@glpinxit - these are common problems. Those of us with Midea units, for example, have almost identical built in uncertainties, including the crucial one of what goes on inside the black or rather white box? Then there is the concern about what is known in this parish as compressorgate, the heat pump manufacturers are in effect marking their own home work. Whether compressorgate will become as big as dieselgate is yet to be seen, but it wouldn't surprise me if something like it does come along.

It is possible to monitor a heat pump using third party kit but it involves £££ and quite a lot of setting up. Rather surprisingly, there is a lack of data comparing results from third party monitoring with the heat pump's own efforts.

Sooner or later validity and reliability will crop up. If you are only looking changes in efficiency in a single system using the same monitoring over time, then as long as the measurements are reliable (when true value X is 2, measurement always comes back as 2.2 etc) then, in an ordinal (putting things in order) way, you can compare two readings, but even then you hit the problem of changing environmental factors eg my COP is better today, but is that (a) because it is milder today than yesterday or (b) because of the system changes I made first things this morning (and of course (c) neither of these and (d) both of these). Getting valid data (getting 2.000 when tho true value is 2.000) is much much harder, possibly even beyond realistic in a domestic setting without spending a lot of money.       

Posted by: @glpinxit

↑

I am concerned by how the energy consumed and produced figures displayed by the Ecodan control panel are derived. I don't know if this is just a Mitsubishi thing or also a feature of other ASHPs.

Our ASHP was installed last summer (July '22). Since then I've been learning how to get the most out of it. But I've now realised that isn't as simple as I first thought.

Based on my MCS documentation I'd been expecting a CoP around 3.6 so I was very happy with data for December that gave the CoP as 3.75. However, the answer for January was 2.6 and since then I've been recording data daily. This throws up my problem.

I didn't look at the aggregate results for January until the third week of February. Since then I've recorded the Ecodan's daily readout of energy consumed and energy produced. I was already recording the energy consumed based on a supplementary electricity meter that was installed at the same time as the heat pump.

I quickly noticed that the metered figure for energy consumed was not consistent with the Ecodan's figure. This inconsistency is itself not consistent- so far it has varied between - 51.3% and +21.5% . A positive (+) difference means the heat pump has recorded less energy consumed than the meter recorded.

Overall, the right answer doesn't come out in the wash when the daily figures are aggregated. The Ecodan generally suggests higher consumption (5.1%, 5.9% and 7.6% for December, January and February). My inclination is to trust my freestanding consumption meter, although I can't compare it with the meter that my bills are based on.

This is bonkers! My assumption is that the heat pump doesn't measure energy consumption but infers it from other parameters- does anyone know where it gets its numbers from? Equally, how does the heat pump derive its 'energy produced' figure? If it is as wobbly as its consumed figure then I lack confidence in it but might it use a methodology that means the resulting CoP is more reliable?

My wider concern is that I'd like to be able to track whether changes I make improve or worsen the efficiency of the pump. But I currently feel that I can't do that with any confidence (including whether I should be concerned and try to improve performance or be happy that it is good enough). Certainly it seems clear that any changes need to be monitored for at least a week before drawing conclusions about their impact.

I should add a short note about data. The meter data is to two decimal places while Ecodan only gives whole numbers- I assume that it rounds conventionally (ie >0.5 is rounded up otherwise round down). This is a potential source of inconsistency and will compound the overall errors. Any difference is more significant when consumption is low and this will be exacerbated by rounding. But the observed inconsistency can't just be explained by rounding.

Thanks for reading this far! Guy.

 

The external electricity meter will no doubt be more accurate than the Ecodan produced data. To measure the true power level (kW) requires measuring the voltage, current and phase angle of the electricity supplied. To quantify the electrical energy (kWh) supplied, requires integrating the instantaneous power level readings over a period of time. The accuracy of the data produced will be dependent upon the accuracy of the measurements taken, and the sampling frequency, since digital system take instantaneous measurements which may not reflect the average reading.

