Maximising the COP of an ASHP

[joe90]

I am sure this has been discussed before but can’t find it. Now that I am getting nearer completion my mind has come back to this question . As we all know the COP of an ASHP goes down if the unit defrosts, so as per Others I plan to limit the water temp to about 40 -42 degrees and top up with immersion’s. It occurs to me that the defrost is weather dependant and I was wondering if there is some way for the ASHP to keep raising the water temp till it gets near to defrosting, and then turn itself off, that way in good weather you may end up with 50 or 55 degrees and spend less on electricity for the immersion’s and only pay for top up when necessary?

Edited November 24, 2017 by joe90

[dpmiller]

Does your ASHP do weather comp?

[joe90]

25 minutes ago, dpmiller said:

Does your ASHP do weather comp?

No. Would this be comprehensive enough? , I just thought if the ASHP could switch off rather than go on to defrost itself the warmest water temp could be achieved ?.

[ragg987]

Provided the cop remains above 1 despite the defrost cycles, why would you use an immersion?

 

With external temperature of around 10C in October, and dhw target temperature of 50C I was seeing COP of between 4.2 and 4.5. This was before the hearing season started.

[Nickfromwales]

1 hour ago, joe90 said:

I am sure this has been discussed before but can’t find it. Now that I am getting nearer completion my mind has come back to this question . As we all know the COP of an ASHP goes down if the unit defrosts, so as per Others I plan to limit the water temp to about 40 -42 degrees and top up with immersion’s. It occurs to me that the defrost is weather dependant and I was wondering if there is some way for the ASHP to keep raising the water temp till it gets near to defrosting, and then turn itself off, that way in good weather you may end up with 50 or 55 degrees and spend less on electricity for the immersion’s and only pay for top up when necessary?

I think the unit would spend too much time idling, and 'stuck' in limbo trying to satisfy DHW without freezing and the space heating would get neglected for excessively long periods of time. I know you have a buffer so that may make it feasible if it stores enough energy to bridge these periods where the house would go 'unsatisfied'. 

CoP has to be maximised by design, from the outset, so you really should know what to expect before even plugging it in  

[Nickfromwales]

Just now, ragg987 said:

Provided the cop remains above 1 despite the defrost cycles, why would you use an immersion?

 

With external temperature of around 10C in October, and dhw target temperature of 50C I was seeing COP of between 4.2 and 4.5. This was before the hearing season started.

To boost the uvc . 

Thats a good figure for your CoP !

[SteamyTea]

That is an interesting problem @joe90

It is easy enough to make a sensor that can check temperature and RH, from that it is a simple calculation to work out the dewpoint temperature.

Once that it known, and a few performance parameters of you ASHP installation, it is then just a matter of turning down the wick on the ASHP.  I am not sure how easy that is to do, but probably not hard.

I think @JSHarris ASHP has sensors already built in that can be used for this.  SO other ASHPs probably have the same thing as it is pretty crucial to their performance (they need to know they are frosted up).

[ragg987]

2 hours ago, Nickfromwales said:

To boost the uvc . 

Thats a good figure for your CoP !

If you mean the weekly boost to prevent legionella then I would use the ASHP to the maximum possible, in our case 55C, then immersion after. Or PV of available.

 

Yes the cop is surprisingly good, I am guessing we did not have defrost cycles at that point. So far with space heating plus dhw on in November I am getting cop of 4.3 on average, despite external temperatures being closer to freezing.

[Jeremy Harris]

COP barely changes with temperature, as long as the air is dry.  What seems to kill the COP is humidity - damp air condenses out and ices up the evaporator.  Our Carrier ASHP (exactly the same unit is badged Kingspan and Glowworm, and probably a few other brands too) has a humidity and temperature sensor mounted right on the rear of the unit, that senses the air flowing in to the evaporator.  The temperature and humidity can both be displayed on the Command Unit. 

 

From what I've been able to work out, from reverse engineering and experiments, the unit doesn't directly detect when ice has formed on the evaporator, but instead determines when the risk of this is high, by a combination of comparing the evaporator temperature (which is measured by an internal sensor in the refrigerant loop), the outside air temperature and the relative humidity of the air flowing into the evaporator, the logic within the unit tries to predict when icing is likely and uses that to trigger a defrost cycle.

 

The defrost cycle has a greater impact on the true COP than might at first be thought, as I have a sensor half way up our buffer tank and when the ASHP goes to defrost mode it rapidly draws heat from the buffer over the ten minutes of so the cycle takes.  The way it works is that the 4 way valve operates and the ASHP goes into cooling mode, drawing heat from the house and using it to heat up what was the evaporator heat exchanger (but is now the condenser) in order to melt any ice.  So, when measuring the electricity drawn by the ASHP (I have a separate DIN rail mounted energy meter on the supply to ours) you also have to try and estimate the amount of heat energy that's been drawn from the house, and add that to the total defrost energy use, as clearly the ASHP has to work to replenish the heat it has pumped out of the house when it returns to normal operation.

