@cathoderay it's this month, those are just the first few days of November. Below is this week. To be honest the only numbers you would care about are the heating and domestic hot water (DWH) as those are measured.

Below is the snap shot from yesterday. As you can see you have the demand, power draw and heat delivered. COP is then a simple calculation. 62/19 = 3.2 on heating 

@batalto - I think the problem may be the charting. The numbers for 2022-11-18 ie yesterday/Friday (those in second screen grab) don't appear too unreasonable, and the calculated COPs seem in the right ball park, though I still think the proportion used for DHW seems high, but the yellow/orange chart bars ie heating and DHW combined, dont really make sense unless there is a valid reason for your usage to vary greatly day by day. Assuming the displayed ambient is correct, and it isn't inconsistent with what we have had here in the south, and we have had a slow but steady decline in ambient over the week, then heating energy consumption should, presumably go up, slowly but steadily. What the chart shows is it dropped on Monday (despite ambient being down a fraction), then rocketed up to well over double Monday's use by Thursday (despite only a small drop in ambient) before falling back again yesterday (despite yesterday's ambient continuing the slow decline). 

If the app can output a csv or whatever of the raw daily data then we can plot is ourselves and see whether it makes sense. It is one inviolate law of data interpretation for me that if the workings aren't available, so that a competent person can verify they make sense, then the interpretation is always suspect. If we can download the data, and plot it ourselves in a transparent way and get plots that make sense, then I am usually inclined to lean towards believing what I can see, at least some of the time!

Does the app allow data download? if it does, that might be enough to bury my security concerns, at least for a short while. It is such a pity you can't just connect a USB lead to the controller, and get the data that way.

@cathoderay we use a lot of hot water due to bathing two children every day. On the heating I'd say it's accurate. Our heating is actually driven by the sun, not ambient. It's just how our house is. Solar gain is a massive driver for us. Warmer, rainy days will use far more heating than a cooler sunny day.

@batalto - ah, that makes sense, I'd forgotten you have a considerable solar input. Downloading the raw data from the app - do you know if that is possible? If it isn't would it be practical to manually copy sample periods, ie does the daily data go back a long way in time, or does it only have say the last 3 months or whatever. I also rather assume the data is stored 'in the cloud' (or more likely on Midea's servers?) rather than in the controller itself, meaning if I do set up the app, I will only have data from when I set it up, rather than historical data. 

@cathoderay

Based upon the information you supplied some time ago about your system, a predicted COP of 3 could be expected with ambient temperatures in the 3C to 4C range. With a 'perfect' system at an ambient temperature of 10C, a COP of over 6 may be possible. Obviously in the real World these figures may be difficult to achieve.

Measuring COP accurately is not very easy, and would involve the correct installation of accurately calibrated equipment, to record the electrical power input, the LWT, the RWT and the water flow rate. I suspect that the standard equipment fitted to an ASHP, will not be that accurate, and may not even be measuring the true values, since the sensors may not be located in the most ideal locations.

I fear the best you can expect from the installed equipment would be to monitor variations in performance over time. Taking readings over a 10 or  15 minute period are unlikely to produce any meaningful results.

Even the History Data, kindly supplied by Batalto, cannot be deemed 100% accurate, because it is showing rounded integers, so the Electric Consumption could be anywhere between 18.51kWh and 19.5kWh, and the Cooling/Heating could be 61.51kWh to 62.5kWh, and that is assuming that the readings are actually accurate.

To monitor your system I would suggest that you record all available parameters in a spreadsheet for a period of at least 1 week, after which you can post the results and we can all have a go at interpreting the data.

As I expect you all have recognised, I tend to be sceptical of data that I can't verify at every stage, especially when the data is in effect the manufacturer marking their own homework. Even reputable manufacturers (if that is not an oxymoron) can get up to very dirty tricks. One only has to think of VAG and Dieselgate.

I wonder if there might be a way of estimating COP using a different approach. If we assume heat loss calcs are reasonably accurate (a very big if, but it may be that errors are random and so cancel out), and if we assume that the heat loss/ambient temp is linear between say -2 and 18 degrees ambient, then in my case the heat loss at 9 degrees ambient will be about 6.2kW (as suggested above). If the house stays at design temps, then that suggests that, on average, the heating system is delivering 6.2kW, because the system is in equilibrium over time, even if there are small ups and down over short time frames. We can say the heat pump is delivering an average of 6.2kW power, and it it does that for 10 hours, then it has delivered a heating energy of 62kWh over that 10 hour period.

If I then see how many kWh the heat pump has drawn via the inline heat pump kWh meter (the external one, next to my main meter) over that 10 hour period, I can then estimate the COP, heat energy in kWh delivered to the house divided by electric energy consumed by the heat pump in kWh.

