@iancalderbank Realised that I would have to get into the control panel right after posting 🙂 I remembered that when I changed pumps, I had to alter the connections. thanks!

Posted by: @derek-m

↑

Quite. It is necessary to ask a prospective installer to provide a copy of the capacity data tables for their proposed heat pump and see how that compares with accurate heat loss calculations. If an installer cannot provide such data, and justify their heat pump selection, then it may be time to search for a different installer.

To answer your query relating to buffer tanks, the first thing to clarify is if your prospective installer actually knows the difference between a buffer tank and an expansion vessel. If they don't then again it may be time to seek a different installer. A buffer tank will not deal with the expansion of the water as it is heated, whilst an expansion vessel, by virtue of its name, will.

A typical system would normally have two expansion vessels, unless the hot water cylinder is a vented type. One expansion vessel for the primary water side, the water that carries the thermal energy from the heat pump to the radiators, UFH loops, DHW cylinder etc. A second expansion vessel would be required on the hot water system if the hot water cylinder is un-vented. If you Google 'expansion vessel' you will be able to see some photo's and probably an explanation of how an expansion vessel works.

Your installer's comment about requiring two buffer tanks, one for radiators and one for UFH, would cause me concern. Is your installer also proposing a system that runs at a high, fixed, LWT, along with zone valves, room thermostats and TRV's? If so, that would probably be the reason for also installing buffer tanks. Maybe time for a different installer, since your prospective one does not appear to understand heat pumps and how they should be operated and controlled.

Let me explain the possible problems that a buffer tank and the other equipment may create and then you will be better armed for discussions with prospective installers.

Installing a buffer tank would create the requirement for an additional secondary water pump, since the primary water pump 'pushes' the water into the buffer tank whilst a secondary water pump is required to 'suck' the water out of the buffer tank.

The speed of at least one, if not both, of the water pumps needs to be controlled, so that the flow rate going into the buffer tank is the same as the flow rate coming out. It may be necessary to measure the flow rates in and out of the buffer tank, or at least the temperatures of the water flowing, so that the flow rates can be balanced.

If the flow rate going into the buffer tank is greater than that coming out, then not all the thermal energy being produced by the heat pump will be transferred to the heat emitters. This may therefore cause the IAT to fail to achieve the desired level. To achieve the desired IAT would require the LWT to be increased, which I think is now fully understood will reduce the heat pump efficiency. It could also lead to premature cycling.

If the flow rate going into the buffer tank is lower than the flow rate coming out, some of the cooler water returning from the heat emitters, will mix inside the buffer tank, with the warmer water flowing out to the heat emitters. This again would mean that the heat emitters would not receive the full quantity of thermal energy, the LWT would need to be increased, and the heat pump efficiency would be reduced.

There is a good 'heat geek' video on youtube which explains this better than mere words can do.

Having zone valves opening and closing will unbalance the flow rate through the buffer tank, as will having TRV's throttling in or opening up.

So I would suggest that if at all possible, buffer tanks, plate heat exchangers, zone valves and most TRV's should be avoided. TRV's in bedrooms may be permissible and room thermostats for possible cycling reduction during milder weather conditions.

I am a firm believer in keeping heating systems as simple as possible.

 

Thanks - that is very helpful.  On reflection I'm pretty sure the installer was talking about expansion vessels being needed, and it's me that's got confused between this and the buffer tank chat.  They sound similar; your explanation is super helpful. 

Finally on the topic of expansion vessels, the Daikin indoor unit with a tank seems to say it has one inbuilt (page 13 here - https://my.daikin.eu/content/dam/document-library/catalogues/heat/air-to-water-heat-pump-mid-temperature/epra08-12ev/Daikin%20Altherma%203%20H%20MT-HT_boiler%20replacement%20range_Product%20catalogue_ECPEN22-767_English.pdf) so does that remove the need for additional ones as part of the install?  If so it's certainly a much tidier end result

If my installer does suggest a buffer tank (I've yet to get a detailed breakdown from them so don't know) I'll definitely dig into that in more detail, your detailed info is much appreciated.There is lots of info online where people say they are needed for heat pump installs.  Here is one I found (not my installer) https://www.nu-heat.co.uk/blog/a-closer-look-at-buffer-tanks/   Would be good to understand why this thinking is incorrect.