 

I've mentioned this before, but the worst case conditions seem to be when the outside air temperature is cold and damp.  Cold, misty, weather seems to cause our unit to defrost if it's asked to deliver a high flow temperature, but cold dry weather (like a cold and frosty morning) won't make the unit defrost at all.  I ran ours at 50 deg C when the Command Unit display was showing a very low RH on a cold and frosty day and it didn't defrost, which suggests that there is some pseudo-intelligence built in, at least to these models.

[SteamyTea]

@JSHarrisHow hard would it be to reprogram the unit to modulate the heat pump down a bit when it predicts the risk of frosting?  Then it is just a matter of creating an algorithm (rather than a curve) that can control that for near enough all temperatures.

Or it could just turn the heat pump off and let ambient temperature defrost naturally, may need a larger buffer tank though.

 

Where @joe90 is not too bad for temperature and RH, so may not hit the frosting condition as often as you do.

Edited November 24, 2017 by SteamyTea

[ProDave]

I have often thought that "turn it off and wait" would be a much better defrost method, providing the air temperature is above 0, and you are not in a hurry to got your HW tank up to temperature.  Something to experiment with when it's all built and I suddenly have "spare time", something I don't recall having for a while.

[PeterW]

I always wondered if you could stick some of that trace heat tape on the condenser and as the temperature reached 0c then trigger it to come on for a short while. It would mean that you’d have to be careful in winter but the logic was it would self control. 

 

Or am I talking rubbish ..?! 

[pdf27]

8 hours ago, PeterW said:

I always wondered if you could stick some of that trace heat tape on the condenser and as the temperature reached 0c then trigger it to come on for a short while. It would mean that you’d have to be careful in winter but the logic was it would self control. 

 

Or am I talking rubbish ..?! 

Doesn’t that mean whenever the outside temp is below zero you’ll be on resistance heat and be warming up the outdoors?

[Jeremy Harris]

"Turn off and wait" would work well when the air temperature is well above zero, the house has a long thermal time constant and you have enough stored water to meet the hot water demand for the next hour or two.  The snag is it wouldn't work at all if the temperature was below zero, and would take a long time to work if the temperature was only just above zero.

 

I can't say for sure, but all the tests I've done on ours suggest that icing is most likely when the air temperature is just above zero and the conditions are damp, a combination that is common here, as after a warmer day, as the air cools during the evening it tends to become saturated with water vapour, which then looks to condense out on any cooler surface, hence the dew and frost we see.

 

I think positioning an ASHP so that the rear face (where the evaporator is) faces the sun, particularly later in the day, might help.  Anything that warms that area up, even by a small amount, would make a difference, and even in the very coldest weather solar radiation can be pretty powerful (a look at how the ice and frost on the side of a car facing the sun melts pretty quickly in the morning shows that).

 

I remember reading a thread ages ago, not sure whether it was on ebuild or somewhere else, where someone proposed fitting a ASHP inside a south facing conservatory, fitted with ventilation grilles to allow air in, sort of built in to the wall, with the exhaust facing outwards.  The idea was that it would be able to draw in slightly warmer air, even in winter, when the conservatory wasn't being used, which would make it less prone to icing.  Anything you can do to lower the humidity around the unit would help, too.  I keep meaning to add a small overhanging roof over ours, to both stop it getting wet in the rain (which must raise the local humidity a bit) and also to slightly reduce the radiative heat loss, by shielding the unit and the ground around it from the cold night sky.  The idea is that the water vapour in the air would condense out on colder surfaces nearby first, as the lose heat to the very cold night sky, so the air reaching the ASHP would already be just a little bitdrier.

Edited November 25, 2017 by JSHarris

[SteamyTea]

I don't think that it is really possible to affect the local humidity as an ASHP draws in too much mass of air.  Sticking one in the sun should help though as even at quite modest levels of direct sunlight will soon warm it up.  Just think how easy washing steams on a cold day when the sun hits it.

 

It may be useful, but would need experiments to find out, to blow air though the unit with the HP part turned off.  If the HP part was stopped before freezing starts (should be easy to establish) then just pass ambient temperature air though, any condensation should evaporate, or get dislodged, fairly quickly.

[Jeremy Harris]

Ours certainly runs air through the evaporator when the compressor is turned off, albeit fairly slowly, with the fan speed modulated right down.  Presumably that's to try to keep the evaporator as close to the ambient air temperature as possible, which would reduce the risk of icing for as long as the outside temperature is above zero.  Under those conditions, damp air would warm up the evaporator towards ambient more quickly, because of it's greater heat capacity.

