My observation about the discrepancy of the onset of cycling between the model and the measurements was actually not questioning either. The fact the (real) heat pump cycles at all at 4C suggests that something either temporary or permanent is awry. Temporary factors could include solar gain, cooking or something else which temporarily reduces the demand. Permanent could simply be that it is a bit oversized. Perhaps the discrepancy is due either to one of those temporary factors or due to a mismatch between the modelled house loss and the actual house loss.
My first instance would be to check the (real) power output at which the real thing started cycling. If this is consistent with the manufacturer data then then next question becomes, why is this so low in the specific conditions. Only then would I think you are in a position to examine why the model hasn't reproduced this. But all of this is likely obvious to you anyway so I apologise if Im telling you something you already knew.
The reason that I mentioned the cycling was because of the fact that it makes doing an accurate comparison between theoretical and actual that more difficult.
The excellent raw data, at 1 minute intervals, clearly shows that rather than being kept constant, the LWT is performing a 'sawtooth', as the heat pump heats the water from the high 20's / low 30's to the high 30's / low 40's, then stops running whilst the water dissipates the thermal energy and cools.
Averaging the data over 1 hour periods, leads to the possibility of inaccuracies, which could accumulate in the final results. It is therefore important to try to minimise any errors.
I have various theories as to why there may be a divergence between actual and theoretical, the primary one being that in the same manner as most homes in the UK, the data was collected during a period of 'human infestation' 😋, which of course will have upset the balance in the 'actual's' favour. Obviously removing all the humans from the equation would not be practical or even desireable, so making an allowance within the Modeling Tool may be the best option. I believe that the average daily electrical energy consumption (unless used for heating), is in the region of 8kWh to 10kWh, most of which may probably be converted into thermal energy within the home.
What are your thoughts? Should an hourly additional heating value be added to the Energy Supply figure within the Modeling Tool? If so, should this be spread evenly over the 24 hour period, or when humans are likely to be more active?
Decisions, decisions. 😜
In answer to the last question, emphatically no in my view.
More generally there will always be an offset between the modelled net heat loss and the actual net heat loss, and human activity is just one of the several contributions. There is no value so far as I can see in separating them out and attempting to model them makes the model ever more complex with little chance of adding insight. So my view is that the human contribution is just rolled up into 'unknown divergence between theoretical loss and actual loss'. Frankly there are much larger contributions due to unknowns (or ignored) fabric features, so it's doubly not worth bothering.
More generally still, I think one should deliberately limit the extent to which 'second order' effects and beyond are replicated in a model. If you don't then the model becomes as complex as the real world and you lose the insight into the main effects which is the value of the model.
So for any disturbance from simplicity, I think it's necessary to assess whether modelling it is a) necessary to answer whatever question you are trying to resolve and b) likely to add insight. Unless it passes both of these tests I wouldn't bother (more to the point it's actually counterproductive to bother).
This post was modified 8 months ago 2 times by JamesPa
Exactly, but almost no one does it! I suspect most Freedom supplied installers rely on Freedom's heat pump calculator which can get it wrong, and there is no way the average householder in the market for a heat pump is going to dig out the data from obtuse engineering handbooks, nor should they have to, even if they could make any sense of it. I only learnt (entirely thanks to this forum) about the engineering data after my heat pump had been installed...
Imho any installer/specifier who doesn't do this is an incompetent fraud. It's a pretty basic requirement to match the actual declared capacity of the chosen pump at the design conditions, to the calculated load. In fairness to many of the installers I have so far spoken to in this arduous journey, the majority do seem to consult the capacity tables.
The householder shouldn't have to do this, I agree, but the designer absolutely should. This is really basic stuff!
@jamespaTo make things simple, perhaps there should be a law requiring manufacturers to publish accurate performance data that is arrived at during standard test conditions so that we can compare … y’know, like the published data for car emissions … err… oh, perhaps not then!😉 Regards, Toodles.
Toodles, 76 years young and hoping to see 100 and make some ROI on my renewable energy investment!
@jamespaTo make things simple, perhaps there should be a law requiring manufacturers to publish accurate performance data that is arrived at during standard test conditions so that we can compare … y’know, like the published data for car emissions … err… oh, perhaps not then!😉 Regards, Toodles.
Some manufacturers do publish operating data performed during standard test conditions. That is the source of data behind the Modeling Tools I have created.
