Here's a four hour period from the same morning last November. Why does the set LWT change when it does? The OAT is lower when the heat pump is active (actual LWT is rising), is this simply the exhaust air cooling the local OAT a bit? What is the real OAT? That which appears when the heat pump is not active?
That does look like the OAT is dropping slightly due to the cooling effect from exhaust air. I see similar on my Samsung, who's OAT thermocouple is shielded in a black plastic sleeve and clipped in location on the rear of the heat pump in front of the fins, so measuring intake air temp passing over the fins. Mine samples regularly (every 30 secs), and I can observe a 1-2C drop when the heat pump starts up which I assume is the cooling effect of recirculating air.
The change in set LWT appears to correlate with OAT, decreasing as the OAT increases and increasing as the OAT decreases, between 10-13C, as you'd expect from a system running weather compensation. Again, this is most likely driven by changes in OAT caused by recirculating cold air than any changes in actual weather conditions. And the system appears to be cycling every ~30mins.
Samsung 12kW gen6 ASHP with 50L volumiser and all new large radiators. 7.2kWp solar (south facing), Tesla PW3 (13.5kW)
Solar generation completely offsets ASHP usage annually. We no longer burn ~1600L of kerosene annually.
Homely has been discussed a fair amount on the forum, and it seems to a bit of a Marmite product. The developers also have some rather odd business practises, which are of course their choices to make, but at the same time they are not popular with some end users, eg making it difficult to get at the raw data, which after all does belong to the end user. For all the marketing hype, Homely is just a load adaption algorithm, with a bit of added look ahead at the weather forecast and time of use tariff tweaking, hooked up to the heat pump over modbus. I'm inclined to think their savings claims may be exaggerated. These things are extremely hard if not impossible to measure accurately in the real world, because conditions are constantly changing.
The 'can save £444/year' report is as you suggest, though don't quite go as far as to say, disingenuous, because the non-Homely setups used higher than normal set LWTs, either fixed at 55°C, or on a WCC with the left hand end (higher) LWT set to 55°C. This will, all other things being equal, increase running costs over a more normal setup using a lower max LWT. Furthermore, these is no data on the actual LWTs that Homely used (though there is an unexplained table that shows LWts ranging between 35.2 and 43.8°C, if these were the LWTs Homely used then of course it was cheaper to run), or the actual IAT (room temperatures) achieved with and without Homely (if one was on average cooler than the other then of course it would be cheaper to run). Homely say the tests 'ran for over 3 months testing different heating scenarios in both UK winter and shoulder season conditions' but there is no information on the conditions during the various tests eg if Homely was on average used in warmer weather then of course there will be savings. Nor is there anything on the Energy House Labs website about this study. Given the paucity of background and detailed information, I have to conclude that in my opinion that this is a marketing exercise based on Homely marking its own homework.
Midea 14kW (for now...) ASHP heating both building and DHW
As you say in your earlier post OAT is fundamental because it drives everything else so at least we need something consistent. I doubt it materially prejudices the day to day operation putting the sensor inside the unit, so I can understand why some manufacturers do albeit that it makes understanding whats going on a right pain.
Have you got a plot similar to the one above when its a lot colder outside?
4kW peak of solar PV since 2011; EV and a 1930s house which has been partially renovated to improve its efficiency. 7kW Vaillant heat pump.
That does look like the OAT is dropping slightly due to the cooling effect from exhaust air.
That has always been my most likely explanation, in which case the Midea recorded OAT when the heat pump is running is a bit lower than the actual OAT the building sits in.
The change in set LWT appears to correlate with OAT
I'm not so sure about this. The drop in set LWT 0440 does happen after a rise in OAT, but a similar rise happened ~30 mins earlier, and didn't trigger a change in set LWT. Likewise, the fall at 0740 happens after a rise in OAT, but that rise had been preceded by two earlier almost identical rises that didn't trigger a fall in set LWT. And what triggered the rise in set LWT at 0510? It happens on the second or even third of three very similar OAT variations.
The WCC settings at the time were 55 @ -4 / 30 @ 15 which is roughly 1.3 degree change in set LWT for every degree change in OAT, but the changes in OAT are not followed as one might expect by corresponding changes in the set LWT, which leads me to wonder what the control logic is. That said, a set LWT or around 35° is about right given the OAT and WCC settings.
And the system appears to be cycling every ~30mins.
Indeed it is, Midea's do this, they only run steady at around 7-8° OAT. But it is slow cycling, and I guess Midea and Midea clone owners just have to accept that's the way these heat pumps work.
Midea 14kW (for now...) ASHP heating both building and DHW
Have you got a plot similar to the one above when its a lot colder outside?
Yes, here is a zoomed in plot for a cold night in January this year:
Note that the cycles in this plot are defrost cycles, when the heat pump exhaust air is presumably warmer, and so we may be seeing two things: exhaust air cooling the local OAT when it is running normally, but warming it when the heat pump is defrosting. Another thing I have noticed is that in milder weather, the actual LWT varies above and below the set LWT, with the average being close to the set LWT, whereas in colder weather, the actual LWT is at or below the set LWT, giving an average that is below the set LWT.
Midea 14kW (for now...) ASHP heating both building and DHW
The refresh time of climate-related curves in heating mode can be changed but not less than 0.5 hrs.
This explains the LWT behavior. Further, since the OAT sensor is influenced by defrosting, it effects the WC and hence the whole operation. I don't think this is desirable.
I would either use a reliable external sensor and feed it in or discard it altogether and "tweak" the LWT setpoint according to IAT.