 

The local humidity thing is really more about trying to prevent the area around the ASHP from being wetter than elsewhere.  Although they are designed to be left uncovered, having rain running down the evaporator will make it ice up more quickly, I'm sure, than if it was kept sheltered from the rain.

[SteamyTea]

Do we know what temperatures the evaporators run at, I have -23°C in mind.

If the air flow was faster though the evaporator, would that increase the temperature of the evaporator (of the fins really), helping to reduce icing.

Would make for a noisier unit.

 

Of course you can just get a larger ASHP and run it slower.

[Jeremy Harris]

30 minutes ago, SteamyTea said:

Do we know what temperatures the evaporators run at, I have -23°C in mind.

If the air flow was faster though the evaporator, would that increase the temperature of the evaporator (of the fins really), helping to reduce icing.

Would make for a noisier unit.

 

Of course you can just get a larger ASHP and run it slower.

 

 

It varies a great deal with load.  There are times when the evaporator on ours is barely cool, and other times when it's well below zero.  In general, if our ASHP is running at it's normal pretty low load, and the air temperature is above about 5 or 6 deg C, then the evaporator barely reaches 0 deg C.  Push the load up, by asking it to deliver water at a higher temperature, and the evaporator cools down pretty dramatically, even with the fan speed having increased to try and shove the maximum amount of air through it.

 

From what I've observed, the first thing to change when the load changes on the ASHP is the fan speed.  The internal logic and timing seems set to speed up the fan first, to establish a higher air mass flow and get the evaporator as close to ambient as it can.  It then very gently ramps up the compressor speed, so cooling the evaporator.   Somewhere in the settings I think there is a way of displaying the effective evaporator temperature from one of the internal sensors, I may have a go and see if I can get a better feel for how cold it gets.

 

The lower temperature limit for the evaporator is set by two things, the static pressure inside the evaporator and the saturation pressure of the refrigerant.  Some refrigerants have a pretty low saturation pressure at low temperatures. like CO₂, for example.  This means a CO₂ filled ASHP will still function with the evaporator down well below -30 deg C.  The more typical zeotropic mixes that are used in most modern units (typically R410A, which is a 50/50 mix of difluoromethane and pentafluoroethane)  have saturation pressures that are zero at around -40 deg C, so, depending on the pressure in the evaporator circuit, they may work down to about -25 deg C or so.

Edited November 25, 2017 by JSHarris

[pdf27]

One interesting thing I found the other day is a complete set of what are effectively COP tables (heating and electrical power, so you have to work out the COP for yourself, but it isn't hard) for Samsung heat pumps (attached). The impact of low flow temperatures is huge - if it is supplying water at 30°C the COP is still greater than 3:1 even at -20: supplying water at 55°C doesn't achieve this until the ambient temperature reaches about +10°C.

This is where phase change materials like Sunamp  or thermal stores are potentially very interesting - DHW flows should ideally not exceed 48°C, but stored water needs to regularly exceed 60°C for legionella protection. At 45/2°C (hot water/outside air) the COP is 2.75 - increasing that to 55/2°C drops it to 2, and resistance heating is needed to boost it after that - dropping effective COP down to about 1.8. Same thing applies in summer, although at 20°C the numbers are rather better - 4.85 and 3.2 respectively.

 

That's actually something I want to have a play with in the next couple of weeks if I can find the time with a spreadsheet - heating degree day temperature records are available for download, and Sheffield Solar have downloadable PV generation data. That makes it quite easy to do a crude spreadsheet model of annual COP in a Passivhaus - you know peak consumption is about 10W/m² of low temperature heat with no solar or internal gains, which gives you a W/°C value and hot water demand can be estimated as well. I've got a internal gains figure of ~10kWh/m²/year from my architects (call it 1W/m²). The idea is then to balance solar gains (assuming no gains when heat is not needed and using the Sheffield Solar data integrated over a day as a measure of solar resource) against this to give the 15 kWh/m2/year total consumption.

That then gives a spreadsheet with a couple of years of real world data showing air temperature and thus COP coupled with heat demand on that day. A very similar calculation can be done for a GSHP, just taking the seasonal average temperatures and the associated COP values. I have a strong suspicion that the ASHP will come out with lower power consumption, depending slightly on how much hot water use I assume - the higher the hot water use fraction, the more an ASHP will benefit from the higher summer air temperatures pushing the COP up.

Performance-Data-Technical-Data.pdf

[Jeremy Harris]

Bear in mind that all the COP data almost certainly ignores the energy loss in defrosting.  That's certainly the case for the data I've seen, where the COP has been quoted at a figure for a certain temperature differential, but when you estimate and add on the actual heat energy pulled back out of the house during defrost (which has to be replaced when the heat pump starts running in heating mode again) you find that the real world COP can be around half the quoted figure.

