Posted by: @cathoderayHere are all the charts in one image, to (I hope - you need a large monitor) make direct comparison easier:
This is a comparison, albeit only for one rather unexceptional but not atypical day, of predictions based on observed data (using my regression line) and a model worked up from scratch. I think the comparison is interesting: despite having the worst R squared value, my regression line equation makes the best 24 hour value prediction. If we take the 'true' value to be 35.32 kWh (with the Midea 36 being a rounding error), then my regression over-estimates by 2.6 kWh, while your closest model prediction under-estimates by 3.05 kWh. @jamespa will no doubt correct me, but what I think is going on here is a sort of regression to the mean: the hourly values are a bit out, but over a 24 hour period the 'outness' cancels itself out, if that makes sense. I'm also if I admit it a bit suspicious of the third R squared value...it looks a bit low to me. All thoughts and comments welcome. Please bear in mind the goal is to be able to predict accurately what the 24 hour energy in would have been, had a setback not been applied, which can then be compared to the actual use with a setback applied.
So it would appear that you have already reached a conclusion based on a small number of results.
My prime objective will be to compare the predicted results formulated from the manufacturer's published data tables, to see why they differ from those of a system that has been identified as not operating at its optimum, and see why this may be the case. That could be one reason why the predicted energy input is lower than the actual, assuming that the actual is being measured accurately.
It is always wise not to jump to conclusions.
Posted by: @derek-mPosted by: @cathoderayHere are all the charts in one image, to (I hope - you need a large monitor) make direct comparison easier:
-- Attachment is not available --
This is a comparison, albeit only for one rather unexceptional but not atypical day, of predictions based on observed data (using my regression line) and a model worked up from scratch. I think the comparison is interesting: despite having the worst R squared value, my regression line equation makes the best 24 hour value prediction. If we take the 'true' value to be 35.32 kWh (with the Midea 36 being a rounding error), then my regression over-estimates by 2.6 kWh, while your closest model prediction under-estimates by 3.05 kWh. @jamespa will no doubt correct me, but what I think is going on here is a sort of regression to the mean: the hourly values are a bit out, but over a 24 hour period the 'outness' cancels itself out, if that makes sense. I'm also if I admit it a bit suspicious of the third R squared value...it looks a bit low to me. All thoughts and comments welcome. Please bear in mind the goal is to be able to predict accurately what the 24 hour energy in would have been, had a setback not been applied, which can then be compared to the actual use with a setback applied.
So it would appear that you have already reached a conclusion based on a small number of results.
My prime objective will be to compare the predicted results formulated from the manufacturer's published data tables, to see why they differ from those of a system that has been identified as not operating at its optimum, and see why this may be the case. That could be one reason why the predicted energy input is lower than the actual, assuming that the actual is being measured accurately.
It is always wise not to jump to conclusions.
To be honest I don't see a difference between a 2.6kWh 'error' and a 3kWh 'error' on a value if 36kW. Both are within the quite evident noise.
Also being honest I have somewhat lost the plot and, for now, slightly lost interest. For me the realisation that the system essentially operates 'open loop' (in relation to IAT) has shifted the focus of enquiry from understanding energy in, to understanding why the IAT behaves as it does in response to energy in (however that is measured or predicted). Whilst key details of the heating system beyond the heat pump and the house itself remain uncertain, one can only speculate so far as I can see.
A fairly simple model, representing the house as a single heat store, using the known heat loss rate, assuming a fixed standing heat input from other sources, a fixed standing load when the heating is on only, and taking the reported energy out values as correct ( ie just regarding the heat pump as a source of energy to the house ar the hourly rate measured), is capable of reproducing the measured hourly IAT from June through start of December to within +/-1.5C, with no intermediate 'reset' (so model errors are cuumulative). If you omit any one of these factors, model and actual diverge, and the fact that there is still a +/-1.5c variation suggests that yet another factor (most likely the difference between fabric temperature and iat, which is difficult to model) matters. I haven't bothered to share this in detail because I don't know what it means, other than its clear the house is more complicated than a simple load. However until we do understand the house at least to some degree or have a lot more data from carefully controlled experiment, I can't see how we can distinguish with certainty what part of any mesured 'saving' over a short period is due to setback, and what the consequences of other effects.
Perhaps this is just not a terribly suitable candidate for experimental measurement with our current degree of understanding. Early genetics progress was done largely on drosophila not humans for two simple reasons, they have a simple genome and they reproduce rapidly. In this case neither, sadly, applies.
I will add any insight I can from time to time provided i can do so without being attacked, but I fear that until we have much more data, or data from a simpler example, it may not be possible to reach a firm experimental conclusion on which agreement is possible.