As far as I am aware, the testing facilities are independent of the manufacturers, and perform internationally specified standard tests.
The fact that the OAT reading appears to increase by approximately 5C during the setback period, would make me suspect that your heat pump may be sitting in a 'cold well'. Where is it located?
Could you please provide a copy of the raw data for analysis.
Thanks, and good to be on friendly terms again.
The heat pump is on an east facing wall, and although it more than meets the minimum distances between it and other objects, it is a little enclosed by vegatation, hedges), and I think that means the cold exhaust air can't disspiate quite as fast as it would in a fully open setting, and that is what cools the air when it is running, and that is probably why the oAT rises during the setback - the cold air is turned off. There may also be another more natural explanation, the air temp just got warmer itself and then cooled down again! But I think the steepness of the transitions makes that highly unlikely.
Let me know how much data you would like, and I will post it here on the forum. Please bear in mind I haven't yet done added the correction factor for the calculated energy in values, I want to wait until Monday when I will have three weeks data to compare, including the current cold spell (up till now it has only been mild OATs data), as i don't yet know whether the same correction factor applies at all OATs. It is also going to be bit of a challenge doing it, as I have just the 60 sec window in between the data readings to do the update, in which I have to swap the files and stop and restart the script.
One thing that stands out is that your flow temps drop a lot further than mine during the setback.
Agreed, and it means the have further to recover. As the heat pump is fully off during the setback, the circulating pump should not be running, meaning the temps we see on the chart should be those for the LW and RW in the pump.
I wonder if you have a way of positively detecting defrost cycles? I use the reversal of the LWT/RWT difference, but you only have LWT (flow temp) on your chart. I do wonder id the deep dip and rise shortly after 0600 is a defrost, all the more so as I think I can just about see a concurrent uptick in the OAT. The first down spike before 2100 also has the same characteristics. Certainly defrosts are ruinous to the efficiency of my heat pump.
The householder shouldn't have to do this, I agree, but the designer absolutely should. This is really basic stuff!
I think the problem round here might partly lie with Headroom sorry I mean Freedom Heat Pump's Pile them High and Pump them Fast approach, everything is aimed at the quick installation, from the spreadsheet they provide to do the calculations (which in fairness does include come manufacturers data but unfortunately they had to guesstimate for Midea heat pumps) through the kits the provide to the 'Quick Installation Manual' they give the installers. I'm also extremely grateful to my installer, who pulled out all the stops to get the installation done in time to meet the grant window deadline, which is to say I think the problem lies with the modus operandi that Freedom advocate.
I've previously suggested we should call the scandal of selling 14 kWh pumps that can only manage that sort of output when you don't need the heating on, and only manage much less when you really do want the heating on, should be called compressor-gate. Unfortunately the Daily Mail haven't yet picked up the story...
Midea 14kW (for now...) ASHP heating both building and DHW
I have had a closer look at the raw data that you supplied previously, since it contains 3 overnight setbacks. Just looking at the data itself does not highlight that there may be a problem, it is only when presented in graphical form do the pieces start falling into place.
When the heat pump is operating, the OAT sensor should be measuring the ambient air temperature since it should be located on the intake side, prior to the evaporator. It is also a fact that on at least some heat pumps, when the fan stops the OAT reading can increase slightly because it is being warmed by residual thermal energy leaking out of the heat pump. In an extended the OAT sensor reading should fairly quickly cool to show that true ambient temperature once more.
The fact that today's graph showed a temperature rise, but no corresponding fall, caught my attention, but more importantly was the quite large temperature fall when the heat pump restarted. Assuming that the sensor is correctly installed in the appropriate position, this would indicate that the air going into the heat pump is being cooled by the heat pump itself, which of course could be quite a serious problem.
Looking at the 3 setbacks mentioned previously was not conclusive, but your mention of shrubbery and hedges may hold the key. I suspect that how the heat pump operates and how the OAT reading may vary, could be influenced by the wind speed and possibly direction, in how effective the cold are being exhausted is being dissipated and not being drawn back into the heat pump.
I would suggest measuring the air temperature around the heat pump and elsewhere, to look for variations. Also see if you can obtain wind speed and direction data for your location, for the 4th, 5th, 6th and 7th of November.