[pdf27]

I'm not sure if that's actually a big problem in a well designed house - design low temperature around me is somewhere around -3°C, at which point heat demand will be about 1.5kW. Since it's hard to find a heat pump which can supply pumped heat at a low enough temperature to heat a Passivhaus slab directly (the Samsung units for instance use resistance heat in the 20-30ish temperature range for instance) then it's likely to be cycling on/off. At +2 in damp air, the cycle might be an hour at 5kW followed by 4 hours off to defrost. What you lose in a heat cycle to get rid of the accreted ice you might have gained in the first place by the phase change on the ice releasing latent heat into the evaporator - and even if you didn't the effect over the course of a year may well be quite modest.

 

 

[Jeremy Harris]

37 minutes ago, pdf27 said:

I'm not sure if that's actually a big problem in a well designed house - design low temperature around me is somewhere around -3°C, at which point heat demand will be about 1.5kW. Since it's hard to find a heat pump which can supply pumped heat at a low enough temperature to heat a Passivhaus slab directly (the Samsung units for instance use resistance heat in the 20-30ish temperature range for instance) then it's likely to be cycling on/off. At +2 in damp air, the cycle might be an hour at 5kW followed by 4 hours off to defrost. What you lose in a heat cycle to get rid of the accreted ice you might have gained in the first place by the phase change on the ice releasing latent heat into the evaporator - and even if you didn't the effect over the course of a year may well be quite modest.

 

 

 

 

I've proved it's not - if I just keep the flow temperature at no more than 40 deg C then the ASHP never defrosts at all.  Boosting from the 40 deg C in the buffer to 50 deg C or so for hot water uses very little energy, especially as most of the time that comes from a Sunamp PV, and if the Sunamp PV isn't charged, the variable power input instant water heater (a Stiebel Eltron DHC-E) will add just enough extra heating to bring the pre-heated water up to a usable level for DHW.

 

Our ASHP is oversized simply because a 7kW one came along at a bargain price.  Our house heating requirement when it's -10 deg C outside is only about 1600 W, so the heat pump never has to work very hard.

 

Accreted Ice isn't involved, as the design of the heat pump is such that it defrosts before ice builds up.

Edited November 25, 2017 by JSHarris

[pdf27]

1 minute ago, JSHarris said:

 

 

I've proved it's not - if I just keep the flow temperature at no more than 40 deg C then the ASHP never defrosts at all.  Boosting from the 40 deg C in the buffer to 50 deg C or so for hot water uses very little energy, especially as most of the time that comes from a Sunamp PV, and if the Sunamp PV isn't charged, the variable power input instant water heater (a Stiebel Eltron DHC-E) will add just enough extra heating to bring the pre-heated water up to a usable level for DHW.

Thanks. So on that basis you'll never see heating related icing issues for a well designed modern house with slab heating (radiators may need more thought, but they aren't relevant in my case). That means it's only an issue for heating DHW from 40 to say 55°C: that's of interest to me because we're considering a demolish/rebuild on a house we previously only previously considered refurbishing, which means any complicated options aren't going to be financially viable - if you can do hot water off the HP,  that's what'll happen.

 

The Clark/Grant paper gives measured hot water consumption in real life being in the region of 25-40 kWh/m²/year - for a 150m² house that's 13 kWh/day including losses in the primary circuit, etc. That seems to be a bit of an overestimate - for our current 100m² bungalow the combi uses 160 kWh/month of gas when we're home all summer with the heating off (20 kWh/m²/year of gas - probably about 16kWh/m²/year for actual delivered hot water). Essentially that means the heat pump would only ever risk running in icing conditions for ~1h/day - after which it can defrost naturally - and on that basis I should be pretty safe using the claimed COP figures without making allowances for heat lost to the defrost cycle.

[Jeremy Harris]

What I also found (there's loads elsewhere here about it) is that increasing the flow temperature from even a massively over-sized ASHP makes it defrost a lot.  There's no gradual transition that I could find.  Running at 40 deg C flow means it never defrosts, running at anything over 45 deg C means it defrosts at least twice an hour, sometimes three times an hour, when the outside air temperature gets to around 4 to 6 deg C and the humidity is high (typical English winter day!).  Each defrost cycle lasts about 10 minutes, and seems to work the heat pump hard, presumably to get as much heat from the house into the evaporator as quickly as possible.  The result is that lost energy from the house during the defrost cycle is often a lot greater than energy gained during the normal operating cycle for the same time period, as the ASHP doesn't seem to modulate down in defrost mode.

 

It was quite a shock the first time I saw a defrost cycle, and saw the 70 litre buffer tank temperature plummeting down to well below room temperature as heat was being drawn from it!

[dpmiller]

Having watched a number of youtube videos of defrost cycles, I'd say that a heat cycle of 10mins is waay too long and no more than a minute or two of heating should suffice. @JSHarris have you ever watches the defrost to see how much heat is actually needed?