@jamespa  So, it seems that we have an empirical answer to your question rather than an engineering one. Therefore, from this input and context, I deduce that the optimal scenario is a heat pump with variable pump flow control maintaining a dT of 5C, but if you have a basic modulating heat pump with fixed pump flow, then one should plan for minimum 8C dT at design low OAT temperature such that the dT can be limited to ~3C at 15C OAT with fixed flow. To achieve this, a 10kW heat pump would have to modulate the compressor down to 3.75kW to achieve 3C dT, which would seem to be within the scope of most modern pumps. I presume that this is why Ecodan have temperature sensors on both flow and return pipes. Is this a reasonable conclusion? My preference would be for a clear engineering explanation of heat pump behaviour which would provide for an informed basis for planning.

@jamespa  For the sake of clarity, in my previous response I ignored the house heat requirement at 15C OAT for a 10kW load at design temperature. The heat requirement at 15C OAT would be about ¼ of that at 0C OAT, so a 10kW heat pump would have to modulate down to 2.5kW which in this same scenario would result in a dT of 2C for a fixed flow. For most HP’s, cycling would probably set in before this, I suppose.

Posted by: @heacol

↑

That 2.5% cost increase every degC corresponds to my home-made graph pretty accurately.  I was confused by the graph above at first, but then I understood it; a 25C increase in flow temperature would double your running costs (ie. +100% on top of costs).

Great rule of thumb, like it

Posted by: @heatgeek

↑

Is this a reasonable conclusion? My preference would be for a clear engineering explanation of heat pump behaviour which would provide for an informed basis for planning.

I honestly dont know.  I think there were two questions

1. Is 5C somehow critical  - answer no

2. Does it matter if you have designed for ~5C at the design conditions, your heat pump doesn't modulate the water pump, and so higher OATs than design deltaT will be less - answer not sure

In the absence of a coherent explanation why its not best to allow the deltaT to reduce to whatever lower figure it settles at in scenario 2 & higher OATs, I would fall back on the thermodynamic argument which says that it is better to let it reduce on the grounds of lower DT = lower FT required for a given average rad temp (and thus rad output) = better COP.

I think that means we dont yet have a conclusion because we dont have a plausible engineering argument to counter the thermodynamic one, and we do have the observation posted earlier that the best performing pumps on openenergymonitor all operate at <5C.

@jamespa, please do not make assumptions when you admit that you do not understand the topic. There are reasons for the 5 Deg Delta T but they are exceptionallyy complex. I do not fully understand them yet and I have been working on it for years. It has to do with the refrigerant circuit.

My 7kW Arotherm+ normally runs at 1200 litres/hour but I've seen it show 1800 litres/hour when doing the DHW. It may also use the latter speed when getting near the maximum heat output on the CH. That's a rare situation and I haven't checked the flow rate when it occurs. 5C delta T at the heat pump is compatible with about 4C at the indoors end of the primary circuit.

Posted by: @heacol

↑

@jamespa, please do not make assumptions when you admit that you do not understand the topic. There are reasons for the 5 Deg Delta T but they are exceptionally complex. I do not fully understand them yet and I have been working on it for years. It has to do with the refrigerant circuit.

I am not sure what assumptions you think I am making in my analysis of what we know so far about whether there is anything 'magic' about a 5C DT flow-return, and in particular whether there is an advantage in modulating the water pump (which some heat pumps seem to do and others not) to maintain this DT even at higher OATs

Reviewing the responses above we have

• @heacol arguing that 8C is ideal ( https://renewableheatinghub.co.uk/forums/postid/43261) and but clarifying that anything in the range of 3-10 makes little difference
• @robs saying that 'not a single one of the top performing heat pumps on heatpumpmonitor.org have a constant dT of 8C. At warmer OATs they mostly run with a dT between 2 and 4.  ( https://renewableheatinghub.co.uk/forums/postid/43266)
• @robl who has designed his own saying it operates at 6C, he believes it would be more efficient at 3C but that would be audible so he chooses not to ( https://renewableheatinghub.co.uk/forums/postid/43269)
• @jamespa pointing out that the simple thermodynamic argument says keep it as low as possible on the grounds that lower DT = lower FT for any given average emitter temp = higher COP ( https://renewableheatinghub.co.uk/forums/postid/43292), and also that a linear WC curve, which is what many heat pumps offer, is closer to ideal if the water pump is operated contstant flow rather than constant DT.

The only experimental evidence we appear to have is that quoted by @robs.  I haven't verified it but if he is correct then it would seem to run counter to the argument that its somehow necessary to maintain 5C by modulating the water pump.

The only argument we have so far had that is supported by an engineering explanation amounts to 'to first order lower DT is better (better COP) but given that low DT = fast water/fat pipes, ~5C is often a good practical compromise, sacrificing just a few percent in efficiency in return for pipework which is practical for the typical domestic environment'  

As I say above I think that means we dont yet have a conclusion, and I cant see that in coming to that conclusion I am making assumptions. 

I suppose I am however applying my general rule which is to apply a tinge of scepticism to anything I am told, unless it is supported by rational argument or by authoritative references.  For that I apologise if an apology is necessary, although in todays dis-information fuelled world I would argue it is a sensible approach.

@jamespa There is no consensus as manufacturers will not supply their test data, however, it is very much based on the design of the refrigerant circuit and where it is optimized to. There is not a single heat pump optimised to UK's temperate climate as the market is minute, most are optimized for operation below -10Deg C. All manufacturer data that I have read, reccomend between 5 and 10 Deg C.

I personally have not tested it but I can guanty, every manufactures unit will opperate diffferently so there will never be an industry wide consensous. I have found, the best perfomance to be between 5 and 8.