That doesn't invalidate the exercise, it has taught me and I suspect others a lot and I'm sure will continue to do so. And as we don't have much other data, it's for now the best available. I'm just not optimistic that any comparison of model with experiment over a short period of time will yield a conclusion. Having said that I sincerely hope you prove me wrong, and maybe the fact I am saying this will increase the incentive!
For now I remain fairly convinced, based on what I have seen from experiment, model and theory, that its fairly probable that some modest savings can be had by limited setback in some circumstances, but not all. That's not very useful however and I freely admit that. My conviction that the relatively short/uncontrolled experiments which some people report tell us relatively little is much stronger.
For the avoidance of doubt I applaud your efforts so far and, as I say, will endeavour to provide any insight I can so long as i am not attacked for being unable to explain every observed effect in this evidently complex sytem.
@derek-m - I suggest you have to accept what @sunandair had the balls to say recently, that (I paraphrase) it can feel to people who post observational data here that they will have their nuts cracked by theorists wielding sledge hammers (to which I should add, @jamespa has done some welcome partial back-peddling of late). I posted what I thought were some useful charts comparing the predictions from your model - the only predictions that you had made available at that time - with the observed data. This is a fundamental step in the life cycle of a model/theory: comparing the predictions with observations, for without experimental observations to back up a theory, it remains as no more or less than a faith or religion, a set of beliefs held in the absence of evidence, and all too often, it has to be said, the vehemence of the believers will be in direct proportion to the lack of evidence. The problem is not that the laws of physics might be wrong, rather it is the application of those laws that might be wrong.
What was your response when I posted those comparisons? In a word, dismissal:
Posted by: @derek-mCould you now post the un-redacted raw data so that meaningful comparison can be made.
You chose the word redaction (and have used it more than once) to imply I had something to hide. You know full well I do not hide things, I have posted far more, and far more complete, data than anyone else. The point was that you do the predictions blind to the results of the observations, in the same way that a randomised placebo controlled clinical trial is properly done double blind (neither doctor nor patient knows whether the patient is taking the drug or the placebo), because we humans just happen to be humans, influenced by all sorts of things. In your world, you might just as well call the trial a randomised double redacted placebo controlled trial, an absurd and nonsensical use of language.
Your sentence really is quite the masterclass in scorn. Having implied I was hiding data for all the wrong reasons, you then completely dismiss the charts I have posted by not even acknowledging them, and then double dismiss them by calling for a 'meaningful comparison' [in contrast, by implication, to my dismissed and unmeaningful comparison].
And so it goes on relentlessly. I ask what is unmeaningful. You reply with another masterclass in dismissal:
Posted by: @derek-mNothing.
I am used to looking at the big picture, identifying probable problem areas, and then looking in detail at these areas.
You say 'Nothing' and then do another snide dismissal (one looks at the broader picture from on high, taking in the whole with our masterful eye, and then zoom in for detail etc etc [unlike those tiny ants on the ground who can't see further than their own mandibles]).
I then post the compound image with all six charts with a brief and cautiously worded commentary and then CRACK! Down comes another sledge hammer:
Posted by: @derek-mSo it would appear that you have already reached a conclusion based on a small number of results.
The ant has dared to jump to conclusions! Never mind it has done nothing of the sort, the vehemence of the true believer blinds them to what I actually said (small sample...interesting...do a 24 hour comparison...no mention whatsoever of a conclusion). You then treat me to a patronising sermon about, well, not jumping to conclusions.
The problem with your bullying tone - because that is exactly what it is - is that it stifles free speech. It is not only unwelcome and distasteful, it is also counter-productive.
Midea 14kW (for now...) ASHP heating both building and DHW
Could you please post the complete raw data.
Posted by: @cathoderay@derek-m - I suggest you have to accept what @sunandair had the balls to say recently, that (I paraphrase) it can feel to people who post observational data here that they will have their nuts cracked by theorists wielding sledge hammers (to which I should add, @jamespa has done some welcome partial back-peddling of late). I posted what I thought were some useful charts comparing the predictions from your model - the only predictions that you had made available at that time - with the observed data. This is a fundamental step in the life cycle of a model/theory: comparing the predictions with observations, for without experimental observations to back up a theory, it remains as no more or less than a faith or religion, a set of beliefs held in the absence of evidence, and all too often, it has to be said, the vehemence of the believers will be in direct proportion to the lack of evidence. The problem is not that the laws of physics might be wrong, rather it is the application of those laws that might be wrong.
