@cathoderay

You are indeed correct that, according to their own published data, Midea ASHP's are particularly poor at maintaining their specified output capacity at ambient air temperatures below +5C.

I have attached a copy of a spreadsheet that I produced some time ago for the Midea 14kW unit. The green highlighted cells on the top row can be varied to see the effect of making changes to your system. I have put in the known data for your system, the only parameter of which I am not aware is the 'Total Heating Elements Output'.

It should be remembered that the calculated LWT shown in column D would be better identified as the average heating element (radiator) temperature, so if there is a delta T of 5C between LWT and RWT, then the value shown would need to be 2.5C higher.

The other thing to remember is that heat loss is not just dependent upon the ambient air temperature, but also the weather conditions. The heat loss on a sunny day can be quite a large degree lower than on a windy and/or rainy day with the same ambient air readings.

@derek-m - thanks for posting this spreadsheet again, it is worthy of many postings! Total Heating Elements output (at delta t 50, ie fossil fuel delta t), is around 23,000W. It is just over half that at ASHP delta t values of around 30, around 13,500W.

Posted by: @derek-m

It should be remembered that the calculated LWT shown in column D would be better identified as the average heating element (radiator) temperature, so if there is a delta T of 5C between LWT and RWT, then the value shown would need to be 2.5C higher.

I meant to include this point, in relation to taking measurements as well as implications for emitted heat. Sometimes we think about LWT as it leaves the heat pump, sometimes we think about delta t at the rads, and fail to make the necessary corrections. A LWT 55/RWT 45 does not give 50 degrees at the rads, and so a delta t of 30 at the rads in a room with a design temp of 20, the rad average is going to be 40 something, and the delta t 20 something. That's where the over-sizing caused by rad selection factors (as above) might be enough to make up the shortfall.

Your spreadsheet gives something not very different to Midea's weather comp curve: your spreadsheet gives 53.5@-5/27.7@+15, Midea's standard curve is 55@-2/37@15. I think I am right in saying the higher 37@15 figure comes from GH's pre-emptive management of moaning biddies (complaining the rads were cool to the touch). 

The reason why my unit fails at lower temps is in the CAP columns for higher LWTs amd lower ambients, the upper right quadrant of the table. At LWTs at 50 and above, and ambients at zero and below, the CAP (output, in watts) falls below 12,000 - not enough to meet the 12.4kW demand at -2 ambient. In actual practice, the point at which the unit output falls below demand appears to be at slightly higher ambients, maybe around 4 degrees. That may be losses in the system either generally or more particularly because of the plate heat exchanger.

@cathoderay Ah yes, I remember now... Have you changed your unit into a 16kW yet, after all it's only dip switches 😀

Posted by: @hughf

Have you changed your unit into a 16kW yet, after all it's only dip switches 😀

Indeed it is, but it is currently summer, so no need for 16kW, plus I am wary of voiding my warranty because I have tinkered with the settings, hence my wish to clear it with my installer and more importantly Freedom, who are responsible for the parts side of the warranty, before I do anything.

Freedom have this no need to deal with pesky punters policy, which means I will have to go through my installer anyway. 

@cathoderay

I have attached a copy of the spreadsheet for a 14kW Ecodan for comparison. Ignore the #DIV/0 calculations at lower ambient temps which are due to there being no data table for LWT above 55C.

You will note that the Ecodan can maintain the specified output down to -10C.

I have found in the past that some Chinese manufactured equipment uses components that are very close to the maximum tolerance rather than the +20% often used by others, I therefore try to avoid buying Chinese wherever possible, though that is becoming more difficult in itself. The problem is that most people want equipment at the lowest possible cost, and since the 'bean counters' took over the world, if they can save 1p by using a lower specification component, then that is what they will do.

Posted by: @derek-m

I have attached a copy of the spreadsheet for a 14kW Ecodan for comparison.

Very informative. Unfortunately I was tied to using a Midea unit, by virtue of it being baked into the planning and listed building consents. The Midea unit is lower profile, making it less visually intrusive.

