Rethinking the system engineering for mass retrofit - ideas (probably provocative)

[JamesPa]

Over on the thread 'Rethinking the mindset for mass retrofit - a provocative idea' we have ended up discussing, and I would say largely (but of course not universally) agreeing on, regulatory changes that are necessary if we are to achieve the mass roll out of heat pumps that is necessary to achieve our climate change goals.  That discussion continues, but I thought a bit of a brainstorm of engineering ideas might be useful as well, given that we have quite a few tradesmen/engineers/former engineers here.

 

The essential background is that today we install about 1.6M gas boilers each year of which 1.4M are retrofits.  Each gas boiler installed is an opportunity lost.  In 2022 we installed about 60,000 heat pumps.  The UK is 20th on the European league table of heat pumps per person.  Currently heat pump retrofits cost about 10-15K (or even 20K) after BUS grant, this is clearly not going to achieve mass rollout when a gas boiler replacement is 4-5K. 

 

A part of the reason is that the replacement of a gas or oil boiler with a heat pump not infrequently comes with replacement of much of the rest of the heating system.  These include

 

• DHW tank
  • Cold feed to DHW tank (upgrade 15mm->22mm
  • Feeds from boiler to DHW tank (upgrade 22mm->28mm
• Emitters
  • Primary feeds from boiler to the emitters (upgrade 22mm->28mm)
  • Secondary feeds to individual emitters Uupgrade microbore -> 15mm?)
• Miscellaneous things like mag filters etc
• things that are anyway counterproductive like room thermostats (other than any in the HP controller itself), timers etc

 

and we add – buffer tanks

 

This all adds cost and disruption which are frequently out of proportion to the benefit perceived by the customer. In many (not all) cases they are not necessary.  Sometimes they are done unnecessarily because the whole-house sizing is over engineered, so we need to add this to the list of things for which we need an better engineering solution

 

In this thread Id like to suggest a discussion of system engineering amendments for: whole house sizing (particularly to avoid over-sizing), and the DHW tank.  The first because it has so many knock on effects, the second because its often the biggest single identifiable change other than the HP itself, and quite likely the one causing most disruption.  Im throwing in ideas here, which seem reasonable from an engineering perspective.  I don't claim to be an expert (because I'm not), so I may have missed something fundamental.  I do, however, claim to be capable of understanding logical arguments put by experts if they make them.

 

1.    Whole house sizing

 

This is fundamental.  Get it too low and the customer is cold, get it too high and the system performance is compromised, you may need an otherwise unnecessary buffer tank which is likely to compromise performance, the DHW and primary feeds may need upgrading unnecessarily, and it costs more. 

 

This is what we achieve using the current method (https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/606834/Report_on_compliance_with_MCS_installation_standards_v32.pdf),

 

ie quite frequently a 2 fold oversizing or worse, less frequently undersizing.  Its pretty poor.  Not surprising really, much of the fabric on which surveyors base their estimates is not visible, so they have to make assumptions which are unlikely to be accurate (particularly if they don't listen to the customer when he tells them how the house construction has evolved!)

 

So why dont we do the whole house sizing by simply measuring demand, or at least use demand as a sense check?  At its simplest its really, really, really simple to take annual gas (or oil) consumption and from that work out the load at any chosen design temperature, and thus the required peak output.   Obviously a few questions are required to determine how the customer uses the house, but at its root its real world data thats easily obtained and measures the right thing.

 

Now there are more sophisticated methods than this crude calculation, using smart meter readings (only applicable for gas of course) or making a specific measurement with sensors and calibrated heat sources, but I would suspect that the simple method, supported by a few questions that the householder has to answer, may well be good enough, and almost certainly better than we currently manage.  For my house this calculation comes out about 8.5kW, against a measured peak of 7.5kW at design OAT.  Meanwhile surveyors (two of them) got to 16kW (because they ignored fabric upgrades that I told them about and double counted room to room losses) and if I use MCS assumptions but put in the fabric upgrades, I get to 10.5kW.  

 

I reiterate - get this right (which currently, based on the evidence above, we dont) and lots of other problems disappear 

 

2.        DHW tank (and associated feeds/components)

 

There appear to be three principal reasons to replace this namely

 

a)    because ASHPs running at say 55 need a larger coil to get to say 48-50 without short cycling and with a reasonable recovery time

b)    because the installer 'decides' that the householder will ‘need’ a larger tank if its heated to a lower temperature

c)    because it makes the plumbing simpler if you use a ‘pre plumbed tank’ provided by the system supplier

 

I suspect (c) is often the real reason and, if it results in a net cost saving is justifiable.  However I haven’t seen evidence that it does.  So lets concentrate on the engineering reasons.

