Im doing some work on a friends heat pump installation which I feel is oversized, but until we get some cold weather I cant be sure how much. However it got me to thinking that if I know the total heating energy each day and the average Outside Air Temperature, I could plot a scattergram of hourly heat loss v Mean OAT, add a trend line and extrapolate that down to -3C.
As a test I tried it with some data from my own installation from last winter. I thought the results were quite useful.
I dont run my heat pump 24/7 so divided the total heating energy by 18 instead of 24 to get my hourly rate. That might not be the correct thing to do but the graph shows my anticipated heat loss at zero outside would be be around 3.5kW, and extrapolating it down further would indicate a heat loss of about 4kW at -3C.
This is well within my 5kW output and ties in quite nicely with my observed usage at the OATs we tend to get near the sea. I just need some more data points on my friends heat pump ......
House-2 bed partial stone bungalow, 5kW Samsung Gen 6 ASHP (Self install)
6.9 kWp of PV
5kWh DC coupled battery
Blog: https://thegreeningofrosecottage.weebly.com/
Heatpump Stats: http://heatpumpmonitor.org/system/view?id=60
I did something similar when I was trying to get to grips with our old Frankensystem. I wasn't able to model it based on the actual heat delivered, however, because I'm not aware of any way to track the heat delivered by a heat pump hot water cylinder unless it actually does it itself.
I do think if you're taking the total daily energy, that you need to divide by 24 for the heat loss, not just the 18 your heat pump is running. In theory it needs to make up for the 6 hours it's not running during the 18 hours it is, so it'll run a little harder in that time to maintain the target temperature (or, alternatively, your heat curve is set a little higher than it technically needs to be if it was operating 24 hours a day).
You're measuring loss, using heat-in as a proxy and assuming a steady indoor temperature. Heat Loss doesn't stop just because your heat pump stopped, it just has be made-up when the unit is running.
Im doing some work on a friends heat pump installation which I feel is oversized, but until we get some cold weather I cant be sure how much. However it got me to thinking that if I know the total heating energy each day and the average Outside Air Temperature, I could plot a scattergram of hourly heat loss v Mean OAT, add a trend line and extrapolate that down to -3C
Thats exactly how I sized my heat pump. Surveys (2 off, each taking 3 hrs, one of which I paid £300 for) said 16kW, measurements of gas consumption said 7kW. The challenge was to persuade installers to take notice, but fortunately I found two were willing to do so. I fitted the Vaillant 7kW (which can do about 8.4kW, although drops to 7kW or even a bit less with defrost). Perfect match.
Personally I think it should be compulsory to compare measured loss with surveyed loss and, if they don't reconcile to sufficient accuracy - enough to resolve any major design decisions - work out why. If installers were obliged to do this we would see far fewer oversized heat pump installations.
@steelbadger Yes. I did think that might be an issue but it also occured to me that my house just might not be as warm as if I heated 24/7.
There is also the point that if I used 24 instead of 18 as the denominator then my rate of heat loss would actually be smaller. So actually I might be erring on the cautious side
House-2 bed partial stone bungalow, 5kW Samsung Gen 6 ASHP (Self install)
6.9 kWp of PV
5kWh DC coupled battery
Blog: https://thegreeningofrosecottage.weebly.com/
Heatpump Stats: http://heatpumpmonitor.org/system/view?id=60
@jamespa I agree. Installers go on about the importance of a room by room survey but how can you have faith in them when they often come up with such different results when done by different installers.
House-2 bed partial stone bungalow, 5kW Samsung Gen 6 ASHP (Self install)
6.9 kWp of PV
5kWh DC coupled battery
Blog: https://thegreeningofrosecottage.weebly.com/
Heatpump Stats: http://heatpumpmonitor.org/system/view?id=60
@jamespa I agree. Installers go on about the importance of a room by room survey but how can you have faith in them when they often come up with such different results when done by different installers.
