@jamespa - this all sounds do-able, and I agree it is worth a try. To cover your various points:
The heating was always on (with or without setback, but it is still on) from 1st Nov to 30th April for the last heating season, so I will pull data for that period.
The scripts run to clock time, or more accurately computer clock time, which usually auto adjusts to BST/GMT - at least that's what I always thought, but now I am not so sure, there is no extra hour from when the clocks went back last weekend, but there are two rows for 0200/0201 and no 0100 row on 31st March 2024. I'm not sure it matters too much though, it is only one hour's worth of data twice a year that might be affected. Power cuts take out more data.
The daily data is logged every hour for the past 24 hours eg 1000 today is 1000 yesterday to 1000 today, 1100 today is 1100 yesterday to 1100 today etc and so can be pulled for any hour of the clock (that's what the 'text as data' queries can do eg the period can just as well be noon to noon as midnight to midnight), but the hourly data has more information in it, so I will pull that data initially, plus the 24 hour data if that would help - let me know.
The simplest setback flag, which is in the hourly data, is the energy in and out drop to zero for six hours, another reason for using hourly data. In the daily (24 hour) data, the setback gets lost in the 24 hour totals.
I did apply a 1.18 correction factor to the script calculated total, heating and DHW energy in values on 29 Nov 2023, and then retrospectively applied this correction factor to the historical data already in the hourly and daily data files (but not the historical minute data, because it does not have those variables, even if buried within the minute data the discrepancy remains, but that doesn't matter, because it is never used uncorrected, if that makes sense). End result is all the calculated energy in variables in the hour and daily data files have the correction applied. The correction factor accounts for the fact the calculated energy in values under-estimate the actual independent heat pump kWh meter. If I remember correctly, the 1.18 factor is an average, in reality the error varies over time. I should probably re-check 1.18 is the right correction, God help us if it isn't! Again. if I recall correctly, I think we speculated at the time that the additional energy appearing on the kWh meter may have been used to supply ancillaries.
The previous paragraph also reminds me that not all energy in goes to space heating, some goes to DHW heating. The amount that goes to each is in the data, there are energy in (and out) values for total, space heating and DHW heating. The allocation is done on the basis of the position of the three port two way valve, which means there may be some run on effects, but they shouldn't be huge.
I've suddenly got rather a lot on my plate at the moment, it may take a day or two to pull the data, depending on how the other things go. I will post it once I have it. Just one thing I need to clarify: when you say "Energy to house, Energy to heat pump", are they what I call energy out (delivered to the house) and energy in (energy consumed/used by the heat pump)?
Midea 14kW (for now...) ASHP heating both building and DHW
Posted by: @jamespaMy first step will be to see if how well I can get the energy to house to correlate well with IAT-OAT
It does correlate well already, at least the OAT/energy out are well correlated, that's the basis of the empirical method of determining heat loss. It will be interesting to see if changing from OAT to IAT-OAT changes this.
Here's the data set I think you asked for (let me know if it is not), AIT (called ambient), two energy in columns (kWh_in columns), one for space heating (htg_kWh_in) and one for DHW heating (dhw_kWh_in), and two more for energy out with same naming conventions, and finally IAT, labelled MD02_tmp (MD02 is the type of sensor). Watch out for the Hours when the DHW heating has been on for part of the hour, usually but not always between 1300 and 1400. That hour will usually also have a heating element, but it is smaller than usual because much/most of the hour was used for DHW heating. I usually exclude them when looking only at space heating, as they tend to distort the correlation/regression. The period is 1st Nov 2023 to 30th Apr 2024. Occasional hours are missing, most often because of power cuts. Note the T in the datetime, which keeps it in text format, otherwise it becomes date formatted and runs a merry old game of its own.
Setback periods are flagged as hours when the energy in values are zero for six hours between 2100 and 0300. I can't explain the htg_kWh_out for the last hour of the setbacks, but it is invariably small and sometimes negative.
E&OE of course, but by and large I think it is a sound data set.
This is a public forum, anyone else can download the data and play with it, if they so wish.
