@jamespa Originally, our Daikin was controlled by the WC settings but the radiators all underwent thorough scrutiny with the lockshield valve key first! There was still a certain amount of seasonal tweaking involved I felt.

We then had Homely smart control fitted and left all the monitoring and adjustments to this controller to carry out without me having to micro-manage a single aspect; even extremes of weather are well within such automatic control now.

No system is perfect and there are certain refinements I would like to see in the Homely app., but, on the whole Homely is better than even a retired 77 year old micro-managing nerd with a penchant for a constant 22.5 degrees C. 24/7.😉 Regards, Toodles.

@jamespa I refer to the Midea controller thermostat, not an external one. This said, I do not pretend to know what heat pump designers design, but the basic principle of feedback control, is to compensate for external disturbances, i.e. outside temperature and heat losses. In this sense, I cannot understand what a weather curve can do, if not messing things up.

Also, I don't understand what you mean by "adaptive". I am afraid people are using technical terms with different meanings, adding to the difficulty of communication. In anyway, I have not spotted this setting in my Midea controller. 

@tasos - if you have a Midea heat pump, as I do, I think you will find it can only run in two ways, fixed flow temp (LWT, leaving water temp) with a room stat (which in my case is in the airing cupboard with the DHW tank) doing on/off control rather than adjusting the flow temp (which by definition is fixed in this mode), or weather compensation. I have tried both, and so far as comparisons can be made allowing for different outside air temps, the fixed flow method is more expensive, by quite a lot.

Adaptive controls normally adapt to something. Homely is a sophisticated example, and it needs to be directly interfaced to the heat pump over modbus, which means it can (I don't actually know if it does) change the flow temp to adapt to, for example, the weather, or the building's heat loss profile, which it 'learns' over time. I believe Mitsubishi Ecodans have something similar, though less sophisticated. You can also set up DIY auto-adaption, which I have done using a python script that adjusts the weather curve over modbus when the indoor air temp is significantly lower or higher than it should be.

I do agree with you that the indoor air temp is the ultimate summative variable of the effects of all the other variables, and is ultimately the one that matters most.

PS I should add an omission from my first paragraph - the room stat can also be external to the wired controller, but it is still an on/off stat. This is what I have, and I use it as the on/off master switch eg to turn the heating off for the summer/on for the winter, by setting the desired temp to be either very low or very high. Using the Midea room stat, which is in the wired controller which is in the airing cupboard makes no sense at all!      

@cathoderay OK. It is difficult to argue through messaging. I ask you to bear with me. There are 3 ways to control a Midea heat pump, but in the 3rd mode you have to place the controller in a suitable position, so that the temperature reading is typical of the house temperature.

Homely "adapts" the weather curve, but I claim this is not necessary when feedback is enabled. Feedback takes care of "adaptation" on its own. Either there is something companies do not understand or what we teach at universities is wrong. Perhaps, someone with hands on experience on heat pump design can help.

Posted by: @tasos

↑

@jamespa I refer to the Midea controller thermostat, not an external one. This said, I do not pretend to know what heat pump designers design, but the basic principle of feedback control, is to compensate for external disturbances, i.e. outside temperature and heat losses. In this sense, I cannot understand what a weather curve can do, if not messing things up.

Also, I don't understand what you mean by "adaptive". I am afraid people are using technical terms with different meanings, adding to the difficulty of communication. In anyway, I have not spotted this setting in my Midea controller. 

Adaptive, in this case, means a mode which adapts its behaviour, over a period of time (days), to the house (basically it learns about the house characteristics).  Midea (and many others) does not have an adaptive mode.  Mitsubishi, Vaillant and others do, and Homely and the like are essentially adaptive add-ons to heat pumps, albeit they have a further level of sophistication.  All of them tweak the weather compensation curve, they do not substitute for the basic principle of running at a variable flow temperature determined largely by the outside temperature.

I totally understand the principle of feedback control.  However there is more than one way to do it, and the objective is not simply getting the right temperature at the point of measurement, its so far as possible getting the right temperature throughout the living space and doing so at the lowest cost.  That affects how you operate the feedback.