What was your response when I posted those comparisons? In a word, dismissal:
Posted by: @derek-mCould you now post the un-redacted raw data so that meaningful comparison can be made.
You chose the word redaction (and have used it more than once) to imply I had something to hide. You know full well I do not hide things, I have posted far more, and far more complete, data than anyone else. The point was that you do the predictions blind to the results of the observations, in the same way that a randomised placebo controlled clinical trial is properly done double blind (neither doctor nor patient knows whether the patient is taking the drug or the placebo), because we humans just happen to be humans, influenced by all sorts of things. In your world, you might just as well call the trial a randomised double redacted placebo controlled trial, an absurd and nonsensical use of language.
Your sentence really is quite the masterclass in scorn. Having implied I was hiding data for all the wrong reasons, you then completely dismiss the charts I have posted by not even acknowledging them, and then double dismiss them by calling for a 'meaningful comparison' [in contrast, by implication, to my dismissed and unmeaningful comparison].
And so it goes on relentlessly. I ask what is unmeaningful. You reply with another masterclass in dismissal:
Posted by: @derek-mNothing.
I am used to looking at the big picture, identifying probable problem areas, and then looking in detail at these areas.
You say 'Nothing' and then do another snide dismissal (one looks at the broader picture from on high, taking in the whole with our masterful eye, and then zoom in for detail etc etc [unlike those tiny ants on the ground who can't see further than their own mandibles]).
I then post the compound image with all six charts with a brief and cautiously worded commentary and then CRACK! Down comes another sledge hammer:
Posted by: @derek-mSo it would appear that you have already reached a conclusion based on a small number of results.
The ant has dared to jump to conclusions! Never mind it has done nothing of the sort, the vehemence of the true believer blinds them to what I actually said (small sample...interesting...do a 24 hour comparison...no mention whatsoever of a conclusion). You then treat me to a patronising sermon about, well, not jumping to conclusions.
The problem with your bullying tone - because that is exactly what it is - is that it stifles free speech. It is not only unwelcome and distasteful, it is also counter-productive.
Sorry but this and the like is a part of the reason that I am losing interest.
To me its vital, where complex systems are involved (which this clearly is) that experiments are backed up by a theoretical explanation, and vice versa, before drawing inferences. Observational data is fine and doubtless in most cases accurate, but ascribing it to a particular cause without being able to link the two through some coherent explanation is what-iffery of the worst kind. People may well have demonstrated that they can save energy by making a change to a non-optimum system, that's great, but claiming (or implying), without evidence that this is the case, that its due to a specific cause (as opposed to the many other factors involved) is simply poor science. To use a medical analogy its no better than saying 'I gave the patient aspirin, a few weeks later he died, thus aspirin caused the death, thus proving that they are bad tablets'.
A little more scientific curiosity, caution about the interpretation of experimental results, respect for theory, and consciousness of the limitations but nevertheless value of any model, would go a long way to making this a more pleasant and ultimately more productive discussion.
And incidentally I don't believe I have 'rowed back' as claimed above. I am always delighted when energy is saved and I have rarely if ever doubted the experimental observations reported. However this is not a thread about energy saving in general, its specific to a particular (possible) cause of energy saving. I have throughout been sceptical of the interpretations of extremely limited data on a complex and noisy system and, until I see some reason to be otherwise, will remain so. So to return to the medical analogy I will not be dragged down the route of 'I gave the patient aspirin, a few weeks later he died, thus aspirin caused the death, thus proving that they are bad tablets'. I do however understand why some appear, out of entirely understandable frustration with the time it takes to do the job properly, to want to go down that route. That doesn't make it a good one.
@jamespa - I don't doubt your good intentions, but once again your tone is too close to patronising for my liking. I am a doctor who has done some epidemiology, and to suggest/imply that I don't know about the problems of small data sets, limited observations, bias and confounding, association is not causation, the perils of extrapolation and all of the rest of the textbook of basic epidemiology is simply unnecessary. To give you the benefit of the doubt, you may be making these (repeated) observations for the benefit of others, in which case you need to say so ('@cathodeRay, I know you appreciate these things, but for the benefit of others, I need to point out...').
There is also another tenet that hasn't as far as I recall been aired: absence of evidence isn't evidence of absence. This is yet another common problem in epidemiology: there is an effect, but you lack the methodology to detect it. It wasn't until the middle of the last century when Doll and Bradford Hill first developed the prospective cohort study that we had good enough evidence that smoking causes lung cancer, and even then there were generalisation problems (not to mention others), as the study population/sample was British doctors, hardly a representative sample of the whole population. But what do you do? Dismiss the results because they can't be generalised, it lacks an underlying theory to chock it up, or any of the countless objections that could be raised, or do you treat the result as a 'good enough' result?