@squiff

The size of any required expansion vessel is dependent upon the total volume of the system (radiators, UFH and pipework etc), the greater the system volume, the larger the expansion vessel will need to be.

Concerning the two reasons given by Nu-Heat for installing a buffer tank. If the system requires extra volume, then fit a volumiser, not a buffer tank. A volumiser is similar to a buffer tank, but instead of have two input connections and two output connections, the volumiser only has one input connection and one output connection. A volumiser would normally be placed in the return pipework to the heat pump, where it will store some thermal energy in the form of warm water, which can then be used to assist during any defrost cycle, without taking as much thermal energy from the heat emitters. It may also assist in reducing cycling, since it would provide additional load to the heat pump.

I don't understand Nu-Heat's thinking with regard to hydraulic separation because of different temperature requirements. For optimal efficiency, the warmed water coming out of the heat pump should be fed directly to the heat emitters to provide maximum thermal energy transfer. As explained previously, having the water flow through a buffer tank to be cooled slightly by mixing seems rather pointless and inefficient. If water going to the heat emitters is too warm, then don't warm it as much in the heat pump and improve overall efficiency. Simples. 🙄 

If the heating system contains both radiators and UFH, which may require different water temperatures, the normal method is for the radiators to be supplied directly from the heat pump with the warmer water, this warmer water is also supplied to the UFH system, which has its own water pump and mixing valve, which cool down the water to the temperature required by the UFH loops.

I suspect that Nu-Heat may use the installation of buffer tanks, to avoid the need to correctly design, commission and optimise the systems they are installing.

I must say that the recent posts about buffers are most encouraging.  When I first embarked on this journey it seemed almost heresy to suggest on a public forum that a buffer was unnecessary in most/many cases, and that the preferred number of buffers is zero.   Also the idea of a volumiser as opposed to a buffer was almost  niche.  Now there is little or no debate. That's real progress in understanding what is, in fairness to us all, unfamiliar technology 

From time to time discussions might get heated, but the value of robust public debate is clearly evident.

Posted by: @iancalderbank

↑

Posted by: @bontwoody

↑

The point I am making is that a slightly undersized heat pump may not just be cheaper to run because is smaller, but also because it is running at a higher COP. I am not a heat pump engineer so I am just relaying opinion from people who know more than me and you may not agree with those views.

but a somewhat oversized heat pump that is cycling in low demand, can also give a high COP, if everything else is done properly. Mine from this morning with a very low heating demand, cycles 3 times, COP is 5.2

-- Attachment is not available --

one cannot simply make a blanket assumption that smaller heat pump gives better COP at low load.

Posted by: @bontwoody

↑

The DHW issue is a bit of a red herring in my opinion as I used to recharge a 300lt cylinder to 52C daily in well under 2 hours using a 5kW ASHP in my last house.

disagree about the DHW red herring: with a 16kw, I can reheat my 300L UVC in just over 30 mins. which means less downtime for the heating loop, and less "grumble time" if someone is complaining that hot water has run out.

Posted by: @bontwoody

↑

@squiff As I said in my first post "Its probably best to do your own research for this and go with what you think as you may not get a consensus here." LOL

agree 🙂

Posted by: @jamespa

↑

It's unfortunate that we currently don't have a fool proof way to measure the loss that is generally accepted, if we did then choices could be more finely tuned and better informed.

totally agree. that, in a nutshell, is the root cause. If one genuinely knew, for absolute certain, what the house needed, then you could make an informed decision what size you wanted.

but if the 10kw you've been given by the survey has already been doubled over the actual figure by the uncertainty of the heat loss methods, and then someone adds more "for safety" you could end up with a 16 when your house only needs a 5.