Obviously in any feedback system you must have something you can change to affect the variable you wish to keep constant.  In the case of a wet heating system you have two principal variables which you can use to adjust the output of the system to match the heat required namely, the flow temperature (temperature of the water leaving the heater), or the percentage of time for which the water flows (how much its on). 

The fact is that heat pumps operate more efficiently at lower flow temperatures because of the basic principle on which they work (If you want to look it up its linked to the Carnot equation).  The efficiency gain is roughly 2-3% per degree C, which means that if the flow temperature is lower by 1C they will deliver the same energy to the house (which is what you need) for about 2-3% less electricity in (which is what you pay for).  Thus it makes economic sense to operate heat pumps at the lowest temperature possible, consistent with keeping the house sufficiently warm.  Boilers are also more efficient (to an extent) at lower temperatures, but the mechanisms for the efficiency gain are different.

WC adjusts the flow temperature up (or more importantly down) so that the heat emitted by the emitters matches the heat being lost from the house (which is dependent on outdoor temperature, assuming that the objective is to keep the house at a stable indoor temperature).  Without WC you are forced to run the heat pump at (as a minimum) the temperature that is needed to supply sufficient heat to keep the house warm when its at its coldest possible outside, and then vary the percentage of time it is on to achieve the necessary energy balance.  So in the South of England for example, this means that you will be running the heat pump at a flow temperature suitable for -2C (the design extreme) even when its +7 ( the average and fairly typical temperature), which means you will be running it at roughly 10C higher than you need to most of the time, costing you 20-30% extra.  Most would agree that this is not desirable.  Furthermore at 7C the heating will be turning on and off a lot, which means you will be less comfortable (because there will be temperature variations) and, because the emitters are hotter than they need to be there will also be greater temperature gradients across the room, also reducing comfort.

So in an ideal world matching the heat input to the house to the heat lost from the house (which is dependent on outside temperature) will result in a perfectly stable inside temperature, and averaged over time that must be the case because of the first law of thermodynamics.  So operating WC (which is indeed 'open loop') will result is a stable house temperature with maximum comfort at lowest possible cost and with minimum possible temperature variations in time or space.  Many people find that this is in fact the case even in the non ideal world, because houses have a large 'thermal mass' which means that they do in fact average over time!

However we dont live in an ideal world and there are also dynamic effects, most notably solar gain and wind.  Here is where an input from the house temperature can come in if necessary, to compensate for this.  If solar gain is the main disturbance, as is often the case with modern houses (or older but modernised houses) then a good technique is to use a thermostat as a limit device, ie to set it above the desired temperature by a couple of degrees, so it 'kicks in' in extreme situations.  Another technique, if your pump supports it, is WC + temperature sensor.  In this variant the WC curve is set a degree or so above the optimum (so, if left to itself, will result in very mild overheating), and the temperature sensor switches the pump on and off to achieve the desired temperature.  However because the WC curve is very close to the optimum the on/off cycles are long and the efficiency penalty is relatively small (but still 2-3% per degree C).  This technique of course changes both variables we are able to control, flow temperature and % on time, but retains the concept that the FT does the heavy lifting (in achieving the required energy balance) and the % on time is a fine tune on top.

I hope that makes things clearer, if not please ask again.  You are of course free to operate at a fixed flow temperature if you choose (although an MCS installer is not free to set it up this way), however, unless you are operating already at a very low flow temperature (say 30C or lower), it will cost you money and reduce your comfort relative to either pure WC or  WC 'tweaked' as per one of the above (or another option if your heat pump offers it).

@cathoderay Just to clear up a question; yes, Homely takes over from Weather Compensation and handles the Leaving Water temperature dynamically. Regards, Toodles.

Posted by: @toodles

↑

@cathoderay Just to clear up a question; yes, Homely takes over from Weather Compensation and handles the Leaving Water temperature dynamically. Regards, Toodles.

... which is precisely how it achieves cost savings, by playing with LWT in response to OAT, predicted OAT, quite possibly predicted solar gain and wind, IAT and current energy price and optimising by reducing the LWT wherever possible, unless it makes sense temporarily to raise it to take advantage of cheap energy. 

@jamespa Every house needs a specific heat power to bring it to a set temperature. This means a specific water temperature. If you lower this temperature to be more efficient at the pump side, you will simply not attain the set temperature.

