We have a 14kW Mitsubishi Ecodan installed in September last year. The MELCloud system was pretty disappointing and put me off properly looking at system efficiency. For this reason, after the install, we wish we’d considered different vendors including Vaillant. We’ve recently found MELPump third party monitoring and have been looking a bit more closely at the data using the MELPump online charts. I also wish I’d found Open Energy Monitor before planning my install to include a full sensor kit in the spec (but we don’t currently have this on our install).
It’s spring and outdoor temperatures are > 10 degrees most days. Despite this, the hot water CoP figures that MELPump is producing seem to be far lower than I would expect, in the region of 2, some days 1.7, rarely are there days with readings more than 3. Hot water target temperature is 47 deg. C. with the drop set to 7 deg. Looking at the Mitsubishi spec, efficiencies should be higher than this at these temperature settings.
Can MELPump and the underlying Mitsubishi data be trusted?
Are there any sense checks I can do with my smart meter to work out whether these low CoPs are real? I have a 300L tank and can look at specific use over times when the heat pump is working. How much energy should a heat pump use to bring a 300L tank up by 5 degrees C, for example?
I’m familiar with Ecodan, and will impart some advice that I hope will help…
Two important pointers to start with:
Heat pumps are most efficient heating hot water from “cold”, rather than just topping it up.
The temp sensor is below the middle of the tank, so the top usually hotter than reported.
300L is a big tank - what is your typical daily usage? You may find you get better COP by increasing the offset much more, say 20°. Many owners find they can heat up the tank just once a day, ideally at a time that is warmest or cheapest.
Have you set a schedule for “Legionella prevention”? This hurts your COP in two ways, and is unnecessary for most domestic settings [1].
Next, you could try setting “Eco” mode, which takes a little longer to heat the water in the DHW tank but the energy used is reduced.
Meh… no, not so much. It’s fine for catching gross under performance, and for validating improvements, but there are a few factors that can result in errors. How was the heating performance over the winter?
1.75 kWh of heat energy, which would mean about 0.88 kWh of electricity at COP of 2.0, or 0.6 kWh if COP is 3.0. In practice, only part of the tank needs that much heat, so is likely to be less than that.
I will increase the offset, disable legionella cycling, and enable Eco mode and monitor efficiencies.
Typical use is maximum of 2 baths per day but it’s a larger house so the cylinder was specced up to be able to deal with more use than this when guests stay.
I did not have MELPump installed over the winter months but the limited historical data that is available suggests heating CoP figures during the winter months may have been higher. However, I had a few very high electricity bills using up to 90kWh per day which did seem excessive (>£100 per week which is far more than I remember spending on gas). Surely CoPs on hot water during warm weather should be higher/more efficient than CoPs on heating during very cold weather?
What was the estimated heat loss for the property, and designed flow temperature?
Not as much as you might think - the driving factor of performance is the difference between outdoor temperature and the water temperature. DHW heating cycles will run much hotter than the (ideal) low and steady temperature of house heating. A 40° flow temperature in 5° weather will perform better than 60° flow in 15° weather.
Overall COP will also be pushed down in the summer by the standby usage when the system is idle, whereas winter tends to be better as the heating is on for much more of the day.
A non-heating when the house is empty is a minimum of 8.5kWh.
Estimated heat loss 14.47kW.
Do you know anything about the booster heater in the Mitsubishi Ecodan? It seems to come on occassionally and I’m not sure if this is necessary or a misconfiguration
If that heat loss figure is accurate (they are almost always over estimated), then you’d expect to be needing 350 kWh of heat on the coldest day, which is easily over 100 kWh of electricity. I’d guess your actual heat loss is probably closer to 10 kW.
Booster heater should only come on in extreme conditions when the heat pump isn’t able to provide enough heat for the house, and will certainly impact electricity consumption. I wouldn’t expect to see it operating in milder weather.
@jb3tzghz I have the same dhw setup. 300L, dhw setpoint 47c, max drop also 7c in the past. The COP was around 3.5
I changed this to 48c, max drop of 10c. This is due to dhw planning via Odin.
This is the same data as melcloud is using. I also have a certified energy meter, and mostly that one reports 10% less consumption than melcloud. But it’s a good estimate though.
COP of 1.7 is a bit too low in this weather. (It’s even too low for winter btw). Try to plot compressor frequency when running dhw, also show tank temperature.
I think Tim has some spot on points, i don’t trust it fully but if i can make an improvement it thinks from 2 > 3 then there will be real term improvement of a similar proportion i hope!
I also do 1 reheat per day, pretty much full tank up to 47C
My FTC5 eco mode isn’t as good as the FTC6 and newer approach, but certainly worth enabling
I also increased my max run-time to the longest of 120min so in the summer with no heating demand it can go as low-and-slow as it likes
I have a 250Li Mitsi pre-plumbed cylinder (with plate-to-plate heat exchanger for DHW) and a 8.5kW outdoor
I worked a lot on my flow rates during hot water, because of the pre-plumb with two pumps it made quite a difference to balance the speed on both sides of the HEx,
MELPump is reporting a flow rate of 20l/min on my hot water run and i had to try a few consecutive days with different pump speeds on the DHW pump as there is no flow sensor
Generally i get a DHW CoP between 3.5 and 4.7
Here is my MELPump report from yesterday and DHW CoP was reported at 4.5 today, which i do think is probably a bit higher than reality:
HP set in Quietest mode (this gave an instant improvement of between 0.5 and 1 COP. If the tank was very cold a full 3hrs run time might be necessary, but we rarely dip that low.
When the HP is on DHW only (shoulder months / summer) we can see averaged COPs well over 4.
The summer COPs are from last year before going Quietest mode. I reckon we’ll see higher this year.
Caveat is that our heat pump is big enough to provide all we need on Quietest Mode (=somewhat oversized)..
Even last winter on Quietest it didn’t drag the combined COP down much - not like the winter before.
To expand on this further, the most important factor for measuring delivered heat are the temperature sensors, specifically the delta between them. They don’t need to be super accurate, but they do need to match and have good thermal contact with the pipework. I’ve seen several installations where sensors are simply stuffed under the insulation. Copper wire and thermal paste is most ideal; zip ties and aluminium foil at a minimum.
Some things to look for:
flow temperature should be above return temp when heat pump compressor is on.
it’s unusual (but not impossible) for these temps to be inverted when idle.
flow and return temps should (eventually) match when compressor is off and pumps are on,
or when system has been sat idle for a long time - see report Phil shared above.
flow and return should be above tank temperature while heating DHW.
a typical delta for heating would be about 5°, tuneable by adjusting the pump speeds.
The flow sensor that Ecodan comes with is fine, but isn’t calibrated for glycol so will be slightly out for those systems.
Electrical consumption is estimated from compressor frequency and pump status, so COP is likely to be slightly off. Good enough for tracking improvements though.
The problem is solved. As many on this forum helpfully suspected, incorrect installation of stock temperature sensors was the issue:
The flow temperature sensor was in the incorrect position, just after the hot water loop coming off, so it was still getting some of the increased temperature when the hot water tank but only part, hence the low readings. The installer visited and has fixed it. Yesterday CoP hot water reading was 4.21.
Thank you very much to everyone for all the help. I’m very impressed with the community: it reminds me of better internet times
I am going to invest in calibrated monitoring and hopefully submit data to the OEM board.
