compensation in action

The graph below shows data taken every two minutes during the week from 00:01 on 11th October 2022.

  • The red curve shows the outside temperature
  • The grey dots show the instantaneous flow temperature.
  • The green curve shows the flow temperature averaged over 1 hour
  • The orange curve shows the internal temperature

Click on image for larger version. Weather compensation in action. When the outside temperature falls, the flow temperature in the radiators increases to maintain the internal temperature.

Notice that when the outside temperature falls, the flow temperature in the radiators increases to maintain the internal temperature.

But on day 4 of the period shown in the graph above, I changed the setting of the Weather Compensation from the curve labelled ‘0.6’ to the curve labelled ‘0.5’ in an attempt to lower internal temperature of the house. I’ll explain more about these labels below.

The graph below shows that average for the 4 days before the change was 21.0 °C and the average for the 3 days after was 20.76 °C: so it does seem to have had a small (0.24 °C) effect, but I will need to continue experiments – see the end of the article for an update.

Click on image for larger version. Graph shows the internal temperature of the house detail averaged over a period of 1 hour. The weather compensation parameter was changed on Day 4 and it does seem to have slightly lowered internal temperature.

It is striking to me how stable the internal temperature is given that – as I understand it – it is based entirely on measuring the temperature OUTSIDE the house – not INSIDE it!

COP

To evaluate the COP, one needs to work out the ratio of the heat delivered to the electrical energy used, over some set time period.

The hourly averaged COP is shown in the graph below. The times when the COP is greater than 4 correspond to times when the difference between the flow temperature and the outside temperature is small, and so not very much heat is being delivered with these high COP values.

Click on image for larger version. Graph shows the hourly averaged COP. Considering only use for DHW the average COP was 3.1 and considering only use for space heating the average COP was 3.9. Overall, considering both DHW and space heating across  the entire period the average COP was 3.7. These averages are shown as dotted lines on the figure.

With a little spreadsheet untangling it is possible to extract the data corresponding to periods when the heat pump is heating DHW and periods when it is heating water for space heating. For DHW the average COP for heating water to 50 °C was 3.1 and for space heating the average COP was 3.9. Overall, considering both DHW and space heating across the entire period, the average COP was 3.7.

Electrical and Thermal Power

Calculation of COP requires evaluation of both electrical power consumed and thermal energy delivered. The graphs below show both these quantities measured every 2 minutes throughout the week or so under consideration.

Click on either image for larger version. Graphs show hourly averages of electrical and thermal power. The DHW cycle runs once a night using cheap rate electricity. The separation of the two uses of the heat pump is not quite perfect: sorry.

Tweaks

So far I have just showed a week or so of data. Now I will explain what I hope to achieve with some ‘tweaks’ First let me explain, about how Vaillant implement Weather Compensation.

Their scheme is illustrated in the figure below. The flow temperature of water in the radiators is set depending the temperature outside. The sensitivity of the weather compensation is set by picking a curve labelled by a number from 0.1 to 4. For example, when the outside temperature is 5 °C,

  • the curve labelled 0.6 would result in a flow temperature of about 34 °C but
  • the curve labelled 0.5 would result in a flow temperature of about 32 °C

Click on image for larger version. The flow temperature of water in the radiators is set depending the temperature outside. The sensitivity of the weather compensation is set by picking a curve labelled by a number from 0.1 to 4. When the outside temperature is 5 °C, the curve labelled 0.6 would result in a flow temperature of about 34 °C but using the curve labelled 0.5 would result in a flow temperature of about 32 °C.

On Day 4 I adjusted the weather compensation from 0.6 to 0.5. To see if this tweak is working we can look at the second figure in this article in this article which I have reproduced below.

Click on image for larger version. On Day 4 the weather compensation setting was changed from 0.6 to 0.5. If we look at cold spells before and after the change it does look as though as the flow temperature is perhaps a degree or two than one might otherwise have expected.

If we look at cold spells before and after the change it does look as though as the flow temperature is perhaps a degree or two cooler than one might otherwise have expected. And since this article has taken a day or two to prepare, I now have a couple more days data on the internal temperature with WC curve 0.5. It does indeed seem to have maintained an internal temperature about 0.23 °C cooler than using WC curve 0.6.

