Low loss headers and plate heat exchangers for air source heat pumps

[dnb]

I have been evaluating options for my heating and DHW provision over the last few weeks. Things have changed (probably for the better) since I started the project several years ago and there are now more options but the physics remains the same - low temperature heat provision is more efficient, even with a new R290 heat pump.  I have a few candidate high level designs from various people, at a range of prices and a variety of designs. Some are simple designs, others look like they might need a dose of simple adding. Unless of course the comlicated is adding some efficiency or reliability I haven't yet considered. I am now trying to evaluate things so I get the best value, and ideally take advantage of the government grant - I believe the house meets all the requirements. Add to this that I am trying to take best advantage of 6.5kW of south facing solar generation. Using solar to run the heat pump, acknowledging I can't modulate it via PWM like an immersion heater is an acceptable solution since there are very few scenarios where the COP of the heat pump drops to unity, and the immersion heater on its own can't supply all demands just on the yearly expectation of solar power. This adds a requirement for some kind of buffer (or batteries, but I can't make the cost vs return maths work yet).

 

A couple of the designs contain a low loss header or a plate heat exchanger on the heat pump side of things. The literature on plate exchangers indicates they are nominally very efficient at their job, but it puts an additional pump in the scheme, and makes a temperature difference from primary to secondary outside of the tank. Is this something to avoid? The low loss header puzzles me - the internet suggests they need very careful design to not be lossy. Is this a way of trying to avoid/reduce the need for buffering? I've managed to make (on paper at least) a near passive house with a heat demand of 16 to 17 W/m^2 depending on who's modelling you believe - mine says 17.2, the EPC says 16, but rounds down - but I do have a significant DHW demand so I am trapped between a 5kW heat pump that will do the heating well but struggle for time with DHW or a 12kW pump that could conspire to short cycle like mad just doing heating if it's cold but not freezing. 

 

Most of the above is background so that the context of the actual question is evident. I like a good deal of the solution involving the plate heat exchanger but am keen to check for devils in the detail.

 

Are plate heat exchangers actually bad for heat pumps and efficiency if they are properly sized and insulated? I am in 2 minds because coils in tank don't seem particularly efficient either in some cases, but maybe this balances out and coils are cheaper, but this depends on how much I believe in tank stratification, whether I want to mix antifreeze into the heat pump water or whether I drive the heating directly from the heat pump.  The usual thing... Ask 3 people get 3 different answers...

 

I assume the low loss header is not the right answer for me, although it does meet some of my requirements to separate usage from generation.

 

[JohnMo]

You don't mention zones, if you have any or many or just one, that will make a difference, nor do you mention how you heat radiator or UFH or a mixture?

 

Generally LLH, PHEs add to required flow temperature required from the heat pump which all have a negative impact on CoP.

 

You need to seperate the design of the heat pump and use of PV, they may be related but different subjects really.

 

Hot water cylinder are easy, size to suit your usage, specify a 3m2 coil, done. I heat a 210L cylinder in around an hour from a 6kW ASHP. Have a PV diverter to dump excess solar into water during peak generation times.

 

Anti freeze in the grand scheme of things is cheap. Antifreeze valves require no antifreeze solution in the heating medium. But if you are away and have a power cut, when it's cold, you may have to top up the system to get it going again. Antifreeze doesn't have that issue. I just add about 20% antifreeze, to get freeze protection down to about -6, limited impact on performance...

[Originaltwist]

There's a huge difference between coil in the tank and a PHE. As the latter has to be pumped there is instant destratification in your tank so you call for power and the shower bizzarely goes cold. The wad of heat at the top of the tank put there by your solar diverter and immersion is worth preserving.

BTW an external immersion heater like a Willis should be better for building that stratified wad of heat - easier to service too.

[dnb]

Thanks both. Some good things to think about.

 

8 hours ago, JohnMo said:

You don't mention zones, if you have any or many or just one, that will make a difference, nor do you mention how you heat radiator or UFH or a mixture?