Posted by: @derek-m

I therefore try to avoid buying Chinese wherever possible, though that is becoming more difficult in itself.

Very true. I needed some chain to be re-galvanised a few years back, turned out the only UK plant doing hot dip spin galvanising which is needed for chain was closed indefinitely, and so I had to buy new even though I had perfectly good chain apart from the galvanising. Guess where the chain came from... Instead of recycling perfectly usable chain, we import chain from the other side of the world... Unsurprisingly, the quality of the galvanising is nothing like the old British stuff.

I'm putting this here because it is in effect a continuation of this thread.

For the past week, I have been fine tuning my heating always on weather comp curve. The cooler ambient temps have made it a good time to do this. I have ended up with 53@ -2 / 30@15. This seemed to keep the house steady at design temps throughout the 24 hour period.

However, being always on, it meant a sizeable chunk of energy use happened overnight. Overnight Sunday to Monday, it was using a steady 2kWh per hour, and if we count night as the 8 hours between 2200 and 0600, then that is 16kWh during a time when I am usually tucked up in bed. 

Yesterday I set the timer to have a setback to 16 degrees between 2200 and 0400. This morning, I see that unlike the night before, I didn't use 16kWh over night, in fact I used nothing, definitely a useful saving which will add up over time, but the house is only recovering slowly this morning, kitchen is 17.5 degrees against a design temp of 19 degrees, despite having the heating back full on from 0400. Furthermore, between 0500 and 0600, it used 3kWh. But that is only 1kWh 'extra', against the 16kWh overnight without the setback.

My guess is that the setback means the 24 hour steady state weather comp curve can't quite cope with the extra demand need to bring the house up two or three degrees from the cooler overnight setback room temps. This suggests I need to up the weather comp curve - but that then means when the house is in steady state as design temps, the curve will actually be a little two high...

The problem is the weather curve only knows about outside ambient temp. It doesn't know whether the house is in steady state, and need a little less input, or whether it is in a warm up phase, and so needs a little more input. The ideal answer so far as I can see is two weather comp curves, one for most of the 24 hour period (the steady state curve) and one for the warm up period, say 0400 to 0700, the latter having a slightly higher LWT to give the boost needed.

But, unless I am mistaken, you can only have one weather comp curve. Either the steady state one, which isn't enough during the warm up period, or the booster curve, which will be too much during the steady state period.

I'm trying to think of a way of tricking the system into boosting output in the warm up period. I wonder if any one has any ideas?    

@cathoderay 

I have the same issue and I just switch off between 12 and 5am.  My house is back up to where it should be in 2-3 hours in the morning, which works for us.  It definitely uses less energy doing it this way.   It would be good if my ASHP allowed the curve offset to work on a timer but it doesn't. 

I can use Mitsubishi's auto adapt, which does look at room and outside temp.  It seems to bump the flow up very high when chasing a target room temp and uses a lot of energy.  I haven't yet measured this but will do.  I suspect a lot of the savings overnight will be used up getting the house up to temperature in the morning. 

Posted by: @cathoderay

↑

I'm trying to think of a way of tricking the system into boosting output in the warm up period. I wonder if any one has any ideas? 

Cheat? As well as switching modes, my home automation system (demoticz) can override the flow temperature on my ecodan. I could program that to apply a fixed temperature, or with more difficultly it could increase it from the controller's curve-calculated value by a blanket offset, follow a different weather-compensating curve, or whatever.

But I don't, although applying a fixed high flow temperature is an easy quick fix on the control panel if the house isn't getting warm quickly enough. I think I could do the same with the maker's MELCloud app. Either way, domoticz would wait until the house exceeds the scheduled temperature, then take back control.

What I normally do is have the controller's curve set to be slightly more aggressive than needed, which was mainly so that it held the house steady even when the outside temperature is flattering (such as a howling southeasterly wind, which the outside sensor is sheltered from, in the afternoon on a sunny day when the sun warms the brickwork behind the sensor), but it has a fringe benefit that the house warms up a bit more quickly from its overnight setback temperature. The main trick is that domoticz turns the heating from "night" to "day" earlier if it's colder indoors. I think it can be as early as 0330, to as late as 0630 (or not bother) but it's usually around 0445.