If my theory proves to be correct I would suggest some extensive pruning, or building some form of intake duct to draw warmer air from outside the cooler zone.
my defrosts are at 5pm, 8:15pm 6:30am and just after 9am. The giveaway is the sudden drop in LWT corresponding to a blip in OAT. I assume the temp sensor is affected. I can check the LWT and RWT together on the Melcloud app if I want.
IMO defrosts are the Achilles heel of current ASHPs. Look at this close up of a defrost on a cold morning earlier in the year. You can see the COP dropping as it ices up (same energy in, falling energy out). It stops heating for 5 minutes at 6:46, during which time the house will be losing heat. It then takes 25 minutes to raise the flow temp back to where it was, which takes it to 7:10. On this particular morning it does it again at 7:57 and repeats through the whole day.
I don't think there is anything unusual about my ASHP of set up; they all do it.
the air going into the heat pump is being cooled by the heat pump itself, which of course could be quite a serious problem
I think this is very germane. My heat pump, and I am sure mine is not alone, is operating in a micro-climate of its own making, and sets off on a downwards spiral. As I mentioned earlier, if you stand by the heat pump when it is running, you can feel the air is chillier than the air a few yards away from the heat intake, meaning the air drawn into the heat pump, and that measured by the ambient OAT sensor, will be chillier.
I have been thinking about setting up an OAT sensor away from the heat pump, but as I mentioned earlier, I haven't yet found a suitable sensor. The one I use for the IAT is definitely not waterproof, and I can't think of a location where it would be outside, but not exposed to the elements. The other sensors I have come across all require a higher voltage than I have available on my modbus cable, which is 5V, as it comes from the USB port. Although such a sensor would be very useful for monitoring, I don't think it could be used to over-ride the heat pump's internal OAT sensor, as the latter would constantly over-write the former.
see if you can obtain wind speed and direction data for your location, for the 4th, 5th, 6th and 7th of November.
I think I already have at least some data for this: HA records data for two Met Office weather stations that are about equidistant from my location, but they are not that near (15+ miles away). Nonetheless, if they show similar wind speeds and directions, then it is likely I had similar winds here. There is a WOW hobbyist station nearer, but it doesn't record all parameters. Nonetheless, I can at least indicate the general direction and strength of the wind my location for the dates of interest using HA's History Explorer (which, despite my normal remarks about HA, is really rather good):
Moderate winds throughout, with W/SW prevailing. With winds in the west, my heat pump is in the lee of the house.
I would suggest some extensive pruning, or building some form of intake duct to draw warmer air from outside the cooler zone.
Neither are really an option. The building is listed, which meant I had to get both planning permission and listed building consent (being listed means permitted development doesn't apply) and given the heat pump is on the front elevation, the application relied on, and was successful because of, the visual screening provided by the hedging, and the heat pump itself was chosen because it is low profile (unlike most heat pumps, the Midea ones are wider than they are tall) and so the visual impact has been kept to a minimum. Adding ducting is definitely not going to match the character of the building one bit! There is also no obvious place to duct air from that isn't in effect drawing on the same air for intake.
Once the clock has gone past 0900 I will post the last 24 hours data. I suspect it may be interesting, but it won't be pretty!
Midea 14kW (for now...) ASHP heating both building and DHW
IMO defrosts are the Achilles heel of current ASHPs...I don't think there is anything unusual about my ASHP of set up; they all do it.
I absolutely agree. It's like having an internal combustion engine that throttles itself when you put your foot down - madness! The heat pump becomes a heat thief! I can see why icing happens, but what I can't understand is why such a self defeating solution to the problem has been used, with no apparent effort applied to finding a better solution. Perhaps they need to get ChatGPT on the case. A different material, or maybe a coating, for the coils, that resists icing? How do ships and planes deal with icing these days? An alternative way of generating warm air to do the deicing that doesn't suck heat from the house and at the same time pushes the heat pump to operate at it's most inefficient - maybe a separate heater element, which although it has a COP of 1, is at least not a heat thief?
Midea 14kW (for now...) ASHP heating both building and DHW
While you could defrost without using energy, it’s difficult. You could vibrate the coils, and get ice to fall off intact. I’ve seen an ashp designed to be manually hit with a hammer to get the ice off! Don’t do this with a conventional one, it will break it. Otherwise, I think the efficiency loss is less if you use heatpump created thermal heat rather than a straight electric heater, and slightly less the faster it is - if you did it too slowly, it might use heat but never actually defrost. What’s usually happening is that the ice is being melted off - which takes a fixed energy.