In my cynical opinion the customer having faith is not the purpose of the survey. The purpose is to protect the installer by following a process set our by MCS. Whatever garbage the survey spits out it guarantees the installer cant be criticised for fitting the wrong pump. That's why only the better installers, who also recognise the survey limitations, will consider any other evidence.
This may sound cynical but why else is there so much resistance to using additional data, and why are there surveyors who will refuse to take into account fabric improvements that they cannot see (eg IWI).
In my case two surveyors came up with the same number, but it was still very wrong. I got the detail of one of the calculations and can reconcile the difference between it and the measured value if I make a reasonable assumption (which several of the better installers have subsequently endorsed) about ACH value, ie one that is less than the MCS worst case figures.
@jamespa I guess a degree of butt covering would help explain it.
House-2 bed partial stone bungalow, 5kW Samsung Gen 6 ASHP (Self install)
6.9 kWp of PV
5kWh DC coupled battery
Blog: https://thegreeningofrosecottage.weebly.com/
Heatpump Stats: http://heatpumpmonitor.org/system/view?id=60
As a test I tried it with some data from my own installation from last winter. I thought the results were quite useful.
I dont run my heat pump 24/7 so divided the total heating energy by 18 instead of 24 to get my hourly rate. That might not be the correct thing to do but the graph shows my anticipated heat loss at zero outside would be be around 3.5kW, and extrapolating it down further would indicate a heat loss of about 4kW at -3C.
We have discussed this technique in considerable depth in other threads. One important thing, given the underlying premise (energy in to the property = energy loss from the property), which I am not sure you have covered, is the IAT (indoor air temperature) needs to be stable over time. If the property gets cooler, the heat loss is greater than the energy into the house, and vice versa. As your data is daily rather than hourly, that means the mean daily IAT needs to be stable over time. If it is not, that may account for the spread in your plot, which is considerable. If you also have the IAT in the data, you could try sorting by IAT and doing the plot for days when the mean IAT was the same or very similar.
I'm not sure about dividing by 18 rather than 24. Each dot on your chart represents the energy put in over 24 hours, even if it wasn't spread evenly over each of the 24 hours. That is therefore the energy you need to put in per 24 hours to keep the property in steady state (assuming it was in steady state), so to get the hourly equivalent rate, I think you may need to divide by 24. not 18.
Another problem is how to deal with cold spells and consequent defrost cycles. During a defrost cycle, heat energy is removed from the house — the energy in to the property goes negative — it may be worth looking at how you get the energy into the property data to see how defrosts are handled. My method is based on my minute data (which is then aggregated into hours), and during defrost cycles the energy out does go negative (because the delta t goes negative, the leaving water temperature is lower than the returning water temperature), meaning I think defrosts are dealt with correctly. An alternative way of dealing with defrosts is to exclude them eg plot for OAT 5°C and above, but then you literally lose sight of what is happening at lower OATs. Here's a such plot of mine (hourly data, setback hours and DHW heating hours also excluded, axes the other way round, independent variable on the X axis, dependent on the Y axis). Even with all those exclusions, things start to go wonky at the lower OATs on the chart:
But I am still reasonably confident my heat loss at -2°C (my design OAT) is around 9kW. The spreadsheet based Freedom Heat Loss calculator's estimate was 12.3kW.
Another way of looking at the design heat loss is to look in detail at a period when the OAT was at, or close to the design OAT. Here's a 12 hour period in February of this year when the OAT was around zero, for around six hours. The IAT is stable, and where is should be, and at that OAT (around zero) the scatter plot suggests the heat loss should be around 8kW (kWh per hour), which by and large it was:
But I can't explain the hour on hour variation, but do wonder whether the defrosts (LWT below RWT) had something to do with them, at least in part.
Midea 14kW (for now...) ASHP heating both building and DHW
One important thing, given the underlying premise (energy in to the property = energy loss from the property), which I am not sure you have covered, is the IAT (indoor air temperature) needs to be stable over time. If the property gets cooler, the heat loss is greater than the energy into the house, and vice versa. As your data is daily rather than hourly, that means the mean daily IAT needs to be stable over time.