 

No, I don't. This is a thread about PHE and LLH not the system design itself. The amount of zones is irrelevent to coil vs PHE vs direct. If the house is hot enough, it is hot enough regardless of divisions. For the record, UFH on the ground in 3 zones, towel rails in the 4 bathrooms, FCU to put heat into the A/C ducting for upstairs and MVHR. It is all sized so that 40 deg C flow temperature works down to -1 deg C outside.

 

8 hours ago, JohnMo said:

You need to seperate the design of the heat pump and use of PV, they may be related but different subjects really.

They are already separated. The PV is installed, certified and generating.  I want to make best use of the free electric, so I can't leave them in total isolation and consider justice done to the house design as a whole system. It would be like fitting a 50CCA battery to my V8 TVR and expecting it to start. (Truth be told, fitting a 10,000 CCA battery probably wouldn't make it start reliably, but that's another story.) See later for an example. I know I can be a bit "Dirk Gently" about the fundimental interconnectedness of things...

 

8 hours ago, JohnMo said:

Generally LLH, PHEs add to required flow temperature required from the heat pump which all have a negative impact on CoP.

This is my worry. But surely a coil in tank suffers from a similar drop in CoP as the tank warms up, but it's more disguised because there may be a temperature gradient building in the tank.

 

8 hours ago, JohnMo said:

Hot water cylinder are easy, size to suit your usage

Already done. Assume the answer is "Many" litres.  Which is why I know I have to decide how the compromise falls when choosing a heat pump! It all has to work properly in the reasonably forseeable worst case when all said and done.

 

8 hours ago, JohnMo said:

I just add about 20% antifreeze, to get freeze protection down to about -6, limited impact on performance...

That's good to know. It rarely gets below -4. I didn't really want antifreeze everywhere in the heating system and only being one leak away from the DHW hence the additional PHE.

 

8 hours ago, Originaltwist said:

As the latter has to be pumped there is instant destratification in your tank so you call for power and the shower bizzarely goes cold. The wad of heat at the top of the tank put there by your solar diverter and immersion is worth preserving.

The destratification is a concern. Tank baffles etc sound like nice ideas but probably not something that works well enough in the real world so I would have to carefully consider the times when the tank would accept heat if I went this way. Not impossible, but another constraint/compromise for the mix.

 

I wasn't planning on immersion heaters except as a reversionary mode (i.e. heat pump failure) I would ideally need 6kW (to make full use of all the PV at peak) and it makes little sense to me to use something with a CoP of 1 when I could use something with a CoP of say 2.5 and use the PV to partially power it. To give an example, let's say the PV captures 1kWh in the time it takes to heat the tank. The ASHP could heat the tank from cold to full using  3.5kWh [using an average COP of 4.17 calculated by numerical integration WRT return temperature from a very optimistic Samsung datasheet], but I got one of those kWh for free, so I pay for 2.5kWh. If I did the same with an immersion heater I get 1kWh for free, but I now pay for 3.1kWh to get the tank to the same temperature.  The CoP I used doesn't really matter as long as it is above 1 the maths always favours the ASHP unless I can afford to wait all day to see if there is enough PV to get a full tank. And then the PV has only helped with DHW whereas it could have offset both DHW and heating and taken advantage of the ASHP CoP.

 

8 hours ago, Originaltwist said:

BTW an external immersion heater like a Willis should be better for building that stratified wad of heat - easier to service too.

I do like Willis heaters. I have one driving the temporary heating now.  But wouldn't it need pumping thereby defeating the point?

[JohnMo]

6 minutes ago, dnb said:

This is a thread about PHE and LLH not the system design itself. The amount of zones is irrelevent to coil vs PHE vs direct

Bit confused why you think system design isn't done as whole, talking about bits in isolation, without the whole picture, just becomes meaningless. Each part of the system is there for a reason, with considering everything in the round makes zero sense.

