@jamespa

The baseline was 0C, with the estimated value being at an IAT of 20C, so that would give 13kWh/C.

Note that this was based upon the fall in IAT during just one setback period, I can't remember which date I used.

Posted by: @derek-m

↑

@jamespa

The baseline was 0C, with the estimated value being at an IAT of 20C, so that would give 13kWh/C.

Note that this was based upon the fall in IAT during just one setback period, I can't remember which data I used.

At this stage I will very happily settle for the same to within a factor of 2, given that they are from two completely independent approaches based on very limited data and only a developing understanding of the how this all works.

Posted by: @jamespa

↑

Thanks for the data health warning. 

I'm not sure I like the tone here! Seriously though, data is data, and while it has some known and correctable flaws eg the energy in as calculated from the raw data doesn't match that measured by the independent kWh meter, the weaknesses, if you want to call them that, are more to do with representativeness, eg I have that PHE, and so the data is a sample from the population of heat pumps with PHEs, not all heat pumps. I would also like to remind you that however weakened it is, it is the best data you have got, because no one else has recorded and published their raw data. It also only applies to leaky old buildings with nonetheless high thermal mass, in weather conditions as they actually happened, etc etc etc but these limitations are inevitable.     

Posted by: @jamespa

↑

Id by happy with the raw minute by minute data as you posted for the shorter period.  Preferably just a couple of months though, or alternatively (if you think its representative) I could probably cope with Sept-Mar last year, ie one season.  The formulae are all set up so its really just a copy and paste exercises once the data is imported (which I do via libre office because it natively recognises the date-T-Time format and translates it to a date number, whereas my version of excel currently doesn't).

I said earlier the only complete data set starts from late September this year, when I added the modbus IAT sensor, in time for this heating season. The standalone IAT data logger I mentioned is problematic: it's a different measurement device, in a different location, and its data format (hourly) is not easily merged with the minute data (the 'inner join' problem I mentioned earlier). Bottom line is there isn't a full data set with IAT until we get to September this year, without me doing an awful lot of very tedious work which I am not prepared to do.

I also dislike multiple data sets, and prefer one master record (ie file, with a rigorous back up schedule) from which data sub-sets can be pulled and worked on. The data isn't so huge that this can't be done, and I can always split it into annual chunks if it does get out of hand. I really do need to correct the current master minute data set so all the records in it are comparable, ie have the 1.18 correction factor worked into the data. I will try to do that later today and then post the complete data set from late September this year to date.

I sometimes do a search and replace to remove the T from the datetime and replace it with a space. Best done in a text editor like Notepad++, the moment you put the data in a spreadsheet, the spreadsheet will start messing with the date time formats.

If you think the data is 'very limited', then perhaps I should remind you what that really looks like. It looks like this:

,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

repeated forever, to the end of time.

Posted by: @cathoderay

↑

Posted by: @jamespa

↑

Thanks for the data health warning. 

I'm not sure I like the tone here! Seriously though, data is data, and while it has some known and correctable flaws eg the energy in as calculated from the raw data doesn't match that measured by the independent kWh meter, the weaknesses, if you want to call them that, are more to do with representativeness, eg I have that PHE, and so the data is a sample from the population of heat pumps with PHEs, not all heat pumps. I would also like to remind you that however weakened it is, it is the best data you have got, because no one else has recorded and published their raw data. It also only applies to leaky old buildings with nonetheless high thermal mass, in weather conditions as they actually happened, etc etc etc but these limitations are inevitable.     

Posted by: @jamespa

↑

Id by happy with the raw minute by minute data as you posted for the shorter period.  Preferably just a couple of months though, or alternatively (if you think its representative) I could probably cope with Sept-Mar last year, ie one season.  The formulae are all set up so its really just a copy and paste exercises once the data is imported (which I do via libre office because it natively recognises the date-T-Time format and translates it to a date number, whereas my version of excel currently doesn't).

I said earlier the only complete data set starts from late September this year, when I added the modbus IAT sensor, in time for this heating season. The standalone IAT data logger I mentioned is problematic: it's a different measurement device, in a different location, and its data format (hourly) is not easily merged with the minute data (the 'inner join' problem I mentioned earlier). Bottom line is there isn't a full data set with IAT until we get to September this year, without me doing an awful lot of very tedious work which I am not prepared to do.