Measuring heat energy produced is even more complex, since it involves accurate measurement of LWT and RWT, to calculate the DeltaT, and also an accurate measurement of the water flow rate, with account being made for any antifreeze that may have been added to the water.

Even if an external flow computer is added to measure the heat energy output, the percentage of antifreeze would need to be known.

You could try e-mailing Mitsubishi Technical Help and ask them to specify how accurate the readings should be.

@derek-m another query I have in the back of my mind is how the energy consumed by the immersion heater is being calculated.

in certain circumstances the DHW cycle will activate the immersion heater at the end of the heating cycle to attain the higher temperature. Yet the heater is supplied via a separate fused supply. And in some cases can be physically switched off via a local isolation switch. 
my doubts are: how does it read the energy being used by the immersion heater to top up the DHW and if it’s time driven how does it know if the immersion heater is not operating eg switched off. 
this would potentially have a massive impact on COP accuracy. 

Posted by: @sunandair

↑

@derek-m another query I have in the back of my mind is how the energy consumed by the immersion heater is being calculated.

in certain circumstances the DHW cycle will activate the immersion heater at the end of the heating cycle to attain the higher temperature. Yet the heater is supplied via a separate fused supply. And in some cases can be physically switched off via a local isolation switch. 
my doubts are: how does it read the energy being used by the immersion heater to top up the DHW and if it’s time driven how does it know if the immersion heater is not operating eg switched off. 
this would potentially have a massive impact on COP accuracy. 

I think I have read in some manuals, that the installer (or you) need to confirm that an immersion heater is installed along with its rating. The controller could then infer that if it is switched on for 30 minutes and is a 3kW unit, that it will have used approximately 1.5kWh of electrical energy.

The only way that a controller could realise that the immersion heater is switched off, would be by sensing that the hot water cyclinder temperature did not increase as expected. I don't know if controllers are programmed with this capability.

I've emailed Mitsubishi to how the Ecodan arrives at its figures. (Also, I'm pretty sure that in our system the immersion heater power doesn't go through our heat pump meter. I might check this, however.)

Posted by: @glpinxit

↑

I've emailed Mitsubishi to how the Ecodan arrives at its figures. (Also, I'm pretty sure that in our system the immersion heater power doesn't go through our heat pump meter. I might check this, however.)

Even if the electrical supply to the immersion heater does not go through the heat pump electricity meter, the power used by the immersion heater can be inferred, if the heat pump controller is switching the immersion heater on and off. Of course inferring use, does not take account of periods when the controller has switched on the electrical supply to the immersion heater, but any inbuilt thermostat on the immersion heater is up to temperature, and therefore not drawing power from the supply.

The only true way of measuring actual usage is by using a suitable meter.

I had a reply from Mitsubishi to my 'how does it measure things' question a few days ago. It was quite short and rather disappointing and it said "Our estimated readings are based on a CT coil in the indoor unit for Power and the flow rates obtained by the flow sensor for delivered heat". Since then I've been quietly fuming- it sounds as though I should assume that  the heat delivered numbers are as unreliable as the energy consumed numbers (or 'estimates' as I'm going to have to get used to calling them).

Even recognising the numbers as 'estimates' overstates their implied value because the metered energy consumption is always unpredictably different in either direction. This leads me to infer that the estimates are a guide rather than a measurement. So, while the estimates may provide a general indication of overall efficiency, their inconsistency means they are no use as a reasonably quick way to measure something like the impact of tweaks to my weather compensation curve. 

If my car fuel gauge was made by Mitsubishi then I think I'd always fill up at the half-empty point. And if they made the speedometer then I think I'd have been banned by now!

Posted by: @glpinxit

↑

I had a reply from Mitsubishi to my 'how does it measure things' question a few days ago. It was quite short and rather disappointing and it said "Our estimated readings are based on a CT coil in the indoor unit for Power and the flow rates obtained by the flow sensor for delivered heat". Since then I've been quietly fuming- it sounds as though I should assume that  the heat delivered numbers are as unreliable as the energy consumed numbers (or 'estimates' as I'm going to have to get used to calling them).