 

OVERSIZING. to put a pinch of balance back into this debate.  There is a risk of drawing conclusions based on assumptions about performance of an oversized heat pump when there may be 3 or 4 other reasons for rapid reheat of DHW for example or rapid cycling.

We have what might be considered a small heat pump for a 180sq M  older type property So our Ecodan 8.5 delivers heat to 15 radiators fully adequately down to -8c as at last December’s cold snap.  So to my mind there’s are other things at play rather than just sizing.

(by the way I don’t think our HP is under sized but rather it reflects the uvalues of the building and air leek controls in place) 

Rapid reheating of DHW. We experimented with flow-rate.

The attached 2 graphs show the same, 8.5kw HP producing the DHW to the same temp - note duration between graphs 36 minutes for one and 52 minutes for the other. The only difference between the two setups is the flow rate.

The shorter re-heat was produced with a flow rate of 13LPM and producing a DT of 7c, while the longer reheat had a flow rate of 20 LPM and a DT of 3-4 LPM

Clearly if  some makes of HP are capable of varying automatically the flow rate to achieve a specific DT then this would help with efficiency. However there are several makes which do not operate with PWM controls and rely on a pre set flow rate. so an understanding of the best room heating / DHW flow rate is needed.

This, in turn raises the question of cylinder coil design. Why are the vast majority of HP Ready cyl. coil sizes and connections 22mm when all Primary Pipe sizes (above 6kw) are at least 28mm? This often leads to a narrowing of the DT as the DHW REACHES THE END OF ITS CYCLE.

My last observation on HP sizing concerns pipe capacity through the branches of distribution. This has a major influence on performance of output. If the branches of the pipework do not have the diameter size for the radiators it supplies then he HP output will be compromised. And on a retrofit installation - who actually knows how balanced the pipework actually is?

So in short. A 8.5 rated Ecodan can reheat a 2500 cylinder in approx 36 minutes and does not cycle 3 times an hour in a typical April May month when set up with balanced emitter distribution pipework. However it will cycle 3 times per hour when the outside temperature is 17c and the inside start temp is 20c and we’re trying to reach 23c (obviously)

@sunandair

Thank you for carrying out the detailed testing and producing the excellent results. Whilst it is possible to predict likely outcomes in theory, whenever possible, it is essential to test the theory on a live system.

Interpreting practical results also allows any theoretical models to be refined, so making them more useful for future use. Prior to reading your post I was in the process of adding the likely affects of variations of DT and flow rate to some of the models I produced previously.

So what appears to be happening within your system?

The laws of physics inform us that it will take a certain quantity of thermal energy, to increase the temperature of a volume of water, by a number of degrees. This is approximately 1.16W per litre of water for each 1C increase in temperature.

The rate at which this volume of water will increase in temperature will be dependent upon the rate at which the thermal energy is supplied.

It may therefore be assumed that by increasing the flow rate, the quantity of thermal energy will be increased, so the water will heat up quicker. Your data shows that this assumption is not correct.

The other defining factor is the DT between the flow and return temperatures, since increased flow rate may be supplying more thermal energy, but this will only be of use if the additional thermal energy is being transferred into the water inside the DHW cylinder.

Performing calculations on the data obtained-

Test 1.

Specific Heating Capacity of water x volume of water (lpm) x Delta T x number of minutes = 1.16 x 13 x 7 x 36 = 3800Wh

Test 2.

Specific Heating Capacity of water x volume of water (lpm) x Delta T x number of minutes = 1.16 x 20 x 3 x 52 = 3619Wh

Specific Heating Capacity of water x volume of water (lpm) x Delta T x number of minutes = 1.16 x 20 x 4 x 52 = 4825Wh

So it would appear that during both tests, the quantity of thermal energy transferred from the heat pump to the DHW cylinder, was in the region of 4kWh, which would equate to a temperature rise in the order of 16C.