 

Engaged volume is important, zones and zone sizes dictate the need for a buffer or volumiser or not, hydraulic requirements of the heat source and the heating system could be very different depending on piping layout etc, so may benefit a LLH or buffer.

 

Cylinders are not heated direct as you have fresh drinking quality inside, so has to be heated via either a plate or coil heat exchanger. Or directly from an immersion. A bog standard invented heat pump cylinder, with a big coil is a standard solution for good reason, good stratification as standard, decent CoP and reheat times. Your CoP when cylinder heating is more likely to be an average of 2.5 to 3.

 

Your PV is just a matter of timing.

 

Other than that I am confused what you you are trying to achieve. Other than reinventing the wheel.

[dnb]

16 hours ago, JohnMo said:

Bit confused why you think system design isn't done as whole, talking about bits in isolation, without the whole picture, just becomes meaningless.

You were advoacting for exactly this with the PV...  Requirements decomposition is an integral part of engineering so that pieces can be considered in isolation - I was hoping we could tease out general reasons why the designs look like they do without me having to publish all of them since  I haven't sought permission to publish. In theory I should do this.

 

16 hours ago, JohnMo said:

Other than that I am confused what you you are trying to achieve.

I am trying to select from a number of options. As I said in the second sentence of the first paragraph of the 1st post.

 

On 13/02/2024 at 11:07, dnb said:

I have a few candidate high level designs from various people, at a range of prices and a variety of designs

 

16 hours ago, JohnMo said:

Your CoP when cylinder heating is more likely to be an average of 2.5 to 3

Agreed, but not according to Samsung!! I don't belileve them either with the CoP of 4, but the maths still works out the same - using the ASHP when there is "some" PV generation appears to be the more efficient way to operate. Few people seems to be advertising this sort of control yet, which I can understand because heat pumps are still a minority product compared to displacing burning gas. I need to get my home Matlab licence working again to try making a simulation for this based on my PV logging.

 

16 hours ago, JohnMo said:

Cylinders are not heated direct as you have fresh drinking quality inside

I never said they were.  "One leak away from the DHW" means the coil leaking into the DHW. Not a fan of antifreeze even one exchanger removed from DHW.

 

16 hours ago, JohnMo said:

dictate the need for a buffer or volumiser or not,

I strongly implied it did need a buffer.  But some of my candidate solutions have one and some don't. I believe the house does need one but need to understand the problem a bit more.

 

16 hours ago, JohnMo said:

Your PV is just a matter of timing.

Yes, and timing implies size of heat pump. An extreme case - If I need all the output for 24 hours to heat the house then there's no time to do DHW.  A more practical case: If there are 4 hours of useful PV on average in the spring and autumn, then the solution may well be to oversize the heat pump such that it has a chance of capturing all the energy we need for the day in those 4 hours (ackknowledging that some energy may be needed from the grid).

 

16 hours ago, JohnMo said:

Other than reinventing the wheel

I fear some of the people I've approached may be selling square wheels. Or octagonal at best.

[Wil]

For DHW- It’s a lot simpler to divert PV into an immersion and deal with the loss of efficiency than trying to match the ASHP input to it. You’ll also be able to sink a lot more heat into the tank with an immersion compared to an ASHP. Ideally I suppose you’d ASHP to 45degrees then immersion everything else for efficiency. 

 

If the sun goes behind a cloud while using the ASHP, then you’re paying for the kWh it consumes whilst the suns hidden (or on a cloudy day). If your PV is at 1kW and your consumption at 250W, then you can put just the 750W extra into the tank. Depends if you care about zero grid consumption or self-utilisation of the PV.

 

[Wil]

12 minutes ago, dnb said:

I never said they were.  "One leak away from the DHW" means the coil leaking into the DHW. Not a fan of antifreeze even one exchanger removed from DHW.

You’re only one leak away from inhibitor and the contents of your heating system sludge too in that case?

[dnb]

On 14/02/2024 at 15:29, Wil said:

simpler

Nail hit on head. It's simpler for the installer. It may not be better for me as the end user.