I'm kind of the opposite to you guys. We are currently running 24/7 with temp comp set to 45 @ -5C & 30 @ 10C (Lounge sits at about 19.5C at this), with a demand from 1 or more zones all the times, some rooms are set to 18.5 as we dont need them so hot due to still working on them. As we have E7 I'd be happy with it running harder at night getting the temps higher in the hope we might use less on the daytime on the above settings. To this point E7 is working out cheaper (for all electricity use) than a flat rate but that may flip the other way eventually. Havent seen higher than 37C water temp so far, but this next week we certainly will.

Posted by: @kev-m

↑

@cathoderay 

I have the same issue and I just switch off between 12 and 5am.  My house is back up to where it should be in 2-3 hours in the morning, which works for us.  It definitely uses less energy doing it this way.   It would be good if my ASHP allowed the curve offset to work on a timer but it doesn't. 

I can use Mitsubishi's auto adapt, which does look at room and outside temp.  It seems to bump the flow up very high when chasing a target room temp and uses a lot of energy.  I haven't yet measured this but will do.  I suspect a lot of the savings overnight will be used up getting the house up to temperature in the morning. 

Here is one method you could try if your thermostats have a schedule capability.

Rather than trying to program a change in WC curve or WC offset, program your thermostat to a lower temperature setting. By doing so your heat pump should switch off during the specified period during the night and allow the indoor air temperature to fall, but if it is particularly cold then it will switch the heat pump back on to maintain this lower temperature without it falling any further. Most nights it will probably not switch the heat pump back on. During the day the thermostat should resume it normal higher setting, thereby allowing normal WC control to be performed.

It may also be possible to bring the indoor air temperature back up slowly, by raising the setting on the thermostat in 1C increments.

I realise that it goes against the 'preaching' of running the heat pump constantly on WC, but if it reduces energy consumption and you can live with the temperature variations, then by all means it should be tried.

@kev-m, @mjr, @mookyfoo, @derek-m - thank you all for your comments. I did do the setback over Sunday to Monday night (to 16 degrees on the kitchen room stat which is just a calling/not calling for heat control between 2200 and 0400). It worked insofar as the heat pump was off for that period, and I probably saved around 12kWh of electricity, and it also tells us, because the heat pump didn't come on, that the kitchen stayed at > 16 degrees, so not really a major drop in warmth, but on the Monday (yesterday) morning it took forever to get back up to temperature, and if fact never got to design temp all day. By noon it had reached 18.0 degrees, and pretty much stayed there all day. Design temp is 19.0 degrees. 

Because of this failure to recover, and the forecast colder weather, I turned off the setback last night and had the heat pump running all night. This morning, the kitchen is still only at 18.0 degrees.

I think this failure is because of a combination of things: (a) the heat pump is currently in this cold weather operating at the limits of its capabilities (the old 12.4 kW loss vs a 11.4 kW output from my so called 14kW heat pump) and (b) the relatively high thermal mass of my building. It cools down slowly (didn't even get down to 16 degrees overnight Sunday/Monday), but it also takes forever to heat up again. Basically, when its cold outside, say 5 degrees and below, the heat pump can just about supply enough heat when the house is in steady state to keep it in steady state, but when it has to raise the temp from a setback, it really struggles. I think during this cold spell I am going to have to run it 24/7.

If its a bit milder, I might try a setback again. Previously i asked how to boost the heat output during the rewarming phase (the have a second weather comp curve for this phase question) but I now wonder whether that might possibly be answered/worked around the other way round, ie have the curve set to what is needed for the boost phase, and then achieve marginally lower output in the steady state phase by the use of TRVs in a number of less frequented rooms. These would start to close down a bit more than they do now as the rooms got to temp, and the reduced flow would in due course cause the heat pump to marginally reduce its output, as the warmer RWT from the plate heat exchanger 'fed back' the lower heat demand to the heat pump.

But that might be nonsense, because all the heat pump ever knows about is ambient temp. It has no way of knowing what the room temps are...