 

The householder should have a say in (b).  The installer should give advice, but ultimately it’s the householder’s decision.

 

There are at least two solutions to (a).  The first is what Mixergy do with their Heat Pump kit.  Essentially its to add a PHE in series with the coil and use a pump to pump the DHW through the PHE, thus increasing the effective size of the flow water -> DHW water heat exchanger.  Some downsides, but overall perfectly reasonable.  The second is to run the HP at 55 as normal for DHW heating, (or consider 60), heat the water to say 40-45 on the HP and then use the immersion (or better still a willis heater) to go the last mile.  Yes it will add to running cost a bit, about £120 per year based on the estimate I did.  At £120 per year most people would take this option instead of suffering the disruption and cost of swapping out the DHW tank and associated parts, because the business case for doing so sucks.

 

Upgrading the CW feed to the DHW tank is only necessary if the customer needs (and wishes to pay for) a larger flow, otherwise it isn’t.  So it can be offered as an upgrade to the system not a ‘must do’

 

Upgrading the feeds from boiler to DHW tank (22mm-28mm)  is usually unnecessary so far as I can see.  22mm feeds are capable of 6kW, 8 at a push, at DT=5.  That’s plenty for most (an immersion is 3kW and many get by on that).  The only other problem is if your HP wont modulate down far enough.  However if the whole house sizing is correct in the first place, and it’s a half decent HP, that’s unlikely to be the case unless you have a very large/very leaky house (in which case, sorry, you are going to have to pay the premium).

 

Comments please

 

There are still other elements to discuss, but if we try to cram too much into one thread we wont get anywhere.  And for the avoidance of doubt it is clear that some of this requires more nuanced system engineering which is not going to happen given the fact that demand today exceeds supply.   That latter problem is the subject of the original thread (or at least where its ended up.  And please bear in mind that the requirement is 1.4M retrofits per year at a price people can afford.

 

 

 

[JohnMo]

2 hours ago, JamesPa said:

Upgrading the feeds from boiler to DHW tank (22mm-28mm)  is usually unnecessary so far as I can see.  22mm feeds are capable of 6kW

My heat pump install (6kW) is not next to house so the pipe run up to the manifold for the UFH about 10m (28mm copper and Hep2O), to get to the cylinder is another 8m each way.  The piping I want to use from 3 way valve to the cylinder is already pre installed and it 22mm Hep2O, so the pressure drop is way to high.  My solution is to add a pump to the return leg from the cylinder.  The 3-way valve and additional pump will be supplied from the same supply via the heat pump.  When DHW calls for heat, the 3 way valve is powered so will the pump.  Heat pump demands its own pump runs at 100% speed at all times during DHW heating, so I should be able to fine tune the running speed of the additional pump without it all chasing it's tale.

Edited April 30, 2023 by JohnMo
Missed some context

[JamesPa]

3 minutes ago, JohnMo said:

My heat pump install (6kW) is not next to house so the pipe run up to the manifold for the UFH about 10m, to get to the cylinder is another 8m each way.  The piping I want to use for the cylinder is already pre installed and it 22mm Hep2O, so the pressure drop is way to high.  My solution is to add a pump to the return leg from the cylinder.  The 3-way valve and additional pump will be supplied from the same supply via the heat pump.  When DHW calls for heat, the 3 way valve is powered so will the pump.  Heat pump demands its own pump runs at 100% speed at all times during DHW heating, so I should be able to fine tune the running speed of the additional pump without it all chasing it's tale.

Thats one I hadn't thought of and that's system engineering.  There are multiple solutions to the problems encountered in retrofits, but the current HP installation methodology seems to confine us to a few.  The 'right' one is one which works technically, works for the customer and works financially.

[sharpener]

2 hours ago, JamesPa said:

2.        DHW tank (and associated feeds/components)

 

There appear to be three principal reasons to replace this namely

 

a)    because ASHPs running at say 55 need a larger coil to get to say 48-50 without short cycling and with a reasonable recovery time

b)    because the installer 'decides' that the householder will ‘need’ a larger tank if its heated to a lower temperature

c)    because it makes the plumbing simpler if you use a ‘pre plumbed tank’ provided by the system supplier

 

I was encouraged that the Good Energy online enquiry form gives the choice of opting out of the tank replacement, with a warning that recovery times may not be what they used to be.