Such an approach has it seems to me a number of advantages: it is real world based, based as it is on what actually matters (is my house at the temp it should be, and if it is, how much electricity have I used to get there) and it uses simple and generally reliable and likely to be valid external measuring devices, a room thermometer and an independent kWh meter. No reading at all is needed from the heat pump itself. It's major disadvantage is it assumes heat loss calcs are valid and reliable, a big assumption. Or is it? If an old leaky 3 bedroom house typically has a heat loss of around 12-13kW at -2 ambient (surely by now we most know the likely values), and mine comes out at 12.4, then perhaps it isn't that far out. 

I would dearly love to produce the mother of all spreadsheets but I fear I will fail in the attempt. For starters, I have no plans to set an alarm clock to take readings every hour on the hour during the night, so about a third of a 24 hour day's data will always be missing. Secondly, I know there will be times when readings get missed, or are too early or too late, during the day. Even hourly readings may not be enough, the whole system is usually somewhat volatile. For example, the LWT might happen to be around 30 degrees at two readings an hour apart, and there would be no record of the fact that not just once but twice the LWT peaked at 40 degrees during the hour. This is at least less of a problem with cumulative readings (eh kWh used, two readings an hour apart will tell us what average use was over the hour.

The problem is we are stuffed as long as we don't know how Midea determine the figures displayed on their controller. I think I am right in saying the Comp. Current is always an integer as well, suggesting rounding, which given the low absolute values, say typically around 5, could mean builtin 10% or so errors, unless they are totally random (and so cancel out). We still have no way of knowing how it actually determines 'Heat Pump Capacity' ie claimed moment by moment output and so presumably the interval (daily etc) figures that appear in the app. The fact it says it has used over 5600kWh over a period when the external meter says it drew only just over 4000kWh is enough to make me suspicious of what other figures are also out. 

Hence my suggestion that it might be best to ignore the Midea data, and approach the problem looking as it were from the outside in, as in what went in and what came out of the black box, rather than looking out from inside of the black box to the outside, most likely through a myriad of prisms of Midea's own making.                 

@cathoderay

If you wish to calculate the COP with more accuracy, you could consider having the MMSP equipment installed, which like your power meter is external to your heat pump system, and should provide greater accuracy.

Some time ago I sent you a spreadsheet based upon your system criteria, column D of which gives predictions of the likely COP, LWT and Input Power at various ambient air temperatures.

Obviously the spreadsheet is based upon manufacturers data under specified operating conditions. The fact that you have a PHE and other possible problem areas will no doubt reduce the overall performance of your system.

Posted by: @derek-m

↑

If you wish to calculate the COP with more accuracy, you could consider having the MMSP equipment installed, which like your power meter is external to your heat pump system, and should provide greater accuracy.

I already have an external power meter, the heat pump supply has its own meter. I know exactly what the heat pump input is, recorded on a weekly basis over the heating season. The problem is how to determine what the heat pump output is. Having thought about my steady state assumptions, I am now not so sure the calculations I have used make any sense. If, as an example, I assume the average ambient temp is 9 degrees, and the house is in a steady state with room temps where they should be, then I estimate my heat loss to be 50% of that at -2 degrees ambient, the design temp, in my case 6.2kW (half the heat loss at -2 balanced by the same amount of heat being supplied by the heat pump, thus keeping my house in steady state conditions.

But this gives an absurd whole heating season heat loss: 6 months x 30 days per month x 24 hours per day x 6.2kW = 26,784kWh! Which is way too much! When I used oil for heating, I used around 11,000 to 12,000kWh of oil per heating season, and more recently, using only mains electric heating, with only room in use heating, it was around 9,000kWh per heating season. But at the same time, the 6.2kW load/use doesn't seem that excessive, after all it is the same, on average, as six 1kw convector heaters running all the time, say three upstairs and three downstairs, to keep the house in steady state, which is the sort of ball park usage I might expect.

But the total usage over the heating season doesn't make sense, it is over twice what I actually need to use to heat the house. Something is very wrong with my assumptions and/or calculations, but I can't spot what it is...

@cathoderay I found the design calculations from all heat pump installers were huge for us. When checked against our gas usage for the previous year and using degree days to estimate the -2 dree heat load my estimates where half those of the lowers heatpump quote calcs.. I think they are super cautious.. we have a basement so some walls never experience a real outside -2 more like +10 all year so that might have been  part of it.

Protons for breakfast on you tube has some nice cheats for estimates of peak load based on you previous energy requirements. Worth a look. 

@cathoderay

Measuring the electrical power input is fairly simple, and there are plenty of power meters that can do this with reasonable accuracy. Calculating the heat output energy is much more difficult, in that it involves measuring the LWT, RWT and flow rate with reasonable accuracy. Any inaccuracy in any of these measurements will lead to calculation errors, that are compounded over time. An accurate industrial grade system can cost the same as your heat pump.

As BristoJoe has pointed out, who is to say that the original energy loss calculations supplied to you were indeed correct. Afterall, the calculations are often based upon assumptions of the materials used and their insulation properties.