Epidemiological data is chock full of what you call noise, and, just as we find when studying heat pump behaviour, it is simply not possible to do the ideal experiment, which usually means controlling for everything except the variables of interest. How do you determine whether pandemic lockdowns do good, harm or nothing - or both in equal measure? To answer this questions by experiment, you would need to have two identical countries, infect both with a plague, and then lock one down but not the other, leave them to their devices for a few years, and then come back to determine the outcome. Unsurprisingly, no one has done this, but at the same time we do need a pragmatic answer. What should you do? Do the best you can with the data you have, while at the same time never losing sight of the fact either or both of your data and analysis may be flawed. In this messy world, I use the 'good enough' test, which is analogous to the civil test as opposed to the criminal test in law (balance of probabilities vs beyond reasonable doubt), always accepting that something that passes the balance of probabilities is not as robust as one that passes beyond reasonable doubt. Yes, one day we may have a theory of everything, but in the meantime, we live in a real world, where real decisions have to be made, and sometimes balance of probabilities is all we have. And sometimes, we don't even have that - but lets not go there.
Apologies if that seems like a sermon...
Midea 14kW (for now...) ASHP heating both building and DHW
Posted by: @derek-mCould you please post the complete raw data.
Here it is (E&OE):
Midea 14kW (for now...) ASHP heating both building and DHW
So as predicted i used near 22 (21.9kwh) for the 24hour period compared to just over 13 for the similar weather conditions the previous day, house was up to 23 in kitchen and 22 elsewhere, nice but not necessary. As has been pointed out, my house is proabbly more suitable to setbacks then others so this worked for my house in these conditions. I look forward to next cold snap and i will first try running 9am-9pm and see how house reacts overnight. Thanks to all for valuable input. I have pointed few my friends to this forum to get valuable advice.
Posted by: @newhouse87So as predicted i used near 22 (21.9kwh) for the 24hour period compared to just over 13 for the similar weather conditions the previous day, house was up to 23 in kitchen and 22 elsewhere, nice but not necessary. As has been pointed out, my house is proabbly more suitable to setbacks then others so this worked for my house in these conditions. I look forward to next cold snap and i will first try running 9am-9pm and see how house reacts overnight. Thanks to all for valuable input. I have pointed few my friends to this forum to get valuable advice.
What would have been your energy consumption if you had kept the IAT at 21C?
Judging by the tone of your recent lengthy posts, I can see that there is no point in furthering our discussions. On several occasions I have asked that members refrain from using inflammatory language, but you persist in doing so.
- 25 Forums
- 1,702 Topics
- 36.4 K Posts
- 46 Online
- 1,956 Members
Join Us!
Latest Posts
-
RE: How to make me regret installing a heat pump
My local borough Council amassed this data as the Coron...
By Transparent , 1 hour ago
-
RE: Shutting down heating when out for the day - data
I presume that you heat your hot water at times you are...
By JamesPa , 2 hours ago
-
RE: Daikin Altherma 3 Steady State Operation
@pj_preston The Homely smart controller is connected to...
By Toodles , 3 hours ago
-
RE: E.ON Next Pumped - new electricity tariff targeting as ASHP owners
Thank you @allyfish, as you suggest Cosy (MK2) is where...
By Toodles , 3 hours ago
-
RE: Ecodan Wireless Remote Controller Model PAR- WT50R-E
Perhaps Mars can shed some light on what he was talking...
By Abernyte , 3 hours ago
-
RE: Aira Heat Pump: Stylish Scandinavian Heating
Hi. I too am about to get a technical survey so the thr...
By TimG , 4 hours ago
-
RE: Heat Pump, DHW, UFH and Radiator Install
@jamespa I think you are spot on about the integrated b...
By bontwoody , 4 hours ago
-
-
Thanks for the effort. I will send an email to them and...
By Frits , 9 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 , 11 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 , 1 day 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 , 1 day ago
-
RE: 15 year old Daikin ASHP System
Thank you to all who made suggestion. An update - we ...
By BGF , 1 day 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 , 2 days ago
-
Whilst it seems to work here, this is the message MELCl...
By Morgan , 2 days 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
Latest Topics
-
Ecodan Wireless Remote Controller Model PAR- WT50R-E
By DavidAlgarve 8 hours ago
-
By Mars 24 hours ago
-
Daikin Altherma 3 Steady State Operation
By pj_preston 1 day ago
-
Heat Pump, DHW, UFH and Radiator Install
By bontwoody 1 day 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