From an experience point of view I have been watching our fixed flow rate Ecodan operate for a few years now and have seen how versatile its modulation is. Also when making alterations it was handy having an external circulator. When we removed our originally fitted LLH we also changed the circulator to give us a larger head. Regarding noise the larger circulator wasn’t noisy at all. It had an hand dial vario speed which meant we could precisely set the flow rate we dialled it back to give us 16LPM and the display tells me flow rate and wattage at that dial setting. The 48w output is quiet py comparison to what we had originally, as we now have one pump instead of two the circulator kWh has been reduced overall.

the view that variable speed setups are more sophisticated might not recognise that it would be run by a fixed or less variable compressor. I guess it’s where you like your sophistication. In the circulator or in your compressor?

Mitsubichi has in only the last 15 months changed their data book to give a single optimum flow rate. This is a big change to my heat pump performance. It still has the full flow rate range quoted but the single optimum flow rate gives the HP the best scope of modulation. From 3 lpm up to 10 or 12 lpm.

here is a range of outputs and DTs showing on a minute by minute basis how DTs might be adjusted presumably based on feedback from thermistors or thermostats.

these charts also help with fault diagnosis which may be difficult to pick up on when variable speed circulation happens

anyway just some observations and thoughts with no real bias as I don’t have a variable speed setup

@jamespa @heacol 

Beware disinformation on this forum. I took some time out to verify the statement made about all the top performing heat pumps on heatpumpmonitor.org being fixed flow. I have taken the top 10 performing Vaillants, some Daikins and others and plotted the reported power levels at 4C, 12C & 21C Room delta-T’s plus the associated flow dT’s. From these, the associated water flow rates can be calculated. To boot, the theoretical power level was calculated based on the premise that dT would have been 5C. The data points are not exact and the nearest readings were taken so some variability must be accepted.

I have come to the following conclusions based on a cursory analysis:
i) Many or most of the systems are grossly over-sized.
ii) All of the systems seem to have variable flow control, not fixed flow. I presume therefore that the Vaillant Arotherm, Daikin, et al have internal PWM pumps, otherwise where does the variability come from.
iii) In most cases, the flow dT’s are very low (2-4C)
iv) The calculated power levels at 5C dT match the rating of the heat pump, except for some extreme cases.
Therefore, what seems to be happening here, in my view, is that the HP sets the flow according to its rated output and adjusts this flow in line with WC. However, if the pump is very over-sized, the HP has to squeeze the dT to reduce the heat delivery as full output cannot be used. This is borne out by the calculated power for 5C. I have not done any detailed analysis of this spreadsheet but I thought I would throw it out there for musing, amusing and comment. Please check for calc errors.
So, would performance be even better if the HP was better matched to the required heat load??

Thanks for this, good (and useful) analysis!

Posted by: @heatgeek

↑

I took some time out to verify the statement made about all the top performing heat pumps on heatpumpmonitor.org being fixed flow

As a matter of interest where did you see that statement, the one I saw (which appears to be borne out by your data) was

Posted by: @heatgeek

↑

iii) In most cases, the flow dT’s are very low (2-4C)

Posted by: @heatgeek

↑

I presume therefore that the Vaillant Arotherm,  Daikin, et al have internal PWM pumps,

I can confirm that the 7kW Vaillant has a variable water pump speed.  Unfortumnately I can only monitor it by looking at the UI I dont have any logging of this variable, so dont know what the algorithm is.

Posted by: @heatgeek

↑

However, if the pump is very over-sized, the HP has to squeeze the dT to reduce the heat delivery as full output cannot be used.

Can you explain what you mean by this exactly.  The minimum pump output wont change materially if DT is changed, and if the minimum pump output exceeds the demand from the house, the heat pump must cycle.   

Posted by: @heatgeek

↑

i) Many or most of the systems are grossly over-sized.

I am guessing you are comparing columns D and F rather than D and B.  D and B seem to match quite well, D& F less so but given that heat pumps come in discrete size increments (and quite often models with more than one badge power are in fact the same hardware so there is little point in going for the smaller one).  D & F arent that far out for the Vaillants, although they appear to be a long way out for the Daikins & Mitsi which look like they could be survey overestimation given the values in B

Posted by: @heatgeek

↑

So, would performance be even better if the HP was better matched to the required heat load??

Thats a whole other discussion and will trigger as many replies as the discussion about deltaT.  Basically should one undersize, critically size or oversize a heat pump.  Im not sure I personally want to engage in that discussion again, but doubtless others will

@heatgeek You state "Beware disinformation on this forum", then you state "The theoretical power level was calculated " and finally "Vaillant Arotherm has internal PWM pumps"

You have made some assumptions, therefore, you do not have a full data set to make a determination. Vaillant Arotherm heat pumps do have a variable speed pump, however, unlike the Daiken, it does not track the flow temperature maintaining a fixed Delta T, it sets a fixed flow based on the unit's size.

Without comparing like for like, with a predetermined control, any analysis will not produce accurate results. That is why I built a test rig with Ulster University and produced the document published on this website under articles. This was so that different installation models could be compared under controlled conditions using BS7671 test conditions. Open energy monitor is a resource for the installer only, if you study most of the top-performing installations they are highly tuned, I know one of the top ones has near-weekly tweaks, and they are used for marketing purposes. It is not the resource it makes out to be. I do have an installation on there and will have another at the end of the week, it performs with an average COP of approximately 4, it is one of my commercial installations, done several years ago, installed by a subcontractor. We only changed the radiators, it is in Glasgow, a solid stone building with some single glazing and 50% microbore. It is in an extremely exposed position with the prevailing winds off the Atlantic. It is not difficult to get a good level of performance with good equipment, the units listed here would not be my first choice.