On 14/02/2024 at 15:29, Wil said:

You’ll also be able to sink a lot more heat into the tank with an immersion compared to an ASHP

Why will it be a LOT more? Unless you know that Samsung are exagerating with their R290 heat pump performance - they are claiming 65 degC albeit with a low (but still a lot more than unity) COP? Genuinely interested here because something from this range is very high on my list especially if I can get the grant to fund it. Yes, immersion for anything above this, but I don't see there will be a pressing need if the HP manufacturer claims are genuine under real world conditions. 

 

On 14/02/2024 at 15:29, Wil said:

Depends if you care about zero grid consumption or self-utilisation of the PV.

Self-utilisation for me. I've got no plans for FIT and I will be paying for electricity to heat the house and the water anyway so I may as well multipy the PV generation by the COP to get best use of energy wherever possible. No point trying to not use power in the day time to find it gets used  later in the evening anyway. There's no cheaper energy source available so I can see little point for me in using a PV diverter. The modelling will tell all with this... Not had time write anything for this yet but will share the results when I'm done. Had some work deadlines so my progress on house tasks is slow this week.

 

On 14/02/2024 at 15:31, Wil said:

You’re only one leak away from inhibitor and the contents of your heating system sludge too in that case?

Yes, but iron oxide and inhibitor feels less toxic and easier to notice than glycol based products. Iinhibitor and iron oxide have always been a house feature so it's "normal" whereas glycol feels like something that should be in the garage slowly leaking out of my TVR so I could be convinced I am concerned over very little.

 

Having a think about the various designs, it looks like the simplest is trying to get a good COP by matching the heat pump size to the 97th (ish) percentile worst case outside temperature and rely on modulation then reduced duty cycle when there is less heating demand such that no buffering should be needed. On the worst case days DHW can only from immersion heaters and one would suspect there is no PV on such days... Simple to install and simple to run, but it's probably the more costly option to run because it appears to make the least effective use of PV.

 

Then the more complex designs try to break the link between supply of heat and use of heat in various ways. It would seem from this thread that some ways are more acceptable than others and whatever happens breaking this link will cost efficiency in some way, so it's critical to minimise the loss and know that the PV can provide a greater benefit.

[JohnMo]

7 hours ago, dnb said:

iron oxide and inhibitor feels less toxic and easier to notice than glycol based products

There are several grades of glycol, the stuff used in heating systems is food grade, it doesn't poison you. Stuff used in cars does.

[JohnMo]

On 13/02/2024 at 11:07, dnb said:

trapped between a 5kW heat pump that will do the heating well but struggle for time with DHW or a 12kW pump

Why are trapped into a 5 or 12kW heat pump, plenty of sizes in-between to choose from. Size so you have two to three hours for DHW and rest of the time to do house heating at your design temp. If you have your heat pump running 24/7 it will automatically utilise your PV, time DHW heating at around midday when most solar is available. You also need a PV diverter, heat pump heats cylinder to 50 and immersion takes it to 60.

 

Heat pump will have a minimum flow rate and volume of water that must be engaged at all times. If you have zones that do not meet both requirements, you need to make a system that does meet them. So now you need to read up on volumisers, buffers (2, 3 and 4 port) and close coupled tees (CCT). They can all be used to get what you want, with differing affects on efficiency.

 

The more simple the system the better the efficiency generally, heat pump direct to heating system (via a 3 port diverter), controlled by manufacturer controller is best.  No third party stuff to mess things up.

[dnb]

On 16/02/2024 at 08:39, JohnMo said:

the stuff used in heating systems is food grade

Thanks. That's new information to me. Suspect assumptions were made on both sides (I spent many years playing with classic cars so "knew" glycol was poisonous) so nobody bothered to talk about the COSHH data.