I also dislike multiple data sets, and prefer one master record (ie file, with a rigorous back up schedule) from which data sub-sets can be pulled and worked on. The data isn't so huge that this can't be done, and I can always split it into annual chunks if it does get out of hand. I really do need to correct the current master minute data set so all the records in it are comparable, ie have the 1.18 correction factor worked into the data. I will try to do that later today and then post the complete data set from late September this year to date.

I sometimes do a search and replace to remove the T from the datetime and replace it with a space. Best done in a text editor like Notepad++, the moment you put the data in a spreadsheet, the spreadsheet will start messing with the date time formats.

If you think the data is 'very limited', then perhaps I should remind you what that really looks like. It looks like this:

,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

repeated forever, to the end of time.

 

No offence intended, the comments were meant simply to reflect the facts that:

1. when developing a model for a class of system, its often a good idea to start with relatively simple examples of the class, and work up to more complex examples, if that is at all possible.  I would ideally prefer to do this exercise first on a system with no phe or buffer because its a further complication (which can doubtless be adjusted for in principle but which, in the early days, it would be better to avoid if possible).  However as a more complex system is the one for which data is in fact available, then that's what I will use.

2. so far I have analysed only a limited subset of the total data available, so my analysis is based on limited data.  That was not intended to imply that that the data (in principle) available is itself 'limited'.

Im happy to look at the whole dataset from when it became complete, if you are willing to post it.  Sorry I misunderstood, I thought you were saying that was from September 2022, not September 2023, hence my comment about a season.

LibreOffice seems quite happy with the 'T', it turned the whole string into a date number on import (with time as a fraction), whilst displaying it in the original format.  Having the date/time represented as a number is handy, because it allows date/time arithmetic to be done easily.

@jamespa - I've decided to keep the raw minute data as it is, all corrections (the 1.18 for all derived/calculated energy in kWh and COP values) will be done only in the hourly and 24 hourly data csv files, meaning I can post the minute data as it is:

To maintain comparability, and increase accuracy, any calculations to do with energy in should be increased by the 1.18 correction factor. Personally, I do this for the kWh values when calculating them from amps x volts, then any subsequent calculations eg COPs will have the correction already baked in.

The 1.18 is a best available estimate, based on comparing previous energy in values without the correction factor to the actual external heat pump kWh meter values. The actual correction factor varies slightly over time, I have not managed to spot any particular reason eg how hard the heat pump is working, and the 1.18 estimate is an average value over those minor variations. 

Yes, we would all like all sorts of things in an ideal world, but pragmatists must work with what is already available, or can be obtained without losing an arm and leg and eventually our heads to get it. I have never claimed my data is perfect, far from it, but it currently remains as far as I know the only publicly available detailed minute by minute real world heat pump data set available. As long as the data survives sanity checks, and the limitations are borne in mind - and that does not mean screaming them from the roof tops at every opportunity, they just need to be borne in mind - then that data set remains better than nothing. A lot better than nothing.

@jamespa - I'm not having a good day today! I have just noticed there were two copies of the minute data script running from about 1350 today, meaning there are duplicate entries in the data file I posted earlier from about then to the end of the file. Either delete them using a regex in Notepad++ (Find ([^\n]*\n)[^\n]*\n Replace $1) or just delete the whole duplicate section, it's hardly a large part of the total data set.

Having tidied up and double checked my data, here is the last 48 hours, perhaps a useful illustration of how ruinous cold outdoor air temps and associated defrost cycles are to performance: 

Up until 0100 last night, when the outside air temp started to rise, the heat pump was running flat out, and shooting itself in the foot with defrost cycles, resulting in both very high hourly kWh energy in, and dreadful COPs, often less than 2. Once the outside air temp got to 4 degrees, the the defrost cycles stop, the energy in drops and the COP climbs to a more respectable above 3 value, currently 3.45.      

@jamespa

Let me know if this arrives at your end.

@jamespa

Let me know if this arrives at your end.

Posted by: @derek-m

↑

@jamespa

Let me know if this arrives at your end.

yes it has thanks.  I was apparently logged in, but I logged out and logged back in and this magically appeared

Pray tell, what is this magical this?

Posted by: @cathoderay

↑

Pray tell, what is this magical this?

It was a 'fishing hook' to see if anyone would bite. 😋