Avoid the Heat Pump Villain: Why Low-Loss Headers and Buffers Can Sabotage Your Heat Pump’s Efficiency

Pretty much but worth unpacking for absolute clarity

There are too many dTs in this game

Yes!

Derek rightly corrected me on average flow temp through the rads,

Exactly,  Heat loss from rad to room is dependent on average flow temp through the rads.  The DT across the rads will be driven by this and flow rate not the other way round.
 

so to get back up to the same average flow temp the input flow temp (LWT) has to be increased. Hence one of the sources of inefficiency in the ashp/llh combination 

 
If there is a reduction in flow temperature (flow to flow) across the LLH then yes.  If the LLH is perfectly set up the reduction in flow temperature will be small, so possibly no material inefficiency.  But frequently it isnt perfectly set up and anyway, in the absence of a concrete reason (and in most domestic setups there isnt a concrete reason), why bother with the LLH because it costs money and is the source of problems.

BUT on a boiler system with a higher LW temp say 60C the dT across the rad is a smaller proportion of the dT to the room (thus showing too many dT!) and anyway with a boiler the LWT -efficiency curve is not as severe.

The key is that LWT (and thus the average emitter temperature) is higher, although in a modern system possibly not by that much, if at all.  The latest BRs require the design to be at FT=55.  If you persist with the typical  20C deltaT flow to return, the average rad temp will actually be 40C, a bit lower than the 45/5/40 combination used frequently for an ASHP, but in practice essentially the same.  
However typical practice until recently (and I suspect still in the vast majority of cases) is to set (or rather leave) a boiler FT at ~75 which does lead to higher average rad temp.  Actually this very likely also causes quite a severe efficiency hit because its entirely possible/likely that the boiler wont condense because the return temp is too high – you can tell because a boiler that isnt condensing will have plumes of water vapour emitted from the flue – which you see all the time!  Thats worth, I believe, ~10%.  However almost nobody takes any notice of the efficiency of gas powered central heating systems.  One might, very cynically, argue that the market actively distracts them from doing so by providing lots of ‘controls’ with various degrees of smartness that purport to improve efficiency but may not do so by anything like as much as we are led to believe. 
Just like an ASHP, one of the best tips for getting greatest efficiency (and comfort) from a gas boiler is to turn down the flow temperature to as low as you can consistent with the house being sufficiently warm!

After around 6 years in, after reading up on efficiency, I decided to remove the buffer tank. I reasoned that the system was wasting lot of energy simply keeping the buffer tank up to temp. A fact confirmed by our Plumber, who is experience with ASHP systems as it happens. Our system is quite large, so surely there was enough water buffering in the pipework with out the need for buffer tank, we opined.
So, we duly removed the buffer and made some tweaks to the pipework as necessary token the system flowing.
Results were staggering. Roughly 25% drop in energy input required to give same heat output. I have measured this over several years to eliminate a mild one off winter. 
So no Buffer or LLH in our system seems to be the answer.

Thanks for sharing this and very pleased to hear the result.  Well done for having the courage.
 
Its worth being aware that there are (at least) two reasons why removing a buffer tank saves energy, and its not necessarily heat loss from the tank that is the major contribution:
1. You aren’t having to replace the heat lost from the buffer tank – however if the buffer tank is inside the heated envelope the energy isn’t really ‘lost’, and anyway with a well insulated buffer tank it should be no more than 1-2kWh/day
2. A buffer tank, unless very well set up, results in a temperature drop between the flow from the heat pump and the flow to the emitters because flow and return mix in the tank.  This means that the heat pump must operate at a higher flow temperature to get the same flow temperature at the radiators, and that means it is less efficient (lower COP) by very roughly 3% for each degree drop.
The latter is a function of the thermodynamics of heat pumps so doesn’t apply to fossil systems, which may be why some heating ‘engineers’ don’t understand it!
It is possible to set up a sufficiently tall buffer tank to avoid the flow and return mixing, and thus circumvent (2), but there are far too many reports where this has not been done (or is not even possible given the equipment/wiring used) 
 
 

 

Both however are still a bit lessay fair about the 1m distance and have only said it needs to have the fan pointing away.

. Makes technical sense and might be something along the lines of what I hear the MCS are looking to remove from requirements… but current guidance/requirments still says 1m.

This a matter of law not an MCS decision or even guidance.  End of unless you want to apply for planning permission.

https://www.legislation.gov.uk/uksi/2015/596/schedule/2/part/14/crossheading/class-g-installation-or-alteration-etc-of-air-source-heat-pumps-on-domestic-premises

 

One had said the buffer was needed not technically but to cover manufacture warranty as thats how they specify the design (which made sense but would value other’s experience/input).