You have correctly identified the importance of flow rate, and DT, when it comes to the efficient transfer of thermal energy, since it is not always optimal to push the water around the system at a rapid rate, and thus not allow sufficient time for the transfer of thermal energy to take place. I suspect that this may be true at both ends of the equation, since the heat pump will have a finite rate at which it can put thermal energy into the water, while the DHW cylinder and heat emitters will have a finite rate at which they can absorb thermal energy from the circulating water.

It should be noted that transferring the thermal energy more efficiently does not mean that the overall system performance will be more efficient, in fact I suspect the opposite may be true.

The tests carried out and results obtained would indicate that the design of a heat pump system is not just a matter of sizing of the heat pump itself, but also the type and size of the heat emitters along with the possible limitations of the DHW cylinder chosen.

Even more confusion to discuss. 🙄 

@derek-m 

Thanks for your comments Derek. As an aside,We have altered our system so we now have removed the LLH, removed the second pump and checked out all pipe sizes and junctions. I’ve also had a flow setter fitted but probably won’t  be necessary. This has given us control of flow rates without the hazard of mixing temperatures which would have happened through the Low loss header as it was originally set up.

After copious sketches and discussion our installer did all this alteration completely Free of charge so there is a lot to be said for healthy relationship with your installer - and frank knowledgeable discussion.

We even got a slightly bigger 8m head pump since the new single circuit needed to cope with both the Primary Circuit and the heating circuit.

Posted by: @sunandair

↑

@derek-m 

Thanks for your comments Derek. As an aside,We have altered our system so we now have removed the LLH, removed the second pump and checked out all pipe sizes and junctions. I’ve also had a flow setter fitted but probably won’t  be necessary. This has given us control of flow rates without the hazard of mixing temperatures which would have happened through the Low loss header as it was originally set up.

After copious sketches and discussion our installer did all this alteration completely Free of charge so there is a lot to be said for healthy relationship with your installer - and frank knowledgeable discussion.

We even got a slightly bigger 8m head pump since the new single circuit needed to cope with both the Primary Circuit and the heating circuit.

 

Have you noticed any differences since the alterations were completed?

Posted by: @derek-m

↑

The tests carried out and results obtained would indicate that the design of a heat pump system is not just a matter of sizing of the heat pump itself, but also the type and size of the heat emitters along with the possible limitations of the DHW cylinder chosen.

Well exactly that.  Much of the discussion above focussed on 'delta T' being in this case the difference between flow and return temperatures.  But this is a secondary variable in most cases. 

The driving variable is how much energy is transferred from emitter to room (space heating) or emitter to water (water heating).  This is not determined by delta T (flow - return), it's determined by delta-T (flow - room/water).

If you want, for whatever reason, to change the amount of energy that transfers to the room or the water you will have to change the _flow temperature_, not the flow-return deltaT.  This definitely changes  the heat pump efficiency and may or may not change the flow-return deltaT depending on the heat pump design.  Some react by changing the flow rate (by changing the pump speed) to keep the flow-return deltaT constant, others don't bother controlling the pump speed so the flow-return deltaT will change as a result.

Basically, if you want to predict the behaviour of a heat pump _start with the emitters not the heat pump_!

Posted by: @derek-m

↑

while the DHW cylinder and heat emitters will have a finite rate at which they can absorb thermal energy from the circulating water.

As above it's really more about the rate that heat is transferred to the water in the dhw tank or the air in the room, which depends principally on flow temperature not deltaT (flow - return) (to be absolutely precise there is a small dependence on flow-return deltaT, but it's much less than the dependence on flow temperature itself which, in any given tolerably well set up system, is the primary thing you will need to vary to change the heat transfer to room/dhw).

I say again, if you want to predict the behaviour of a heat pump _start with the emitters not the heat pump_!