Midea 14kW (for now...) ASHP heating both building and DHW
Thanks, I have grabbed the data and done some work, but want to spend at least a couple of elapsed days so I can think about it very carefully, including thinking about the question - "what is the question" (we know the answer is 42 - what we don't know is what the question is!)
Posted by: @cathoderayIt does correlate well already, at least the OAT/energy out are well correlated, that's the basis of the empirical method of determining heat loss. It will be interesting to see if changing from OAT to IAT-OAT changes this.
I would absolutely expect this because the main control loop is linked directly to OAT and, whilst IAT will slightly affect the output from the rads (and thus the amount of energy the heat pump must supply) this is a relatively small effect for the range of IATs we see. So as the heat pump just varies the FT based on OAT the correlation of energy out with OAT is not surprising.
However I don't think it tells us what we need to know. Because there is no 'room temperature compensation', the house absorbs the discrepancy between the energy it loses and the energy it is supplied with by changing its temperature (we can see that in the data). But the whole point of a heating system is to keep the internal temperature constant except when we want it to vary. So I think what we want to know (in order to determine the effect of setback on energy consumption) is
- what would the energy output/input be if IAT is kept constant 24*7
vs
- what would the energy output/input be if IAT is kept constant for the hours of the day we want it to be warm, and allowed to deviate during the setback and recovery hours
which then begs the question, given that we can reproduce external variables from run to run (and thus cant do a proper control) what is the 'synthetic/pseudo' control we should be comparing with and which, in order that we can generalise results from one house to other houses, attempt to explain?
Of course there are also the human factors namely to what extent will we tolerate a variation in IAT during the hours outside setback/recovery and (taking it even further) does slow and low change our tolerance so in fact we behave differently with a heat pump than with a boiler. These are equally, perhaps more, important but not something we can extract from the data nor expect to explain with thermodynamics.
Of course its entirely possible that, given the human factors, there is no answer to the question (whatever it is) that is anything other than personal. That would be disappointing, but its quite probably the case unless we can define the question rather carefully.
@jamespa - I suspected all along the answer might be 42.
Now we have the answer, we can now work out the question, and I think you have phrased it well:
Posted by: @jamespawhat we want to know (in order to determine the effect of setback on energy consumption) is
- what would the energy output/input be if IAT is kept constant 24*7
vs
- what would the energy output/input be if IAT is kept constant for the hours of the day we want it to be warm, and allowed to deviate during the setback and recovery hours
You then identify two other key things:
(a) the fact that in practice we can't do a proper experiment (controlled trial), instead we have to observe what nature does (observational study). This type of study is extremely common in epidemiology, precisely because we can't control nature let alone human beings. You have to make do with whatever you've got, and sometimes you end up with a sort of pseudo-trial, sometimes called perhaps misleadingly a natural experiment, but underneath it is still an observational study, not a true experiment. The original Doll/Peto smoking and lung cancer study was just such a study. Smokers were the exposed group, non-smokers the control group, and the subjects were doctors, overwhelmingly a male homogeneous group at the time, which helped control to some extent for some of the other variables, but also of course limits generalisability. The thing I am getting at here is that when you are stuck with only observational data, there are ways of analysing the data to make sense of it. You just have to work out how to do that analysis...
(b) the human factor. This is certainly key to the comfort factor, and indeed it is subjective, and it comes into the second half of the original question on the other thread, do setbacks save energy without compromising comfort? But there is also the first part of the question, saving energy, and that is an objective question that at least in principle has an objective answer, providing we can find a way to get at it.
I agree, thinking time at the outset is always a very good idea!
Midea 14kW (for now...) ASHP heating both building and DHW
Posted by: @cathoderay@jamespa - I suspected all along the answer might be 42.