You mention it’s a Vaillant.  As it happens I was looking at their their standard plumbing diagrams yesterday and certainly the one I looked at shows a 2 port volumiser (but not a buffer tank).  What I’m not sure is whether the text clarified the conditions under which it’s necessary (basically if you haven’t got sufficient system volume). Unfortunately installers appear to have a habit of assuming that if it’s in the diagram it’s necessary, which is not always true. 

Vaillant tech support are very helpful so you / your installer could ask the question. 

Don’t accept a 3 or 4 port buffer, there are definitely diagrams in the Vaillant approved set that have only a volumiser so you are being given bs if they say a 3 or 4 port is necessary to cover manufacturers warranty.  If they insist look elsewhere.

These are all very good questions.  Unfortunately most people are reliant on installers to do the ‘design’ and decide what valves are (or are not) put in, so don’t have this option.  

Those who, like me, had the fortune to ‘lurk’ here for a few months before getting their heat pump installed (not out of choice – it was serendipity) can do/did just this and are thus in a position to reject offers from installers that want to do silly things.

@jamespa Invoking Occam’s razor in this context does not help. In engineering, there are right answers and wrong answers, not alternative facts.

I am referring to Occams razor in the sense of cutting out the unnecessary and I explain this in my post.  If its unnecessary (in a particular application) then its unnecessary.  Thats a fact, not alternative.

That said, in my personal experience of engineering, there are often a range of ‘right’ answers with different trade offs (as well as some wrong ones) and the trade offs you choose depend on the circumstances and customer.  A 2CV and a Bentley are both valid solutions to the problem of creating a working car to get you from A to B, appealing to different markets.

In your first example I think you conclude that an LLH/buffer tank is not necessary but that the downstairs rads could be replaced with UFH to give adequate system volume for defrost/to meet the spec of the heat pump.  The latter may well be true but it is rather disruptive and expensive unless the work is going to be undertaken for another reason.  A much simpler and cheaper way to solve the problem is to add a volumiser (a 2 port buffer tank if you wish) which has minimal side effects because it is 2 port.  You will  note that, in my post, I draw a distinction between buffer tanks and volumisers for precisely this reason, a volumiser has few side effects.  Of course it only solves one problem, system volume, but if that is the problem to be solved a volumiser is the right choice.

In the second (larger house) variation you suggest adding a 50l buffer tank in order to supply a bathroom as a zone to run on its own.  I would question the wisdom (or utility) of designing a system to do this in the first place but agree that, if you must do this for some reason, its likely that a form of separation will be needed.

Finally you comment that

So these components are not used “willy nilly”. They each have a purpose and do not cause efficiency problems.

Unfortunately the evidence from this forum, over the two years plus that I have been following it, is that buffer tanks and LLHs are used willy nilly by some installers.  Furthermore a badly designed buffer tank/buffer tank with badly designed controls certainly does cause efficiency problems, typically involving a sacrifice of 15% or thereabouts and/or poor heating capacity.  Hence the title of this thread (not my title incidentally)- “Why Low-Loss Headers and Buffers Can Sabotage Your Heat Pump’s Efficiency.".  Many of the instances of poor performance on this forum can be traced to unnecessary buffer tanks /llhs.  Its sad that this is the case, but that is the fact.  

As a general reflection on your comments on ‘control philosophy’ it is pretty clear that the benefits (or not) of zoning and setbacks are both variable and themselves the subject of some discussion.  Heat Geek (the website) gives a worked (and very plausible) example where physical zoning increases cost, because the reduced efficiency (due to smaller emitter area=higher flow temperature) outweighs the saving in heat loss.  There are many variables, but I think its fair to say that its far from clear that zoning in a modestly sized, roughly cubic, house with little or no internal insulation may not save much or any money, and may even cost more.  Setbacks are probably clearer, in many cases a modest setback will likely save money, however it is possible to come up with circumstances where this is not the case.  Due care is thus necessary before assuming that trying to finesse which parts of a house are heated at which times is actually beneficial.

 

therefore in my view there is absolutely no need for a LLH unless there is flow mismatch.

Agree, and if there is a flow mismatch the question is ‘why’ and can it be eliminated using something without undesirable side effects.  In most cases the answer will be yes (or that the flow mismatch arises only because of the LLH).

Finally, a volumizer provides increased water content for defrost but it is in series with flow so a large one will cause hysteresis and will not provide a water bypass flow path.

A correctly sized volumiser will bring the system volume up to the figure required to satisfy the defrost requirement, and thus will have no more hysterisis than is necessary for that purpose.  Obviously if its oversized it will have more, but if your installer cant properly size a volumiser then they surely cant properly design and fit a buffer in a way that does not cause significant degradation of performance.  A buffer connected from flow to return has the potential to short circuit the heating system altogether, and indeed we have seen this in posts on this forum.  A volumiser fitted either in the flow or in the return is fool proof which is an important feature in practice, it seems.  A pressure activated release valve can be fitted if deemed necessary (I have yet to hear a convincing argument for this which does not simply mask a problem, but I grant it does depend on the details of both the heat pump and the system.)