Looking at the graphs, for some reason (presumably dependent on the specifics of the control loop) the flow temperature rises faster in the first graph than in the second, which I am fairly certain is what accounts for the reduced reheat time.  Presumably what is happening is that the hp is reacting to a larger deltaT by raising faster the amount of energy it supplies by raising the flow temp faster in the first graph than in the second.  I'm not sure that we can tell whether this is more or less efficient (in energy terms) than the second case, it's certainly more effective (in terms of reheat time).

Posted by: @sunandair

↑

This, in turn raises the question of cylinder coil design. Why are the vast majority of HP Ready cyl. coil sizes and connections 22mm when all Primary Pipe sizes (above 6kw) are at least 28mm? This often leads to a narrowing of the DT as the DHW REACHES THE END OF ITS CYCLE.

I think because 6-8kW, which a 22mm pipe can deliver at a typical DT of 5, gives a sufficiently fast recovery time for many people and...

22mm gives a sufficiently fast recovery time for most people at dt20, typical of a boiler (influence of history)and...

If you already have 22 mm feeds to the dhw cylinder, it avoids replacing them which is hellishly disruptive.

Given that many dhw tanks are heated by a 3kW immersion, a 6kW feed from a heat pump is considered sufficient by many.

It is true that the DT will narrow towards the end of the cycle, but increasing pipe size won't help.  This is because, at this point, you are limited by the rate of transfer from coil to water not the rate of transfer from heat pump to coil.  The only ways to change the rate of transfer from coil to water are a. Increase the flow temperature or b. increase the coil size.  Changing the pipe diameter won't help at this point.

@jamespa

I partially agree with your assessment, but I would say that the determining factor of how much thermal energy is transferred is the average temperature (between flow and return) to the air/DHW DT, rather than merely the flow temperature to air/water DT.

The flow return DT and flow rate therefore come into the equation.

 In a system with a fixed flow return DT of say 5C, and a flow temperature of 35C, the average heat emitter temperature will be 32.5C. If the flow temperature is now increased to 40C, the average heat emitter temperature will increase to 37.5C, which in turn will cause more thermal energy to be emitted, and more thermal energy to be extracted from the water. If more thermal energy is extracted from the water then the return temperature will start to reduce, causing the flow return DT to increase. To re-balance the system and restore the require flow return DT to 5C, the heat pump will need to increase the flow rate and also produce more thermal energy to maintain the higher flow temperature.

In a system with a fixed flow rate, an increase in flow temperature from 35C to 40C would initially raise the average heat emitter temperature to 37.5C, but as the thermal energy being emitted starts to increase, the return temperature would start to reduce from 35C, which would in turn cause the average heat emitter temperature to also fall. To supply the same quantity of thermal energy as the fixed DT example above, the actual required flow temperature may need to higher or even lower than 40C and will be dependent upon the fixed flow rate.

Posted by: @derek-m

↑

I partially agree with your assessment, but I would say that the determining factor of how much thermal energy is transferred is the average temperature (between flow and return) to the air/DHW DT, rather than merely the flow temperature to air/water DT.

I completely agree.  However DT can only change by a small amount (~5C) in a tolerably well adjusted setup and each degree it changes is equal to half a degree change in average temperature.  So, if you want to change the output, you change FT not DT.  That is, after all, how WC works.  A small change in DT may or may not follow, depending on whether (or not) your heat pump modulates the water pump to maintain a fixed DT.

The mistake some (certainly not you) make is to concentrate on dt*mass flow rate when trying to do the analysis of efficiency etc, and then think about the effect of changing these variables (or what changes occur to these variables).  This is, in most analyses, cart before horse.

The exception of course is when you adjust individual emitters for constant room temp (as suggested by heat geek) rather than constant DT.  In this case you are potentially throttling the output of the individual emitter by throttling the flow, thus significantly increasing DT for that emitter only, albeit not for the system as a whole.