 

On 16/02/2024 at 08:59, JohnMo said:

trapped into a 5 or 12kW heat pump

They were the extreme examples. Nothing more. There's an 8kW in the range that might work if I can wait the delivery time. The 12kW does have some advantages if the efficency issues can be managed effectively but if not then there's no point.

 

On 16/02/2024 at 08:59, JohnMo said:

heat pump heats cylinder to 50 and immersion takes it to 60

So you completely discount the Samsung Gen7 R290 heat pump claims of 60 to 70 deg water? Very interested in knowing because it's high on my list.

 

On 16/02/2024 at 08:59, JohnMo said:

So now you need to read up on volumisers, buffers (2, 3 and 4 port) and close coupled tees (CCT)

Any suggestions about good sources that aren't sales pitches? General information from Wikepaedia et al. is OK but often doesn't have relevant examples. By what you've said, all of the candidate designs will work. Some will be more efficient than others and until I can model them effectively I am doing little more than rolling dice to decide what to buy.

 

On 16/02/2024 at 08:59, JohnMo said:

They can all be used to get what you want, with differing affects on efficiency.

This is what I am trying to tease out from here and my candidate installers. Thanks for the discussion so far.

[JohnMo]

1 hour ago, dnb said:

So you completely discount the Samsung Gen7 R290 heat pump claims of 60 to 70 deg water? Very interested in knowing because it's high on my list.

No don't discount anything. But why heat on a routine basis above 50 and take a kicking with rubbish CoP when the sun's not out and you have no PV to use. A PV diverter will power the immersion but be a CoP of one, but it's free energy when it would otherwise go back into the grid.

[Dan F]

Not read the whole thread yet, but PHE (plate heat exchanger) is only required if you use glycol in the system and you don't want glycol in UFH/radiatos. My understanding is you have thee options:

1) No Glycol, use anti-freeze valves instead.

2) Glycol externally, PHE, no glycol internally in radiators/UFH.  (potentially an additional pump that might not be needed otherwise too)

3) Whole system has glycol. No PHE required.

We were recommended 1).  Simpler, no (albeit small) efficiency losses and pure water is better for heat transfer.

 

PHE (instead of a coil) for heating UVC is different discussion to the use of PHE to seperate external glycol circuit from internal.

[sharpener]

On 13/02/2024 at 21:04, dnb said:

For the record, UFH on the ground in 3 zones, towel rails in the 4 bathrooms, FCU to put heat into the A/C ducting for upstairs and MVHR. It is all sized so that 40 deg C flow temperature works down to -1 deg C outside.

 

You need to keep the flow temp down to that level for best CoP. So if all flow temps are matched then heating not a big problem - if you can run it with all the UFH engaged.

 

Then you will not need a buffer or LLH, they both involve mixing hence an increase in entropy and so loss of thermodynamic efficiency, Heat Geek call it "distortion". You might not need a volumiser either if the UFH volume is enough for defrosting. There is a good discussion of all these arrangements on the Caleffi website, mostly they are to make the system more bullet-proof for the installer not to benefit the end-user. Avoiding call-backs is a big incentive for installers which is why they over-specify HPs and leave them set ridiculously high.

 

Much of the above discussion seems to be about DHW. My own planned installation has got a secondary circulation pump to improve heat transfer in an existing tank, but starting from scratch it should be easy so keep it simple.

 

As @JohnMo says first select yr size using some rule like 45l/day * {greater of [actual occupancy, # of bedrooms +1]}, specify 3 sqm coil, job done. (Newark will quote for custom configurations inc 2 x immersions and many other options. Ideal cylinders are identical to Gledhill but much cheaper. HP mfrs own cylinders are more expensive again.)

 

PHEs involve an extra temp drop typ. 5C, so much the same as a coil, useful for corner cases but generally offer no advantage over a good coil and will de-stratify which you do not want to happen.

 

I have recently looked at the trade-offs between storing energy as hot water and battery storage here. In summary thermal storage has a payback of 12 - 20 years in which time you are likely to have replaced batteries once or twice which makes them more expensive. In my case I have already got them, they are not big enough to run the HP entirely but will be useful to fill the gaps caused by the sun going in unexpectedly.