Now we have the answer, we can now work out the question, and I think you have phrased it well:
Posted by: @jamespawhat we want to know (in order to determine the effect of setback on energy consumption) is
- what would the energy output/input be if IAT is kept constant 24*7
vs
- what would the energy output/input be if IAT is kept constant for the hours of the day we want it to be warm, and allowed to deviate during the setback and recovery hours
You then identify two other key things:
(a) the fact that in practice we can't do a proper experiment (controlled trial), instead we have to observe what nature does (observational study). This type of study is extremely common in epidemiology, precisely because we can't control nature let alone human beings. You have to make do with whatever you've got, and sometimes you end up with a sort of pseudo-trial, sometimes called perhaps misleadingly a natural experiment, but underneath it is still an observational study, not a true experiment. The original Doll/Peto smoking and lung cancer study was just such a study. Smokers were the exposed group, non-smokers the control group, and the subjects were doctors, overwhelmingly a male homogeneous group at the time, which helped control to some extent for some of the other variables, but also of course limits generalisability. The thing I am getting at here is that when you are stuck with only observational data, there are ways of analysing the data to make sense of it. You just have to work out how to do that analysis...
(b) the human factor. This is certainly key to the comfort factor, and indeed it is subjective, and it comes into the second half of the original question on the other thread, do setbacks save energy without compromising comfort? But there is also the first part of the question, saving energy, and that is an objective question that at least in principle has an objective answer, providing we can find a way to get at it.
I agree, thinking time at the outset is always a very good idea!
Thinking further, we cant actually control the variables in either case (setback or no setback). So in both cases we have to build a model, and we are then back into comparing a model with a model. I am thinking that there is no way round this and that the key step is to validate the model against the observed data and do some sensitivity testing on the conclusion against variations (uncertainty) in the parameters of the model.
Hmm.
Posted by: @jamespaThinking further, we cant actually control the variables in either case (setback or no setback).
This is why I mentioned epidemiological observational studies, you have no control over the variables whatsoever. The main way you deal with that nowadays is by collecting all the variables that you think are relevant to the outcome of interest from both the exposed group (setback users/smokers) and non-exposed group (no setback/non-smokers) and then do a complicated regression analysis. I'm wondering if we can adapt an epidemiological method that tackles this very problem, no control over what happens, only measurements of what happens. I keep on thinking about it...
By the way, I've just reminded myself that the early papers on the British doctors study into smoking and lung cancer were authored by Doll and Bradford Hill, not Doll and Peto. Very sloppy of me.
Midea 14kW (for now...) ASHP heating both building and DHW
Posted by: @cathoderayPosted by: @jamespaThinking further, we cant actually control the variables in either case (setback or no setback).
This is why I mentioned epidemiological observational studies, you have no control over the variables whatsoever. The main way you deal with that nowadays is by collecting all the variables that you think are relevant to the outcome of interest from both the exposed group (setback users/smokers) and non-exposed group (no setback/non-smokers) and then do a complicated regression analysis. I'm wondering if we can adapt an epidemiological method that tackles this very problem, no control over what happens, only measurements of what happens. I keep on thinking about it...
Well indeed, but I dont think we have any chance of collecting enough data to follow the epidemiological approach. The only people who might are Octopus and British Gas, who both sell energy and heat pumps. I am going to keep going for another day and then share what I have, but I'm already fairly sure that
a) I can come up with a model (essentially of the response of the house to OAT, IAT and energy) that is readily reconcilable with conservation of energy, with adjustable parameters that are empirically determined from the data and that should take into account the main variables in a way that is reasonably justifiable
b) that the uncertainty in the adjustable parameters is such that I could probably produce a wide range of 'answers' from the model (now we know what the question is), probably too wide to be useful
c) that there remain some material unexplained anomalies which cast doubt on any conclusions
d) that uncertainty in actual OAT is very definitely not helping, but knowing it for certain might not resolve the unexplained anomalies
I acknowledge that that's not particularly helpful.
The good news (perhaps) is that, if the problem is in fact so difficult to solve, there will be people who want to 'buy' our data in order to solve it using the epidemiological result.
@jamespa - first to deal with the point I raised in @bontwoody's thread - the flow rate drops to zero at the start of the setback and stays that way until the end of the setback. The slow decline in LWT/RWT during the setback happens at the heat pump, but no extra heat gets to the house because the flow rate is zero, no extra heat because the rads will still transfer a bit more heat during the early part of the setback, but that energy has already been accounted for, having been put into the house before the setback.