I think this can be dealt with without the need for OAT to be stable (which rarely happens). You can either use average OAT over a reasonable period of time (eg a day), or you can use degree days which are easy to get from the internet for a weather station near you. You then plot this against energy output to house averaged over the same duration, but possibly offset in time by up to 24 hrs, to allow for house response time to changing outdoor conditions.
Here is the plot I did before I got my ashp. This uses 2 seasons worth of data and every day is included even if the OAT swing was large. The source energy data (gas consumption) is actually half-hourly, but I have averaged over a calendar day to correspond to the degree-day data. I downloaded the latter from the internet as I didn't have local temperature recording. Its plotted as power to make it easy to translate to ashp capacity, this is just daily energy divided by 24hrs.
Best fit (which is what the plot shows) was with the energy data delayed by 12hrs relative to the degree-day data. This is the graph on which my ASHP was sized. Conscious of the fact that 'professional' surveys said the loss was 16kW, I also collected some additional evidence to provide additional confidence.
If your source energy data is the ashp then you do, as @cathoderay says, need to use energy out (to house) not electricity in, and take account of defrost. Depending on the ASHP that may already happen in the stats presented.
This post was modified 3 weeks ago 7 times by JamesPa
4kW peak of solar PV since 2011; EV and a 1930s house which has been partially renovated to improve its efficiency. 7kW Vaillant heat pump.
I think this can be dealt with without the need for OAT to be stable (which rarely happens).
It is the IAT (Indoor) , not the OAT (Outdoor), which needs to be stable! See my paragraph that you quote, it says IAT.
Using degree days data instead of OAT (Outdoor) for the plot is an option but to my mind it is far less intuitive (I know what an OAT of -2°C is, but, a degree day of 18???), and it introduces an additional step (getting the degree day data), all without any real benefit.
This is the tricky one, as I have previously said, along with how to handle setbacks. If you are using energy out (to the house), you need to know how that number was determined, so you can see how defrosts are dealt with (or perhaps not dealt with). In my case, the energy out to the house is calculated (in a python script) from the raw minute data (flow rate and LWT/RWTdelta t) and fluid specific heat capacity, and does go negative during the nadir of a defrost, because the LWT/RWT delta t goes negative, and I think this does reflect reality (as I said above), eg over an hour a heat pump supplies 8kWh in to the house, but during that hour sucks back 1kWh during a defrost, meaning the net energy in is 7kWh,
Midea 14kW (for now...) ASHP heating both building and DHW
Using degree days data instead of OAT (Outdoor) for the plot is an option but to my mind it is far less intuitive (I know what an OAT of -2°C is, but, a degree day of 18???), and it introduces an additional step (getting the degree day data), all without any real benefit.
I agree, but if you dont have access to records of daily average OAT and do have access to records of degree days (which was my situation and remain easier to access if you have no local records) then you have to use what you can get!
[taking account of defrost] is the tricky one, as I have previously said, along with how to handle setbacks. If you are using energy out (to the house), you need to know how that number was determined, so you can see how defrosts are dealt with (or perhaps not dealt with).
Again I agree, but also again you have to use what you can get!
The intent of my post was just to make the observation that its not necessary to be too prescriptive about the data, there are several options which work that depend on less granular or more accessible data. Some are admittedly better than others, but any are better than nothing. I apologise if I didnt make that clear.
This post was modified 3 weeks ago 4 times by JamesPa
4kW peak of solar PV since 2011; EV and a 1930s house which has been partially renovated to improve its efficiency. 7kW Vaillant heat pump.
I agree, but if you dont have access to records of daily average OAT and do have access to records of degree days (which was my situation and remain easier to access if you have no local records) then you have to use what you can get!
Good point.
Midea 14kW (for now...) ASHP heating both building and DHW