A buffer should only be needed exceptionally.

Agreed, hence the whole thrust of this thread.  Unfortunately buffers, it seems, are fitted more often than ‘exceptionally’.

So these components are not to be used “willy nilly”.

Absolutely agreed that they are not to be, unfortunately the evidence is that they are!

 

My personal view is that we are ‘in transition’.  There has been an unfortunate history of fitting buffers/LLHs principally to avoid the risk of call outs due to low flow (which is masking the problem not solving it) and the need to do so has become to an extent embedded.  There was a point where fitting a buffer was more or less the default for many installers, and there are probably some where it still is.

More recently the problems they cause have been recognised and so the previous practices are being called out and some who used to advocate them admitting the reasons and changing their position.   Its going to take a while before we get to the point where buffers etc are not ‘fitted willy-nilly’ and for that period the public need to be warned, to the extent that this is possible.

Hummm……larger system volume would have more energy stored in the water so the bulk average water temperature reduces less when the defrost cycle happens.  I suppose the heat pump controller would then spend less time going flat out to increase the bulk average water temperature, so that should mean less ice accumulation in the fins.  Interesting!

Bob

I too have been thinking about this, not least because the only disappointment with my heat pump is that its noisy when recovering from defrost.

I think what you say above is right.  The energy needed for defrost remains the same irrespective of the system volume, but larger system volume means that the system water temp will drop less for the same energy extracted.  This (a) means that it can start putting heat back into the house earlier and (b) means the heat pump can afford to spend a bot more time ‘catching up’, which should (i) should mean less ice accumulation in the fins and (ii) means it doesn’t have to be as noisy. 

Of course the recovery algorithm will not necessarily be sophisticated enough to take advantage of this opportunity, so in practice it may well be that system volume makes little difference once it is sufficient.

 

@tim441 if you want to you can work out roughly how much it’s costing by measuring the temperature drop across the buffer flow to flow.  The cost is roughly 2-3 percent per C (more according to some, but I think the argument for more depends on deltaT flow to return being higher (than the typical 5C) on the emitter size.

It’s entirely possible to set up buffer tanks not to mix, but unfortunately most installers appear not to do so, or fit a buffer tank that is to small to stratify.  Of course some do,  @toodles has for some while reported a buffer tank that doesn’t mix to any great extent.

The real point is that they shouldn’t be fitted in the first place because they are, in almost all cases, superfluous.  Obviously once they are there it’s a different calculation.

Logically the electronic motorised valve.. I think… stops buffer being used when heating DHW/hot water tank. So will be on flow to BT?

Is it a 2 way valve or a 3 way diverter valve.  Normally the (3 way) diverter valve is the only motorised valve needed.

As you say taking some temperature measurements when its working harder would be good, but your initial ones are encouraging.

people who recommend the removal of the buffer tank on efficiency savings that can be made should also urge caution and get people to check their warranty conditions first .

Agreed

However if the above is the official lg position it’s a good argument in future to avoid lg, which is disappointing.  I wonder if their quoted performance takes into account a typical buffer tank as installed.  Oh, wait a minute, I think I can guess.

@fillib  It would be very interesting indeed to determine your actual house loss by plotting ASHP heat delivered vs OAT (eg on a daily basis).  I looked in the spreadsheet but couldn’t find the figures needed to do this.

The reason this would be interesting is to understand how oversized the 7kW model was.  Based on 10MWh/yr gas consumption I would estimate your house loss to be 3.5-4kW, so perhaps as much as a factor of 2 oversized (more if your boiler is very inefficient) given that the 7kW model can actually do 8kW at typical FT and design OAT.  I have heard others say that the Vaillant machines are reasonably good at coping with oversizing, but maybe not by a factor of 2!  This plot from a John Cantor video illustrates the problem that heat pump designers have to grapple with:

 

It occurs to me that the fact that the heat pump was oversized may possibly have reduced the negative effect of the buffer.  My understanding is that, by default, Vaillant heat pumps run mostly at fixed water pump speed which is adjusted at commissioning time (by the heat pump controller) according to the heat pump capacity.  So yours will likely be running to achieve DT of ~5 at 7kW, but as your peak load is half this your DT will likely be in the 2-3 range even at peak load or less at a more normal load.  This will reduce the temperature loss across the buffer due to mixing, thereby reducing its negative effect on COP.  All of course assuming (amongst other things) that the Vaillant water pump does what it appears to do on my machine and what some others have said it does.