Even recognising the numbers as 'estimates' overstates their implied value because the metered energy consumption is always unpredictably different in either direction. This leads me to infer that the estimates are a guide rather than a measurement. So, while the estimates may provide a general indication of overall efficiency, their inconsistency means they are no use as a reasonably quick way to measure something like the impact of tweaks to my weather compensation curve. 

If my car fuel gauge was made by Mitsubishi then I think I'd always fill up at the half-empty point. And if they made the speedometer then I think I'd have been banned by now!

In fairness to Mitsubishi, I suspect that they are doing no less than most other heat pump manufacturers.

As stated, the 'energy in' measurement is derived from a CT (Current Transformer) located within the indoor unit, but it can only measure the energy flowing through it. In your system the immersion heater is powered from a separate power supply, which would appear does not go through the CT. To obtain reasonably accurate 'energy in' measurements would require the installation of an additional power meter, which would need to measure all the power supplied to your heating system. Fitting such a device would probably cost at least a couple of hundred pounds.

The flow sensor, along with LWT and RWT measurements, can be used to calculate the 'energy out' value. The accuracy of the calculated value will be dependent upon the accuracy of the three measurements involved. I don't know of any heat pump manufacturer's who state the specified accuracy of displayed measurements, which would probably mean that the accuracy is not tested and verified. A more accurate 'heat energy produced' value can be obtained by correctly installing a heat meter of specified accuracy. To do so would probably cost several hundred pounds.

At the end of the day it is all down to money. Before the large increases in the cost of energy, I doubt that most people monitored their energy usage, provided that their bills seemed reasonable. So why would a heat pump manufacturer increase the cost of their system by £500 or more, to install a more accurate measuring system. This would not make good business sense, since it would increase costs, possibly reduce sales, and could affect the large bonuses paid to the directors.

Whether this changes in the future only time will tell.

I sent a query to Vaillant tech support asking pretty much the same question.  What is the accuracy of your performance monitoring system.  Funnily enough, I am still an answer.   If I do get one I will post the reply for sure.  

Posted by: @derek-m

↑

I don't know of any heat pump manufacturer's who state the specified accuracy of displayed measurements, which would probably mean that the accuracy is not tested and verified.

What I find perplexing, and this is obviously something that varies from manufacturer to manufacturer, is that most/all collected data is hidden. Our heat pump collects vasts amounts of data, but it’s hidden behind ‘secret’ password protected admin areas. I find that kind of practice a bit dodgy and it’s like they’re trying to hide something, and they don’t want complaints from homeowners that their systems aren’t delivering what’s stamped onto the side of the heat pump units.

Posted by: @glpinxit

↑

I had a reply from Mitsubishi to my 'how does it measure things' question a few days ago. It was quite short and rather disappointing and it said "Our estimated readings are based on a CT coil in the indoor unit for Power and the flow rates obtained by the flow sensor for delivered heat". Since then I've been quietly fuming- it sounds as though I should assume that  the heat delivered numbers are as unreliable as the energy consumed numbers (or 'estimates' as I'm going to have to get used to calling them).

Even recognising the numbers as 'estimates' overstates their implied value because the metered energy consumption is always unpredictably different in either direction. This leads me to infer that the estimates are a guide rather than a measurement. So, while the estimates may provide a general indication of overall efficiency, their inconsistency means they are no use as a reasonably quick way to measure something like the impact of tweaks to my weather compensation curve. 

If my car fuel gauge was made by Mitsubishi then I think I'd always fill up at the half-empty point. And if they made the speedometer then I think I'd have been banned by now!

Although I now have independent electricity and power meters on my Ecodan, I didn't always. I haven't done any direct, controlled comparisons but as far as I remember, the measured COP is reasonably close between the internal and external methods. Also, when I read of COPs from others with similar systems to mine who don't have external meters, they are similar to mine.

Where the Mitsubishi measurements do seem to fall down a bit is the consumed energy at low outputs; people who have compared using external meters report that the Ecodan overestimates in these conditions.