Methods: certainly more data is usually better, and certainly big firms like Octopus and British Gas will have huge datasets (n = very large number), but they will limited in scope, because the only way they get data is through their spy smart meters. I think at best this is half hourly energy in. They don't know energy out, OAT, IAT, how the heating run (weather comp? thermostat controlled? setback or no setback), occupant numbers and habits etc etc) and to that extent we have much better data than the big firms. Furthermore, one of the advantages of n=1 is that my house becomes it's own control, which means a lot of the material (fabric) variables are controlled for.
On one level then n=1 but at the same time I have hundreds of 'subjects', a 'subject' being a 24 hour period. Just as 10 person-years can come from either one person over ten years or ten people for one year, we could have 100 houses all collecting data for one 24 hour period, or one house collecting data over a hundred 24 hour periods, we still have the same amount of data. Thus a single heating season will generate hundreds of 'house-days'.
OAT - consider the 48 hour chart I posted recently:
The green line labelled OAT is actually the IAT, and it shows the expected step pattern, with the AIT cooled when the heat pump runs. I think it may be possible to correct that to a better estimate of the OAT using something like a 'moving maximum' eg OAT(t) = max(AIT(t-3) to AIT(t+3)) or something similar. I am also going to put my stand alone temp data logger in an improvised Stevenson screen (coal bunker on the opposite side of the house) and see what it records. It only records data hourly, but it will give us something to compare with the AIT.
I still see this challenge (as formulated in your question we are trying to answer) as something worth pursuing, both because it is an interesting intellectual challenge in its own right, and also because the answer, if we can get it to our satisfaction, has real world utility.
Midea 14kW (for now...) ASHP heating both building and DHW
Posted by: @cathoderayOn one level then n=1 but at the same time I have hundreds of 'subjects', a 'subject' being a 24 hour period. Just as 10 person-years can come from either one person over ten years or ten people for one year, we could have 100 houses all collecting data for one 24 hour period, or one house collecting data over a hundred 24 hour periods, we still have the same amount of data. Thus a single heating season will generate hundreds of 'house-days'.
Yes but -they are all for the same house which means we dont have any experimental handle on how general the conclusions may be. On the other hand we dont have to deal with the differences between houses so its a good place to start.
Posted by: @cathoderayThe green line labelled OAT is actually the IAT, and it shows the expected step pattern, with the AIT cooled when the heat pump runs. I think it may be possible to correct that to a better estimate of the OAT using something like a 'moving maximum' eg OAT(t) = max(AIT(t-3) to AIT(t+3)) or something similar.
Agree in principle however... I presume this is during heat pump cycling, obviously when its not cycling the correction wont be available.
Posted by: @cathoderayI am also going to put my stand alone temp data logger in an improvised Stevenson screen (coal bunker on the opposite side of the house) and see what it records. It only records data hourly, but it will give us something to compare with the AIT.
I was hoping you might do this. My worry is that the OAT-AIT relationship is not stable because it depends on how hard and how continuously the heat pump is working, a factor which I would prefer not to contemplate adjusting for. If we have a couple of weeks of independent OAT measurements we can get a feel for it, although I would also want to see what it does when its very cold and also particularly when its in defrost mode.
Posted by: @cathoderayI still see this challenge (as formulated in your question we are trying to answer) as something worth pursuing, both because it is an interesting intellectual challenge in its own right, and also because the answer, if we can get it to our satisfaction, has real world utility.
Agreed. Im currently focussed exclusively on a model of the house and at present tidying up so it can (a) be understood by others and (b) can be understood by me when I come back to it. Once we have a convincing model of the house (which I dont think we will have this time round, but might just once we understand OAT vs AIT better) we can think about how to translate it to the heat pump.
It will be a couple of days before I post anything publicly (but I think its going to turn out as I speculate above) but I will PM you with the current status.
Perhaps I’ve missed key points in this discussion but why would the OIT-IAT vary with how hard the pump is working in the general case?
I can understand why it would if the outside sensor is within the pump case, but that’s not best practice and ours is many feet away on a north facing wall.
