Posted by: @jamespetts

↑

I have just had a Vaillant Arotherm+ 7Kw installed in my two bedroom terraced house in London. The installers seem to have been generally quite competent, certainly better than some that I have read about here.

One thing that arose from the heat loss survey is that my radiators were supposedly undersized -=anything from 80% down to 42% of what was supposedly required (at Delta-T 30), with the most used room, the study, being at 69%. These are lovely cast iron radiators that I had had specially installed and matched to my decor when I moved in in 2015, so I did not want to change them. I had a Honeywell Evohome system installed when I moved in, with smart TRVs, which I used with a gas boiler until the end of 2023.

The installers set the heat curve at 1.0 with the maximum flow temperature of 55C and minimum flow temperature of 30C. In the conditions that we have had recently (circa 9-12C), this was more than adequate, and, indeed, the heat pump seemed to be cycling quite frequently.

So, I ran a test - I set the TRV in the study to 30C, and set the desired temperature on the Vaillant controller to 20C with the default 1.0 heat curve and waited to see to what temperature that the study would be heated. It reached 22.5C before the afternoon hot water heating session stopped the increase.

I then found the heat curve tables online, and matched the target flow temperature that the system was giving me with the target flow temperature for the the heat output that I was getting at the same outside temperature, and found that this corresponded to a heat curve of 0.6.

I initially set my system to 0.7, but have since found that even at 0.6, if I set the Vaillant controller to 20C and the Honeywell controller (affecting only the TRVs) to 21C, the study will maintain 21C even with a heat curve of 0.6. Likewise, I have set an overnight temperature on the Vaillant system to 17.5C, and 19C in the bedroom (which is supposed to have only 58% of capacity at Delta-T 30), the bedroom actually maintains 19C all night. I am not using the Vaillant thermostat to adjust temperatures, operating in pure weather compensation mode.

As I understand it, with the correct weather compensation curve, the rooms should maintain the temperature set on the Vaillant controller in any room where the radiator valves are left open and should do so to the same extent no matter the outside temperature (so long as it does not fall below -20, which is extremely unlikely in London). So, if the temperature on the Honeywell controller is set to a temperature in excess of that set on the Vaillant controller in any room, the radiator valve in that room should stay open and the temperature set on the Vaillant controller should be achieved but no more.

Looking at my power usage data from my solar panel app, it appears that the compressor is cycling every 1-2 hours, although the shortest cycles are around 30 minutes. Total space heating power usage for to-day (circa 11C) stands at 3.1kWh with a COP of 3.2 according to the Vaillant app. This includes the practice of slightly over-heating the house between 1400h and 1600h (21.5C) and then letting it fall back (19C) between 1600h and 1900h to try to make sure that the compressor does not run during peak power time (I have the Octopus Agile tariff) unless it is essential, which I anticipate is slightly less efficient than running at the same temperature if the fluctuating price of electricity not be taken into account.

Has anyone else had similar experiences with more heat output than expected? Do people think that it may be sensible to reduce the heat curve even further?

Heat loss surveys are at best a reasonable estimate and at worst total nonsense. Although the science on which they are based is sound, there are too many unknowns about the fabric (in a retrofit) for the answers to be accurate.  Many tend to overestimate.

So if the house is too warm, turn down the heat curve.  It's not at all surprising.

@jamespa Our daughter lived in a flat with tenants below and above; her heating requirements were minimal as the below tenants kept their rooms very warm and the heat percolated upward through the floor.😊 To a certain extent, a mid terrace dwelling is augmented by heat from the party walls. Toodles.

Our 7kW Vaillant is also performing better than expected in that October gave us an average COP of 5.8 and extrapolating from spot calculations would imply a SCOP of 4.5 instead of the 3.9 the installer predicted. We are currently using a coefficient of 0.55 when it was handed over with 0.6. So somewhat above efficiency prediction but not a lot.

On https://community.openenergymonitor.org/t/vaillant-inbuilt-monitoring-vs-mid-meters/24230 the conclusion seemed to be temperature monitoring (from which Vaillant calculate their COP) can have some measurement inaccuracy. So I’m very pleased with the values we see but not trusting them until we have enough data to completely compare with the heat generated (again inaccurate because of assumptions) over a season using gas.

If your style of radiators work then great! We had several of ours changed rounding the sizes up rather than down so we can run cooler, and hopefully more efficiently.

Its very mild so far this year so extrapolating is bound to be inaccurate.

I think that I may have found the lower heat curve limit: I had turned it down to 0.45 yesterday, but I have had to turn it back up to 0.5 again as the study was not coming up to full temperature even some hours after the heating coming on. Interestingly, however, the bedroom was able to maintain temperature during the night in excess of what is set. This suggests that the optimum setting is a heat curve of 0.5, but an even lower night time temperature setting.