Click on image for larger version. Updated version of the second graph in this article with 3 extra days data. The graph shows the internal temperature of the house in detail averaged over a period of 1 hour. The weather compensation parameter was changed on Day 4 and it does seem to have slightly lowered internal temperature.

Conclusion

My main conclusion is that the weather compensation adjustment does seem to be sort-of working. I will continue experiments and let you know how they go.

My second conclusion, is that observing these effects is really hard and it takes hours of analysis to unearth this kind of insight!

My third conclusion – which you may have already spotted – is that my 5 kW heat pump is just too big. It only needs to output 1,500 W to maintain a temperature of just over 20 °C in my home when the outside temperature is 5 °C i.e. with 15 °C of demand. This seems to indicated that a 3 kW heat pump would have been adequate to heat the home down to (say) – 5°C.

This oversizing is probably responsible (at least in part) for the rapid cycling on and off of the heat pump – exactly what weather compensation was supposed to avoid!

 

Tags: , , ,

14 Responses to “Weather Compensation: Experimental Tweaking”

  1. Simon Duane Says:

    Interesting stuff.
    Can I ask, does your heat pump “modulate” its output?
    I believe this involves the compressor running at a variable speed rather than running always at one speed or not running at all. I did go to the manufacturer’s website, but even after reading the information there I wasn’t sure But guessed that it doesn’t modulate (if only because if it did, I would expect them to boast about it somewhere).
    I only have a gas boiler, not a heat pump, but, in that case, modulation means changing the rate at which gas is burned, and therefore also the rate at which heat is output. The old boiler it replaced didn’t modulate. Both boilers would allow the flow temperature to be set, but only in the sense that the boiler would cut out if the flow temperature would otherwise be higher than the set value. Its flow temperature would look like your grey dots. The modulating boiler is capable of adjusting its heat output until the set value is achieved. Its flow temperature would look more like your green curve (including the exponential decay when the boiler is off altogether for an extended period).
    Perhaps I’ve just worked out the answer to my original question (which would be that that guess is correct).
    Perhaps confusingly, my heating controller completely ignores the boiler’s fancy “modulation” capability, and merely asks the boiler to fire up once every 10 minutes (or not at all), and then shuts it down again, but at precisely the right moment, a minute or more later. It seems crude, but appears capable of remarkably good temperature control.
    I know that one day my still almost new boiler will need to be replaced by a heat pump but, in the meantime, I’m trying to learn what I can from the boiler’s behaviour to answer the question behind your third conclusion – how much heat does my house actually need to remain comfortable?
    Best wishes
    Simon

    • Ioannis Palaiologos Says:

      Hi Michael
      Your videos and articles are a brilliant source of information. Thank you.

      I wonder if the pump would cycle less if you set up the house temperature 1 degree higher and what the cop would be in that case. Perhaps the pump consumes less energy to maintain increased temperature by cycling less.

      Best regards,
      Ioannis

      • Simon Duane Says:

        In the same spirit, but in the interests of preserving comfort, perhaps increase the demand by instead letting in more fresh air – open window(s), etc?
        Though the experimentalist in me sees the drawbacks of complicating the interpretation of results by deliberately enhancing the effect of confounding variables. I would probably test my patience by waiting for colder weather 😉
        BW
        Simon

      • protonsforbreakfast Says:

        Ioannis, Thank you for your kind words.

        Yes, I think it would cycle less if I increased teh house temperature, but it’s already warmer than it was! For most of 2019-2021 we lived with the thermostat at 19 °C!

        Anyway: best wishes: Michael

  2. protonsforbreakfast Says:

    Simon,

    Good Evening. I trust you are well. My advice is to follow https://www.urbanplumbers.co.uk aka Szymon Czaban on Twitter and especially YouTube where he has great videos about both heat pumps and boilers. The ability to modulate is key to healthy operation.

    As I understand, Heat Pumps can modulate from 100% to typically 40% of their rated output. For the 0% to 30% of output they need to switch on and off. My heat pump will probably begin operating continuously when the temperature gets closer to 0 °C.

    Similarly, many boilers can modulate but typically have a minimum output of around 6 kW. If the average heat required is below that, they switch on and off.

    Using a low flow-temperature and modulation will prolong the life of the boiler(corrosion is related to temperature) and increase efficiency into the 90%+ range.

