it is necessary to replace the 34.2kWh of heat energy lost
@derek-m - I am not sure about this. That's surely the delivered kWhs needed to keep it at a constant temp during the set back, not the energy needed to reheat after a set back. I think you may also have missed the fact the heat loss is on the output side of things (post-COP as it were) while the energy saved is on the input side of things (pre-COP as it were), getting back say 36kWh post-COP only needs 12kWh pre-COP, given a COP of 3, if that makes sense. Or perhaps I am being very thick.
There is also @kev-m's interesting real world measurements, suggesting setback followed by boost does produce real world measurable savings. For higher heat loss buildings, the pro rata savings could, perhaps should, be even greater.
I am pleased that you are questioning my thought processes, since it is always useful to have ideas and theories double checked.
I did state that it was 'heat loss', not energy supplied to the heat pump. If when running under balanced conditions, the building is losing 6kW of heat energy each hour, then that would need to be the amount of heat energy supplied by the heat pump to maintain a constant indoor temperature. When the heat pump stops, this heat energy supply is removed, but the heat loss does not stop, so hence the building starts to lose heat energy.
One of the problems is where in the heat energy gradient are the temperature measurements to be taken.
There is a temperature gradient, from the hot gas leaving the compressor, through the internal heat exchanger into the central heating water, from the water into the heat emitters, from the heat emitters into the indoor air, from the indoor air into the internal fabric of the building, from the internal fabric of the building to the external fabric, and eventually from the external fabric into the surrounding air. All along that route the rate at which heat energy is being dissipated and absorbed will vary, which in turn will cause variations in temperature, and vice versa.
Actually, having just reread Kev's post, it is not an accurate comparison, since if I interpret his words correctly, he carried out the calculation from the data of the 5-9am period, but in WC mode the room was not back to the original temperature until 1pm. The boost did allow the system to achieve the desired room temperature much sooner, but I suspect not more efficiently.
Whilst I certainly hope that the estimated saving of an overnight setback is indeed correct, I would caution about taking one set of results as gospel. There are too many unknowns which can affect the results, such as assumed energy saving, differences in ambient conditions such as air temperature and solar gain.
it was 10 minutes but I've since changed it to 30. This was yesterday (at 30). Not a lot of difference although less stops.
I would suggest trying 60 minutes, since I suspect that the 'time interval' is used to regulate the rate at which the LWT is increased or reduced. By increasing the time interval I believe that it should slow the controller response, and hence reduce the changes in LWT, giving smoother operation and possibly less cycling.
I agree. As long as the discussion remains civilised, as it does on this forum, to everyone's credit, then we can only all benefit. I often find that when I am questioned about things I am supposed to know about, it helps me to understand whatever it is that little bit better, because I have to think about what I think I already know.
I'm still not quite comfortable with the logic that the boost period will need to replace exactly what was lost during the setback period, because the two possible ways of running things - (a) steady state and (b) setback followed by boost - are not the same in how they maintain the building temperature hour by hour. Assuming other things like outdoor temp don't change significantly, in steady state mode, the building stays at the same temp throughout, in setback and boost mode it cools and then reheats. On the face of it, the steady state mode loses more heat over say a 24 hour period than the setback and boost mode, ergo it needs more energy to maintain it.
Consider also this (extreme) example: the setback duration is one week. The building loses heat until it has reached equilibrium with its environment. It won't be a linear decline in heat, probably more a sort of half life type of decline, as the delta t between the building and the environment gets smaller, but lets say it gets to that balance point in three days, by which time it it has lost a substantial amount of heat. Four days later, the heating comes back on with a boost. Will it really need all that lost energy put back in, over and above what it would have used anyway without the setback and boost, once it started up again?
I think the key thing here might be the difference between the energy needed to maintain a building at xx degrees vs the energy needed to raise its its temperature by 1 degree (eg from xx to (xx + 1) degrees). Using boiling a saucepan of water and then simmering it as an analogy, the getting to the boil uses a higher setting on the stove than that needed to keep the pan simmering, but the key question, when applied to setback and boost, is how much extra energy is need to raise the building temp by a few degrees in a reasonable (say 3-4 hours) time frame after a setback period?
Lastly, I agree the assessment needs to be done over a long enough time period, 24 hours would appear to be the obvious interval. Once (if) my heat pump monitoring system gets functional enough, I will be able to run real world experiments, and record 24 hour energy use is steady state vs setback and boost modes, and then compare days when the outside weather conditions were similar. My HA setup already records outdoor temp (from a local Met Office weather station), it can also be set up to record weather (cloudy/sunny etc) and so it shouldn't be too difficult to make sure I compare like days with like days.
Midea 14kW (for now...) ASHP heating both building and DHW
You are indeed correct that the overall heat loss, with a setback, will be lower than constant running, which in my hypothetical example would account to 1.8kWh total.
The point I was probably making badly, was to achieve this quite small heat energy saving required the indoor air temperature to fall by 1.1C.
To recover this 1.1C loss in temperature requires the heat pump to work harder, and operate at a higher LWT, which would be at reduced efficiency.
I feel that it would require quite detailed testing, accurately measuring numerous parameters, to be able to quantify any possible losses or savings. I still feel that greater savings could be achieved by merely lowering the desired indoor temperature by 0.5C on a permanent basis. This is what we have done at home.
it is necessary to replace the 34.2kWh of heat energy lost
@derek-m - I am not sure about this. That's surely the delivered kWhs needed to keep it at a constant temp during the set back, not the energy needed to reheat after a set back. I think you may also have missed the fact the heat loss is on the output side of things (post-COP as it were) while the energy saved is on the input side of things (pre-COP as it were), getting back say 36kWh post-COP only needs 12kWh pre-COP, given a COP of 3, if that makes sense. Or perhaps I am being very thick.
There is also @kev-m's interesting real world measurements, suggesting setback followed by boost does produce real world measurable savings. For higher heat loss buildings, the pro rata savings could, perhaps should, be even greater.
I am pleased that you are questioning my thought processes, since it is always useful to have ideas and theories double checked.
I did state that it was 'heat loss', not energy supplied to the heat pump. If when running under balanced conditions, the building is losing 6kW of heat energy each hour, then that would need to be the amount of heat energy supplied by the heat pump to maintain a constant indoor temperature. When the heat pump stops, this heat energy supply is removed, but the heat loss does not stop, so hence the building starts to lose heat energy.
One of the problems is where in the heat energy gradient are the temperature measurements to be taken.
There is a temperature gradient, from the hot gas leaving the compressor, through the internal heat exchanger into the central heating water, from the water into the heat emitters, from the heat emitters into the indoor air, from the indoor air into the internal fabric of the building, from the internal fabric of the building to the external fabric, and eventually from the external fabric into the surrounding air. All along that route the rate at which heat energy is being dissipated and absorbed will vary, which in turn will cause variations in temperature, and vice versa.
Actually, having just reread Kev's post, it is not an accurate comparison, since if I interpret his words correctly, he carried out the calculation from the data of the 5-9am period, but in WC mode the room was not back to the original temperature until 1pm. The boost did allow the system to achieve the desired room temperature much sooner, but I suspect not more efficiently.
Whilst I certainly hope that the estimated saving of an overnight setback is indeed correct, I would caution about taking one set of results as gospel. There are too many unknowns which can affect the results, such as assumed energy saving, differences in ambient conditions such as air temperature and solar gain.
At least it has everyone thinking. 😀
I know but it depends what you want to compare. I was trying to answer @cathoderay's question, which (I think) was whether a boost on its own would wipe out setback savings. I think what @cathoderay wants is a boost then a return to WC.
Sorry if I sounded a little critical, it was not meant to be so, though I do have to perform my duty as 'Devil's Advocate' from time to time. 😡
Trying to analyse the inner working of an ASHP system is very much a thankless task, since all the necessary data is not available and the accuracy of the data that is available is not known. It is often a matter of making assumptions to fill in the blanks.
You are doing an excellent job carrying out tests and supplying data, for which I am very thankful. 👍
Sorry if I sounded a little critical, it was not meant to be so, though I do have to perform my duty as 'Devil's Advocate' from time to time. 😡
Trying to analyse the inner working of an ASHP system is very much a thankless task, since all the necessary data is not available and the accuracy of the data that is available is not known. It is often a matter of making assumptions to fill in the blanks.
You are doing an excellent job carrying out tests and supplying data, for which I am very thankful. 👍
Not at all, sometimes only I know what I'm thinking and need to explain better!
I think we're all doing a great job with one hand tied behind our backs and one eye closed. 😀
I wonder if there is any way of hacking into the controller to find out what's really going on?
Trying to analyse the inner working of an ASHP system is very much a thankless task, since all the necessary data is not available and the accuracy of the data that is available is not known.
@kev-m - yes, that is correct, all I would add is "with a view to reducing overall energy consumption". That is the whole point of the exercise, no point in doing it if it doesn't achieve that. For want of a better generic name, let's call it load adjusted weather compensation, or LAWC. Some heat pumps (Ecodan) already have it, or something very like it, others (Midea) definitely don't. If we who don't already have it want to have LAWC, we will have to add it ourselves. If/when I have HA running smoothly with both read and write access to the Midea controller, it shouldn't be that hard to do eg IF room desired temp - room actual temp > 3 THEN set LWT = set LWT + 5 ELSE set LWT = set LWT (else bit is redundant but put here for clarity).