2kW + Growatt & 4kW +Sunnyboy PV on south-facing roof 9.5kWh Givenergy battery with AC3. MVHR. Vaillant 7kW ASHP (new & still learning it)
Posted by: @judithI can understand why it would if the outside sensor is within the pump case, but that’s not best practice and ours is many feet away on a north facing wall.
In this case it's in the case. There is also conclusive evidence that it reacts to what the heat pump is doing.
@judith - as @jamespa says, my Midea unit has its OAT (Midea call it 'Outdoor ambient temperature' which originally I thought meant what it said) sensor in the case, so it is actually measuring something more like the AIT (air intake temperature). If you look at my chart above (first one you come to if you scroll up), see how the green line labelled OAT (used data from this sensor) steps up and down line line with heat pump activity as seen in the LWT/RWT lines.
I do now have the separate RC-4 temp data logger in a well ventilated coal bunker and in a few days time we will have a better idea of the hourly true OAT, and can then compare that to the Midea reported OAT.
Midea 14kW (for now...) ASHP heating both building and DHW
- 25 Forums
- 1,699 Topics
- 36.3 K Posts
- 58 Online
- 1,954 Members
Join Us!
Latest Posts
-
RE: Shutting down heating when out for the day - data
Interesting. It is still worth turning the heating off ...
By jamespetts , 5 mins ago
-
RE: How to make me regret installing a heat pump
@sunandair Theres a lot going on here in this post L...
By Lucia , 7 mins ago
-
Ecodan Wireless Remote Controller Model PAR- WT50R-E
I am struggling to find a reliable Mitsubishi contracto...
By DavidAlgarve , 42 mins ago
-
-
Thanks for the effort. I will send an email to them and...
By Frits , 2 hours ago
-
RE: Vaillant Arotherm monobloc error f.788 - ASHP is not working!
@editor We have a 200 litre buffer between the hydrauli...
By bobshopsupreme , 3 hours ago
-
RE: Installing a heat pump in a Grade II listed property
@cathoderay Yes, it is a plotly plot and I suppose the ...
By Toodles , 18 hours ago
-
RE: Heat Pump, DHW, UFH and Radiator Install
@jamespa Its a lot more than I had managed to work out ...
By bontwoody , 19 hours ago
-
Daikin Altherma 3 Steady State Operation
New Heat Pump Owner and First Post on the Forum We ha...
By pj_preston , 19 hours ago
-
RE: In general - are shorter hotter cycles more costly or longer cooler cycles?
Good idea. You will probably find that the cycling 'pr...
By JamesPa , 20 hours ago
-
RE: 15 year old Daikin ASHP System
Thank you to all who made suggestion. An update - we ...
By BGF , 22 hours ago
-
Hot Spring Ground Source Heating
Anyone who help?I have access to a hot spring, temperat...
By AllenC , 1 day ago
-
RE: Samsung ASHP scheduling help
@minted_lamb news to me. I'll probably end up using the...
By morh , 1 day ago
-
Whilst it seems to work here, this is the message MELCl...
By Morgan , 1 day ago
-
-
@scotty229 I'm sorry to hear of your issues. It looks a...
By Mike H , 2 days ago
-
RE: Is your heat pump insured?
@declan90 I tried that after the phone call and they di...
By benson , 2 days ago
-
RE: Is my Ecodan well installed and setup?
Eh? It's just a coil of wire.How the heck do you get ...
By Transparent , 2 days ago
-
RE: How to set up Home Assistant for GivEnergy?
Not hard work at all, @johndunlop, but rather an exerci...
By Majordennisbloodnok , 2 days ago
Latest Topics
-
Ecodan Wireless Remote Controller Model PAR- WT50R-E
By DavidAlgarve 42 mins ago
-
By Mars 16 hours ago
-
Daikin Altherma 3 Steady State Operation
By pj_preston 19 hours ago
-
Heat Pump, DHW, UFH and Radiator Install
By bontwoody 22 hours ago
-
Hot Spring Ground Source Heating
By AllenC 1 day ago
-
Shutting down heating when out for the day - data
By jamespetts 3 days ago
-
How to make me regret installing a heat pump
By Lucia 3 days ago
-
By Toodles 3 days ago