    Best wishes

    Michael

  3. Simon Duane Says:

    Hello Michael (and I’m sorry for omitting any greeting from my first response!) and yes I’m well, thanks. You too, I infer (from all your activities…)

    But, from what I’ve seen of manufacturers’ literature, although heat pumps _can_ modulate, far from all heat pumps actually _do_ modulate. And I was genuinely unable to find any clear statement in Vaillant’s various pdfs. Hence my question whether you know any more about yours…

    As you say, in this mild weather, it will be running at whatever minimum output it can achieve and will go below that only by cycling. For what it’s worth, my boiler will modulate down to 10% of rated output (i.e. 3.2 kW instead of 32 kW) but it takes tens of seconds for the burning to stabilise enough for it to get that low, and so short burns of a minute (in practice, a burn can be shorter than that) tend be dominated by the initial output which is between 10 kW and 15 kW. I too am looking forward to colder weather, when I’m hoping to see more sustained burning at a lower (intra-burn) average rate.

    On another tack, and re-reading your post, I wonder if it would help to distinguish between the temperature of water in the radiators and the temperature of water in the heating circuit. I assume that your system still has radiator valves which (partially) open according to the need for heat in the room? Adjusting the (flow) temperature of the circulating water by weather compensation should mean that those valves open more, and more often. Which is all good for delivering the heat to where it’s needed. A related question would be, does the pump circulate the water in the heating circuit all the time, or only sometimes (related to when the heat pump is heating it)?

    BW

  4. protonsforbreakfast Says:

    Simon: I have replied by direct e-mail because I can’t post pictures in the comments. But, yes, the heat pump does modulate when the demand increases.

    M

    • Simon Duane Says:

      Thanks very much for the extra graph, Michael – it is entirely convincing. In words, for those reading this without seeing the graph, my reading is that when demand is below the range covered by modulation, the heat pump only runs at minimum output, and cycles in order that the demand is just met (on average).

  5. Frank Says:

    A large thermal store can smooth out cycling. You’d really rather have the heat pump operate over 1hr.

    I also like the idea of having constant low flow temps for central heating, with dhw priority low temp with an immersion sterilisation cycle every few weeks (copper tanks reduce frequency, and off peak rates could be used to reduce cost of heat/kwh).

    The way you get peak thermal loads would be to increase air flow over a convector radiator. This could modulate up to 50% output Vs natural convective heat transfer (this figure is a guess but you get the point)

    • protonsforbreakfast Says:

      Frank – Thank you: some nice ideas there that I will file away. I have seen sets of fans that clip on to radiators to improve heat transfer but they look a little ugly in my opinion.

      Anyway

      All the best Michael

  6. Andreas tober Says:

    Hello Michael, I have a question which has little to do with the main points of your post, so feel free to ignore it but as a physicist to may understand my idle curiosity. In your first graph, the blue curve shows some significant variation in slope on the upward sections, i.e. when the circulating water is heated. It strikes me as odd that this should take so much longer to reach its maximum when the outside temperatures are higher. Is this curve an actually measured curve or one generated in some other way? Or maybe sourced from elsewhere?
    Best wishes
    Andreas

    • protonsforbreakfast Says:

      Andreas, Good Evening: please forgive my delay in replying.

      I’m sorry. If you are talking about this diagram than I just drew teh blue line in PowerPoint to illustrate the qualitative difference in behaviour. Sorry.

      Best wishes

      Michael

  7. Dan Grey Says:

    Still think it’s bonkers that the user can’t just set a desired internal temperature, and the controller works out which WC curve is appropriate to achieve that, given the house’s thermal performance. The only manufacturer I know that has that feature is Mitsubishi in their Ecodan models (“auto-adapt” setting).

    Possibly Vaillant’s “influence” setting does something similar?

    Ultimately the controller logic has three inputs (internal and external temperature, target temperature) and two outputs (compressor motor speed and flow pump motor speed). It wouldn’t be hard to code a neural net to solve that, targeting whatever compressor and pump speeds are optimal for CoP.

    Even just a picking pre-calculated WC curves would be a big improvement, rather than (presumably) depending on the installer making multiple visits to select the right curve for the customer’s choice of internal temperature (and what if they change their mind?).

  8. protonsforbreakfast Says:

    Dan: I agree with everything you have said.

    It is also bizarre that there is not a single piece of documentation that explains how the settings work!

    And how can a heat pump not automatically display its COP?

    Best wishes

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 )

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: