Thanks everyone for the information. I can see this is a somewhat debated issue.
I'll try the weather compensation in the next week or so and report back.
I find it tricky to be honest to be logical about what I'm saving, or not saving. My hunch is it will be about the same if I have the ASHP running all day with weather compensation as compared to off-and-on given consumption levels. But I'm not really clued up enough to be much other than pseudoscience.
I don't have a massive amount of information from the Grant unit but I have just setup homeassistant with outside temperature, indoor temperature and energy consumption. I've got an electricity loop on the main electrical cable. Is anyone with a Grant unit using a custom energy monitor (i.e. Shelly pm) and where/how did you wire this?
One thing I have realised is it seems to make no difference if I heat it to 20-21 degrees as opposed to 18.5 because I'm guessing it just gets easier to heat as the house gets hotter. Although we struggled last year to get enough heat in last year when below zero.
This might seem a little hacky but as opposed to having the heating on with the thermostatic control which triggers it for multiple short periods of time to bring it back to temperature to create an automation that keeps it on for at least say 3 hours at a time if the temperature drops? I can see from what I have that there is an initial period of high consumption above 4KW when turning on and then generally it drops down.
My hunch is it will be about the same if I have the ASHP running all day with weather compensation as compared to off-and-on given consumption levels. But I'm not really clued up enough to be much other than pseudoscience.
Have a look at my recent postings showing observed energy use with a setback and an assessment of energy saved. There is a small amount of whatiffery involved (what if I hadn't had the setback), but I have laid out my reasoning for others to assess. I think the general conclusion, based on the limited data available so far, is that a six hour 2100 to 0300 setback in the current Southern England outdoor temperature range does save energy and so costs, at least with my house and heat pump setup, by about 25%. Once I have more data, in particular what happens at lower outdoor temps, I will be able to give a better estimate of overall savings. There is a very minor comfort hit, with my current setup, the room temp at 0700 is about 1 degree below the desired room temp (18 actual vs 19 desired), but I can live with that for now.
I do have to stress that while this is real world data, it is provisional, given the small sample size (six 24 hour periods). I might yet have to eat my words.
Midea 14kW (for now...) ASHP heating both building and DHW
@cathoderay Whilst not for a moment questioning the data, which is really good and really useful, it will be interesting to see if the same outcome occurs when it's colder during the day.
Currently we have (in the southeast) a fairly typical autumn pattern, cold at night and relatively warm during the day. This means that some of the recovery occurs when cop is higher than it would be whilst the setback was in place. I would guess that the savings may reduce on cold days when you are no longer helped out by this effect. This may trigger a seasonal approach to setback. The defrost penalty might also cause some 'interesting' behaviours.
Hopefully you will keep collecting and posting the data! Once we have data from a variety of conditions It might be possible to use this, some logic plus some modelling results to draw some tentative general guidance for the majority who won't or can't do the measurements you are doing! At present all the majority have to go on is heatump 'lore' which we all know to be inaccurate (or out of date) in many respects
Currently we have (in the southeast) a fairly typical autumn pattern, cold at night and relatively warm during the day. This means that some of the recovery occurs when cop is higher than it would be whilst the setback was in place. I would guess that the savings may reduce on cold days when you are no longer helped out by this effect. This may trigger a seasonal approach to setback. The defrost penalty might also cause some 'interesting' behaviours.
I agree, and in the main thread with the data I have been careful to point out this is a very limited data set, less than a week, with nothing exceptional going on weather-wise. I will be particularly interested to see what the defrost penalty is when it rears its ugly head.
Rather than a seasonal approach, which is a sort of proxy for likely OATs, I was thinking of varying the settings based on actual OAT, eg when it is mild, only tweak the WCC end points by a small amount, when it is colder tweak them by a larger amount. Most of this large tweaking will have to happen at the right hand end, as the left hand end is already close to its upper limit (60 degrees), but that is OK (I think...). At the moment, I am still very much in the observing and learning stage.
It may be that this is better presented as a seasonal approach to those who don't want to get bogged down in code, but they will lose the hour by hour fine tuning of the WCC.
The other thing I don't think we should forget is that the current weather is fairly typical for this time of year, meaning that my current estimates of savings are likely to be those achievable when the weather is typical for this time of year, and other similar periods, if that makes sense. We already know that heat pimps start behaving badly in colder weather, and we expect bills to rise during those periods, but if we can make saving where we can, why not do so? I also have a hunch that there will be savings in colder periods, but we will have to wait and see.
The thing that is taking up most of my time at the moment is the DHW heating, which gets 'cancelled' by my modbus writes to the WCC end point settings, and even worse, the cancelling then returns the heating to fixed flow temp, ie takes it off the WCC, which is defo not what I want to happen. At the moment I have to do the DHW manually....grrr.....
Midea 14kW (for now...) ASHP heating both building and DHW
The data that you are providing is extremely useful, because it not only provides a better insight into how heat pump controllers appear to be functioning, but also in the development and testing of modeling tools, since if the modeling tool matches real data, the tool can then be used to investigate how a system is likely to respond under different operating conditions.
Here's the electric consumption Vs heat generation of my Grant 17kW unit - EmonCMS link.
I used a emonTX with CT clamps to measure electric consumption but installed a standalone Open Energy Monitor heat pump meter system as I wanted to calculate the CoP accurately so now have the HP electric monitored by inline energy monitors too. A Shelly PM would work just as well as long as it is rated for the full electric load of the HP.
The HP will always spike at start up, and it is this that you ideally want to limit happening too often as it consumes more electricity if happening too often and increases wear on the compressor.
I had started to get mine to run for multiple hours when at or before 7°C but today noticed it was cycling frequently. Turned out one radiator wasn't getting hot, just so happened to be the biggest radiator in the system, cracked the valve a little more and bingo, it has been running constantly at its lowest modulation since just after 12pm (after finishing a DHW cycle).
It is possible to get some stats from the controller into Home Assistant via modbus, but probably best to just use external sensors (Shelly or OEM depending on what information you want).
I'm interested to see what kind of changes the proposed Homely integration brings, but for my system any major improvements to be had to performance are still to be found in plumbing changes rather than controls in my opinion.
I'm sorry, but I don't have a lot of time for those emoncms how not to do a chart charts. There is no legend, meaning there is no way of knowing what the chart represents: I have to assume, and assumption is the mother of all f-ups. I don't know whether the bars (whatever they represent) are stacked, the more common arrangement, where the upper bar starts where the lower one ends, or overlaid (both start at zero). There appear to be two Y axes, but I have no idea what they relate to, or what the units are. The outer (left hand of the two) Y axis fails to start at zero but fails to point that out - it therefore distorts the picture, a bit like one of Stalin's tractor production is up charts. And God only knows what the numbers at the top are. Electric? Electric what? Flow - flow is normally measured is volume over time, not degrees.
The HP will always spike at start up, and it is this that you ideally want to limit happening too often as it consumes more electricity if happening too often and increases wear on the compressor.
I don't see any evidence for these assertions on your chart (on which I don't really see anything for the reasons given above). More widely, I think both of these frequently made assertions may be old wives tales, on a par with other heat pump 'lore' that may in fact be nonsense.
Cycling is all but inevitable in mild weather, as it is the only way the heat pump can reduce its output sufficiently. That means an awful lot of heat pumps do an awful lot of cycling without any apparent detriment, either to energy consumption or by way of wear and tear. If you look at this chart of my system (feel free to critique, I can get things wrong too!), which shows, at minute intervals for the last few hours, various temperature parameters alongside a lurid magenta amps in line, you can see a spike in amps as the compressor starts up (LWT begins to rise) on some cycles, but the spikes hardly enough to worry about: small amplitude and short duration. I believe this may be due to the way modern compressors are controlled, in effect achieving a soft start. Even the first start up at the end of a setback period, seen at 0300 hours, only draws a small amount of extra amps, an amp or two above normal draw for about 15 minutes.
Midea 14kW (for now...) ASHP heating both building and DHW
I suspect that different heat pumps employ different strategies to control their output, as I do see a spike on my Grant (most times) when it starts up. Certainly in instances where I have a long "on time" this is the pattern I see (Note: it's quite a different pattern in low temperatures, or when there's frequent defrost cycles running).
See attached images for today and 8th Nov. I've obscured overnight times, to hopefully make less confusing already quite busy graphs.
This pattern appears to support the strategy of compressor frequency modulation described in section 8.1 of the Grant Installation Guide.
@mikefl - a useful comparison, even if it is, I agree, a rather busy chart! There is a more marked spike, but there is also a notable dip after the spike, meaning we may want to consider the area under the curve, rather than moment by moment values. And, although it is a bigger spike, it isn't huge, eg the added area under the curve in the left hand chart caused by the spike isn't that large compared to the total area under the curve.
I was vaguely under the impression that modern compressors have a slow start up, ie they wind up slowly, rather than jamming straight away into top gear, precisely to limit start up current draw (and perhaps other things as well). That's why I think the 'start ups carry a heavy current penalty' thing may be a hang over from the older heat pumps, that no longer applies to newer ones.
Midea 14kW (for now...) ASHP heating both building and DHW
I suspect that different heat pumps employ different strategies to control their output, as I do see a spike on my Grant (most times) when it starts up. Certainly in instances where I have a long "on time" this is the pattern I see (Note: it's quite a different pattern in low temperatures, or when there's frequent defrost cycles running).
See attached images for today and 8th Nov. I've obscured overnight times, to hopefully make less confusing already quite busy graphs.
This pattern appears to support the strategy of compressor frequency modulation described in section 8.1 of the Grant Installation Guide.
It is interesting to see how the compressor appears to go up to maximum speed when the actually LWT is quite low when compared to the target value, which is what may be expected to occur.
In the right hand chart did you vary the Target LWT or was this done by the controller?