 

With your smaller HP this will be easier to do, get the smallest HP you can get away with as then batteries will go that much further and you will need a smaller inverter as well. And it will modulate to lower outputs in mild weather without cycling on and off.

 

IIRC there is an issue with Samsung HPs, they use some trick ?hot gas bypass? to improve turndown ratio but at the expense of efficiency.

 

[dnb]

On 19/02/2024 at 18:55, sharpener said:

IIRC there is an issue with Samsung HPs, they use some trick ?hot gas bypass? to improve turndown ratio but at the expense of efficiency.

Thanks for a very useful insight. You do recall correctly. This is why I am keen to understand if the gen7 HPs have this trick like their predecessors. It makes quite a difference to the plan.

 

Relevent to this, I found their datasheet quoting the MCS approved SCOP. For the 8kW and 12kW pump it looks like:

 

Temp 8kW  12kW
35°c 4.66 4.73
40°c 4.35 4.42
45°c 4.04 4.12
50°c 3.72 3.82
55°c 3.41 3.51
60°c 3.15 3.17
65°c 2.89 2.84

 

This is why I am considering if I really need the immersions - if the HP really can deliver a COP way better than an immersion in  most of the conditions I may as well not tie it to the solar. I currently see batteries to fill in the holes as still a lot more expensive than just buying electricity if I can get the heat pump, buffering and control system sized and timed correctly on average. This is obviously a tricky problem!

 

On 19/02/2024 at 18:55, sharpener said:

So if all flow temps are matched then heating not a big problem

I tend to agree here. The MVHR etc will help to even everything out in the house and I'm not setting out to control things by zone other than having a warm towel in the evening even if the heating is nominally off for the purpose of space heating. It's a case of them being there to get ticks in boxes. I don't see a need to deliberately not heat parts of the house when I've spent a lot of time and effort to make it well insulated and air tight.

 

DHW is the main part that's making me consider the plans carefully. The other is separating generating heat from using it may have some timing benefits from using my solar PV. Yes it's more energy used to do the same job, but when much of that energy is "free" and multiplied by a COP greater than 1 then it's a financial win. My annual solar generation should be in the 6.5MWh range so well worth considering.

 

On 19/02/2024 at 18:55, sharpener said:

first select yr size using some rule like 45l/day

I have the size. It's based on our useage data, multiplied by 5/3 to allow for long term parent visits. This calculation has never been part of the problem other than designers trying to make it smaller.

 

On 19/02/2024 at 18:55, sharpener said:

which is why they over-specify HPs and leave them set ridiculously high

I think I am seeing some of this in the candidate designs. It's dressed up as other things of course.

 

On 19/02/2024 at 18:55, sharpener said:

trade-offs between storing energy as hot water and battery storage

Very useful link. Thank you. Not had time to read much yet this week. Can't see where the week has gone!

Edited February 22 by dnb

[JohnMo]

5 hours ago, dnb said:

annual solar generation should be in the 6.5MWh range so well worth considering.

Also remember this is mostly summer generation, Dec and Jan will produce next to nothing, even if the sun is out the production day is very short. During most of Nov to Feb your heating will not come via anything your PV produces, it may offset a little.

 

If you are using a ToU tariff (E7 etc) you really need to utilise that period well. 

[dnb]

11 hours ago, JohnMo said:

Also remember this is mostly summer generation

Tell me something I don't know...  It's one of the reasons I don't see the need to "penny pinch" with a solar diverter and an immersion if I can meaningfully sort out a few hours of storage. I can just take the benefit (small as it is) from offsetting.

 

In the summer I will use the PV for the A/C as well as the DHW.

 

11 hours ago, JohnMo said:

f you are using a ToU tariff (E7 etc)

Again, this drives the need to sort out some form of meaningful storage that doesn't hit efficiency too badly.