It is certainly very possible that setback and boost (SBB) will achieve savings, but they turn out to be trivial, and we may decide it is not worth the bother. The problem is that at the moment we don't have a general answer. The manufacturers' own monitoring of their heat pumps is indeed incomplete and inevitably open to scepticism (dieselgate). If we knew more about the thermal characteristics of buildings (perhaps we do, just haven't got there yet, or perhaps we don't), then we could predict use for both SS (steady state) and SBB modes. Using @derek-m's example, 12kW loss building sitting in a steady 9 degrees temp for 24 hours (all end numbers are energy used, not produced):
SS = (6kWh (estimated hourly loss at 9 degrees) / 3 (COP)) x 24hours = 48kWh used per 24 hours
SBB (with -3 degree setback 2100 to 0500 and a boost recovery time of 3 hours ie by 0800 house is back to design temp) - this has three distinct phases:
(1) 0800 - 2100: (6kWh (estimated hourly loss at 9 degrees) / 3 (COP)) x 13hours = 26kWh
(2) 2100 - 0500: setback temp has been chosen to ensure heat pump will almost certainly remain off during the setback = 0kWh
(3) 0500 - 0800: the boost period: we know that under steady state conditions, the use will be 2kWh per hour x 3 hours = 6kWh. What we don't yet know is how much extra energy is used to accelerate the rate of warming sufficient to have the house back at design temps by 0800. Without including that acceleration energy, SBB has so far used 26 + 6 = 32kWh per 24 hours, compared to SS using 48kWh, which is certainly a worthwhile saving, were it that simple. The crucial question is how much extra energy is needed for the boost.
I am not sure we can reliably calculate it. We would need to know the thermal mass characteristics of the building, specifically how many kWhs are needed (on the used side, not the produced side of the COP equation) to raise the temperature of the building by one degree (and that is very complicated, not only is each building different, there is also ambient weather eg solar gain to consider), and we would also need to know the starting temp of the building at the start of the boost period, and so how many degrees the rise needs to be, and all that involves too many assumptions, and assumptions are the mother of all F/Ups.
But we can measure it, all we need to do is read the dedicated external heat pump electricity meter at 0800 each day, after running the heat pump in each mode. Severe as my HPDHD is, it is not severe enough to make me get up at 0500 every day to set a higher LWT, but I may be able to semi-automate it by (a) setting a timer to turn the heat pump off between 2100 and 0500, and (b) at some time after 2100, before I go to bed, boosting the LWT by lets say 5 degrees (add 5 to each end of the curve) to start with, so when it does come on at 0500, it runs on a boosted curve. Then, at 0800, I manually remove the boost (subtract 5 degrees from each end of the curve).
Of course, weather conditions will vary day to day, but there will be days that are comparable, and over time it should become clearer what the energy use is in each mode, and thereby whether it is worth running LAWC using SBB, because it does cost less than SS mode, or whether it simply isn't worth the hassle.
Thinking about it, I already have the SS data, if I am prepared to accept the Midea estimates of energy used (they are a bit higher than those recorded by the external dedicated heat pump electricity meter). If I accept midnight to midnight, rather than 0800 to 0800, for the 24 hour period, I have daily use data going back to the end of November, the blue line in this chart:
Al I need to do is set up the timer, and remember to adjust the LWT curve twice daily, and within a week or two I should begin to have an idea of how the two modes compare. HA is already recording three hourly weather, though only from this week, but in due course I should be able to derive some sort of 'weather index', something like -2 for freezing fog, + 5 for glorious sunshine etc, and so will have an opportunity to factor that in as well.
I do worry though that all this activity can only aggravate my HPDHD...
Midea 14kW (for now...) ASHP heating both building and DHW
I wonder if there is any way of hacking into the controller to find out what's really going on?
For Midea units, this may be possible, at least for some parts of the system, eg I have found the javascript behind the Midea app. But there are still a lot of black boxes involved:
(a) where are the various sensor, how accurate are they, and how frequently are they read (matters for volatile variables)
(b) the wired controller makes a good number of these variables visible. Does it do any sums on them before making them visible? I already know that the Midea lifetime energy used variable is suspect, because, rather curiously, it is higher than the value recorded on the external dedicated heat pump electricity meter (current readings Midea 9128kWh, external meter 7994kWh, purely by coincidence these are also almost my current annual use, as the heat pump was installed almost exactly a year ago, but it is not necessarily a useful figure, because the heat pump use last spring was all over the place)
(c) the wired controller sends the data to the Midea cloud. Does it get manipulated before transmission, and/or does the Midea cloud do anything to it?
(d) the Midea cloud sends the data to the app on my phone. What, if anything, does the app do to the data? At least we have the javascript for that bit...
Home Assistant and the midea_ac_lan module allow me to query the wired controller locally (it is a query, a request gets sent, it is not passively snooping on the data) but the dataset available is limited, in effect only energy in and energy out at the moment (plus DHW actual temp) but it does mean I can have a stab at estimating a rolling 3 hour (or any other interval for that matter) COP. Here's the current 3 hour one:
This is work in progress, I think the values are on the highish side, and I have no idea why it is so spiky. But it does have some vague credibility, you can just about see it changing slightly in response to ambient temp (higher ambient = lower LWT which then shows up here as better COP).
Come an ideal world, I would have third party sensors everywhere, tracking everything, but that is still a long way off. In the meantime, I want to see whether we can make an assessment of how trustworthy Midea's own data is. If it is 'good enough', then probably no need to re-invent the wheel.
Midea 14kW (for now...) ASHP heating both building and DHW
Thinking about installing a heat pump but unsure where to start? Already have one but it’s not performing as expected? Or are you locked in a frustrating dispute with an installer or manufacturer? We’re here to help.