Your method of using average heat loss I'm afraid is flawed, since the heat loss varies throughout the day. Using the original data that you provided, with the spreadsheet that I created, would indicate that at an ambient temperature of 5C, the heat loss could be 8.3kWh, at an ambient of 7C, this would reduce to 7.1kWh, at 10C, 5.3kWh and at 15C, only 2.4kWh. There are days when the ambient air temperature can vary from 15C in the afternoon to 5C during the nighttime. If the 'standard' heat loss is now changed to 9000kWh as you stated in your recent post, the above figues change to 6kWh at 5C, 5.1kWh at 7C, 3.9kWh at 10C, and only 1.7kWh at 15C. There is no standard day.

A further aspect, that none of the measuring methods considers, is solar gain. From observations and test I have carried out over several years, the heat loss can vary quite considerably throughout the day, even when the ambient air temperature remains reasonably constant. Solar gain not only warms up internal areas of your home, but it also warms the outer fabric of the building, which in turn reduces the rate of heat loss.

There are so many factors to consider, that accurately measuring heat loss and calculating COP is almost impossible. What you can do is record any measured values over time, and see if there are any improvements made by changes to how your system operates.

Posted by: @bristoljoe

↑

I found the design calculations from all heat pump installers were huge for us.

I had a similar experience, but the range was wider, some clearly far too high, others definitely too low. I don't think the 12.4kW figure is that far out, limitations of any heat loss calculation accepted. It's very close to my own calculations (using an MCS type approach, ie areas U values etc but not their actual spreadsheet, I first did the calcs long before MCS even existed, for my original solid fuel and then oil fired central heating, tweaking the numbers along the way as things changed) and it is close to the Freedom calculator figure when the latter is filled in correctly (typos can creep in unseen) and it also accords with the energy use when I used oil. It's also in the same ball park as the figures that come out of generic calculators.

That said, there are indeed many other factors. Solar gain is very real and can be felt. Then there are things like wind chill, and not allowing for an inglenook fireplace. Nonetheless, heat loss is a real, but elusive, figure.  

@derek-m - I did say the figures were average, because clearly the actual hour by hour let alone day by day situation varies. The point I am trying to get at is the translation of a derived design temp heat loss, ie the 12.5kW figure in my case, into actual heat demand. Definitely my calcs were wrong, what I am trying to do is get them to be more accurate. I think @bristoljoe is probably onto the right thing, degree days, but I haven't been able to find a clear recipe  for how to use them. A lot of websites have obviously wrong info (eg degree days is the number of days in the year when the temp is below baseline and then they quote figures like 2000, when there are only 365 days in a year). My current understanding is the degree day is the number of degrees below baseline the temp is times the number of days eg with a baseline on 15.5 degrees (the historical UK baseline), two days at 10.5 degrees would come to 10 degree days. Its a measure of how long and how much the heating has to work. More cold days equals more work etc.

I agree calculating COP is a nightmare. I still think my proposed approach, know the cumulative over a period of time heat supply needed to keep the house in a steady state at room design temps, and that is your heat pump output, has some merit, because it avoids all those pesky not to mention expensive sensors not quite telling us the right information, as well as avoiding manufacturers' black box calculations. Because it is cumulative, it can cope with hourly and daily and longer period fluctuations: a warm day followed by a cold day is the same as a cold day followed by a warm day, and both are the same as two average days etc ad infinitum. The question is how to translate a single heat loss figure into a standardised (which in effect is what the degree day bit does) heat loss, and then it is possible to say, providing the room temps are stable, that the heat loss equals the heat input, which must be the case, otherwise the room will cool down or heat up, and the heat input then equals the heat pump output. The key and very simple reasoning is that if a system is in a steady state, as in room temps are steady, then heat in must equal heat out, otherwise the room temp wouldn't be steady, it would change. Unless that is someone has started bending the laws of physics. 

I think it may even be possible to go even further and do local degree days. In a previous post, I mentioned how it is possible to use local open source historical weather data to get hour by hour ambient temperatures. This data could I think be used to calculate my local heating degree days for a particular period (a 10 day cold spell would have more degree days than a 10 day warm spell etc) and I don't think it would be that difficult to set up a spreadsheet to do this automatically once you have worked out the formulas to do it. But what I can't work out how to do is translate the 12.4kW at -2 figure into actual heat loss given the particular ambient on a particular day. The simple linear pro rata (as indicated above) doesn't work. I will keep digging, but if any one has any pointers they will be very gratefully received!   

I fear I am not explaining things very well. Let me try again.

Let's start with a very simple imaginary scenario. Over a 24 hour period, the outside ambient stays fixed at 10 degrees. During that 24 hour period, the heat pump keeps the house at a steady 20 degrees throughout. The question is, given the particulars of my house, how much heat energy (in kWh) does it lose to the garden over that 24 hour period? Because the house is in a steady state, and therefore heat energy out must equal heat energy in, then the loss from the house must equal the heat energy output from the heat pump. Lets say that heat energy lost to the garden over the 24 hour period is 30kWh, and then external meter shows the heat pump used 10kWh over the same 24 hour period, then the COP is 3.

All that is then needed is to integrate the same basic calculation over time as the external ambient temp varies, and you can determine your real world COP.