Further, the water temperature is controlled by the inverter. A PI controller, will thus control the inverter frequency, so that the house is kept at its desired temperature, irrespective of external temperatures.

If you want to be more efficient, just lower the desired indoor temperature.

On last point: you cannot have the same temperature in all the different places in a typical house with a single control. You must have TRVs installed ( i refer to radiator systems of course, because fan coils function differently).

Posted by: @tasos

↑

I ask you to bear with me.

I will. We tend to have robust discussions here, but they very rarely aggressive or hostile. Most of us are primarily interested in getting to the right conclusion.

We should perhaps also check whether we are talking about the same Midea unit. Mine is a 14kW M-Thermal Monoblock (model number MHC-V14W/D2N8-B). 

Posted by: @tasos

↑

There are 3 ways to control a Midea heat pump, but in the 3rd mode you have to place the controller in a suitable position, so that the temperature reading is typical of the house temperature.

I agree (and my edit - sorry, I should have got it right first time - I hope made that clear), and they are:

(a) weather compensation: flow temp is set by the outside air temp (OAT). In the Midea case, this is complicated by the fact the OAT is in the heat pump, and is therefore affected by the heat pump's running. As a further refinement, third party controls (Homely, my DIY auto adaption script which adjusts the Weather curve according to any significant differences between the actual and desired room temp) can be added, but the primary control is still by varying the flow temp

(b) fixed flow temp, with on/off by the Midea room stat in the wired controller. This will run the system at a fixed flow temp, and control the amount of heat delivered solely by turning the unit on and off. As my wired controller is in the airing cupboard with the DHW tank, it is not representative of the room temps in the main rooms, and so this is a no-no

(c) fixed flow temp, with control by an external room stat. This too will run the system at a fixed flow temp, and control the amount of heat delivered solely by turning the unit on and off.

I do have an external room stat, but as noted, it is not used for day to day control, it is used as the main heating on/off switch, by setting it either very high (so heating is always on eg 26 degrees) or low (eg 12 degrees, for frost protection). Using a room stat as the main space heating on/off controls means the DHW heating is unaffected, and for examples continues to run during the summer.

But the key thing is both (b) and (b) are fixed flow temp modes (apart from when they are getting up to temp, or cooling down because the room stat says off). The settings for whether you use fixed flow temp or weather compensation are in the second user (not FOR SERVICEMAN) menu (PRESET TEMPERATURE), where you can select either PRESET TEMP. (ie fixed flow temp) or WEATHER TEMP. SET (ie weather compensation) You can either select a standard Midea curve, of 'curve 9' which is the custom curve which allows you to set your custom weather compensation curve. The menu looks like this:

You do have to go into the SERVICEMAN menu to set the endpoints for the custom curve (curve 9).

I may of course have misread something in the not easy to understand Midea manuals, but the above does represent my best current understanding:

PRESET TEMP = fixed flow temp, modulation by on/off stat, whether internal to the wired controller Midea one or an external one

WEATHER TEMP. SET = flow temp varied by weather compensation; this is the only way of accessing the weather curves, including curve 9, the custom curve.

PS other have posted while I was writing this: @toodles - thanks, good to know; what I have written above is a Midea specific addition to @jamespa's comments      

@jamespa You don't need to predict anything. A PI controller effectively does all these.

@cathoderay I don't think (b) is fixed flow temp. Where did you see this ? If this is the case, it destroys all the benefits of an inverter driven pump.

Posted by: @tasos

↑

@jamespa You don't need to predict anything. A PI controller effectively does all these.

Maybe, or maybe (depending on the system) you need a differential component also, which is why PID controllers are a thing.  And maybe you want to take into account other variables to achieve optimum control (obviously optimum is dependent in part on the 'customer') 

However that is not the point, the point is that if you wish to operate just about any heat pump in fixed temperature mode using a (P)(I)(D) controller based solely on indoor temperature you are entirely free to do so (but note that an MCS installer is not free to set it up this way unless there is good reason).  In almost all cases however, this will cost you a significant amount more and you will be less comfortable for the reasons set out above.  Therefore it would be poor advice to suggest to anyone else that this is what they should do unless there are very, very, specific circumstances (which are not very common).