This post was modified 1 month ago 3 times by Tasos
Midea MHCV10WD2N7 R290, 4.8kW peak energy community solar power.
The refresh time of climate-related curves in heating mode can be changed but not less than 0.5 hrs.
This explains the LWT behavior.
Thanks, that might explain the when, and possibly the why. My 't_T4_FRESH_H' (the relevant setting, though it is not exactly clear, under Heat Mode settings in the For Serviceman menu) has been set to 0.5 hours and here is my most recent chart with some 30 and 60 minute blue bars overlaid:
Some of the decreases in set LWT do coincide with OAT spikes, and the increases in set LWT may happen at the next refresh after the the OAT drops again. But it is not entirely consistent.
Further, since the OAT sensor is influenced by defrosting, it effects the WC and hence the whole operation. I don't think this is desirable.
I would either use a reliable external sensor and feed it in or discard it altogether and "tweak" the LWT setpoint according to IAT.
For sure, the heat pump affects the OAT, and so the set LWT via the WCC, in all sorts of ways, more often than not in unwanted ways. Just look at them thar spikes! Normal cycling also has an effect, though less striking. In effect, as @jamespa observed some time ago, it is actually an air intake temperature (AIT) sensor. It is certainly not perfect, but ditching it and using an external sensor is no simple undertaking (if it can be done at all - it might be possible by using a python script to set the WCC end points and thereby the set LWT, but it would be very complicated), and so we have to live with it. Likewise using an IAT sensor, plus doing that loses the 'anticipatory' advantage that using the OAT/AIT has.
Midea 14kW (for now...) ASHP heating both building and DHW
Hi all- does anyone know if "T1S" is midea language for the target outlet temperature? I have Climate Curve T1S Calculated Value 1 as what I believe to be representative of what the target outlet temp will be, when the climate curve is on.
What I have noticed is that if the curve set temp is low 30s it will overshoot to around 35/36, which I presume is down to the unit itself not being able to modulate down low enough to achieve this outlet temp? Our temperature drop (dT?) from outlet to inlet is consistent at around 5-6 degrees C. Yesterday I in fact finished plumbing in our new kitchen radiator, turned off climate curve and just set at 45 degrees C to fully test all of the connections for leaks, and this maintained a steady outlet temp of 45/46 so I'm guessing as above that low 30s, whilst this is the set temp, might not be achievable perhaps as said with a 12kW?
@benson T1S is actually the LWT defined by the WC. However, I wouldn't describe the behavior as "overshooting", but rather "oscillating" around the set value. This is obviously very bad set-point tracking.
I am curious as to what the corresponding compressor speed is, as this looks very bad control to me.
Furthermore, in your previous chart,
I notice a similar erratic behavior when the OAT spikes up, the LWT spikes down by a much larger percentage. This, in control theory is termed "kick", resulting from a change in the set point (in this case T1S because of change in OAT). However, in theory this can be remedied, so that the LWT spike is not that large.
@tasos - never one to knowingly miss an opportunity to say weather compensation isn't any good!
@benson - my understanding is this: T1 is the actual LWT (leaving water temperature), T1S is the set LWT, ie what the WCC determines the target LWT to be. That comes from the Engineering Data manual, in the wired controller modbus register table, T1 is at address 110 and T1S is at address 136 (and called 'Hydraulic module curve T1S calculated value 1', 1 being zone 1. The register also has addresses 104, Water Inlet Temperature, TW_in, and 105, Water Outlet temperature, TW_out, which I think, based on observations, are the same as TW_I Plate Water-Inlet Temp and TW_O Plate Water-Outlet Temp on page 4 of the Operational Parameter pages on the wired controller display. As I never knowingly miss an opportunity to say Midea are hell bent on making sure we live in interesting times, I merely observe that I need to use a whole paragraph of five or so lines to describe Midea's terminology for two very basic values, the set LWT and the RWT.
Apart from around zero and a bit degrees OAT, when they use defrost cycles, and the sweet spot of around 5-7 degrees (it varies) when they achieve steady state running, Midea and clone heat pumps spend most of their time doing low frequency 'normal' cycling, and to achieve a mean actual LWT close to the set LWT, they over and undershoot. If you eyeball the mean actual LWT in @benson's chart, that mean is pretty close to the set LWT.
I do record but don't usually chart the compressor speed (too much clutter). However I do have historical charts that show it (and flow rate) including this one from January this year, which shows low frequency 'normal' cycling on the left, something approaching steady state running in the middle, and defrost cycling on the right. The flow rate also shows another Midea characteristic, much of the time the flow rate is fixed, at around 1.3 to 1.4 cubic metres per hour (note it is charted as actual value x 10, to improve visibility):
Perhaps also worth pointing out that when using defrost cycles (right hand side), the actual LWT mostly only reaches the set LWT, ie the mean actual LWT is below the set LWT, whereas in 'normal' cycling (left hand side), the actual LWT varies either side of the set LWT, to give a mean close to the set LWT.
Midea 14kW (for now...) ASHP heating both building and DHW
I notice a similar erratic behavior when the OAT spikes up, the LWT spikes down by a much larger percentage.
I think you are mixing up the chicken and the egg here. Those are defrost cycles in that chart, ie it is operating in what I call heat thief mode, stealing heat from the house to effect the defrost. As a result, the LWT (and RWT, but less so; it is the LWT going below the RWT that marks a defrost cycle) both plummet, and because the exhaust air is now warmer, the OAT or more accurately AIT (air intake temperature) increases. The order of play is increased exhaust air temp > increased OAT/AIT, not the other way round.
Midea 14kW (for now...) ASHP heating both building and DHW