 

re a) you wrote <There are at least two solutions.  The first is what Mixergy do with their Heat Pump kit.  Essentially its to add a PHE in series with the coil and use a pump to pump the DHW through the PHE, thus increasing the effective size of the flow water -> DHW water heat exchanger.  >

 

I also proposed using a pump on its own to stir the tank and so increase the effective coil area by improving the heat transfer from the outside of the coil over that achieved by convection alone. Can't now find the link to this discussion. Still needs the potable water pump, but less plumbing involved and no PHE required so significantly cheaper.

 

It would be nice to find some way of determining how much  benefit this might bring, my feeling is that it might be 2x. Downside is no stratification so you would need to re-heat the whole tankful.

 

Personally I would heat the water last thing before E7 ends in the morning. This might not suit people who have lots of children/baths in the evening but what I am suggesting would work no less well at teatime and combine ideally with PV.

 

[JamesPa]

5 minutes ago, sharpener said:

I was encouraged that the Good Energy online enquiry form gives the choice of opting out of the tank replacement, with a warning that recovery times may not be what they used to be.

 

That is encouraging and the first time I had heard of this being an option in a quote.  The industry mantra, as I have experienced it, is 'you need a big coil for recovery time and to stop short cycling'.  Obviously GE dont use pre-plumbed cylinders in their installations so dont have the motivation others have to find a reason to recommend cylinder replacement.

 

7 minutes ago, sharpener said:

I also proposed using a pump on its own to stir the tank and so increase the effective coil area by improving the heat transfer from the outside of the coil over that achieved by convection alone. Can't now find the link to this discussion. Still needs the potable water pump, but less plumbing involved and no PHE required so significantly cheaper.

 

It would be nice to find some way of determining how much  benefit this might bring, my feeling is that it might be 2x. Downside is no stratification so you would need to re-heat the whole tankful.

You did indeed and thanks for reminding us. 

 

My worry about this one is that Im not sure it entirely solves the short cycling problem.  But I grant it might, and of course the seriousness of the short cycling problem is dependent on several parameters, some of which are field-adjustable.  The pipework required for all three solutions is very similar so if one didn't work you could convert to the other.  Obviously that's not a satisfactory way for a third party installer, they need a solution which works.

[JohnMo]

My cylinder is a 160L thermal store, already fitted with a PHE, on the heating side in addition to the really small coil designed for 80 degree flow.

 

It is currently only heated as buffer from UFH flow to temper water going to a combi boiler in winter via the DHW coil. And heated in summer with PV.  My intension is to heat to 50 (or the max temp I can get) using ASHP between 10 and 11 using excess solar (more efficient than immersion) then let excess solar drive the immersion to whatever temp I can get.

 

I am leaving the almost new combi in place for now, but it will not be hooked to the UFH.

[JamesPa]

Thats interesting, any chance of a diagram.  Is it intended that it will still work with ASHP? Personally I am not so keen on a large store of mains pressure hot water but had the impression that getting a thermal store to work with an ASHP to heat DHW real time was going to be tricky.

 

James

Edited April 30, 2023 by JamesPa

[sharpener]

1 hour ago, JamesPa said:

My worry about this one is that Im not sure it entirely solves the short cycling problem. 

 

If it were to improve the heat transfer by 2x then this would double the output the HP would run at for the same delta T from LWT to DHW. I wish I could devise a simple experiment though.

[JohnMo]

1 hour ago, JamesPa said:

Thats interesting, any chance of a diagram.  Is it intended that it will still work with ASHP? Personally I am not so keen on a large store of mains pressure hot water but had the impression that getting a thermal store to work with an ASHP to heat DHW real time was going to be tricky.

 

James

 

Sorry not the best sketch, top sketch is hot water flow - if below 45 all water flows to combi. If above the the water goes via a mixer to the taps.

 

The heating sketch, will now be cylinder heating, the UFH heating pipes are now deleted, and the combi will become an ASHP.

 

[JohnMo]

2 hours ago, JamesPa said:

impression that getting a thermal store to work with an ASHP to heat DHW real time was going to be tricky.

I think you may be correct, but I have it, so I will try to see what I can do with it.

[dpmiller]

TS works fine for us with the ASHP but sensor pocket position and hysteresis setting is critical, as well as minimising the time it takes for the HP to ramp up towards full power. Which means plenty of flow and the flow modulated by the HP...

[JohnMo]

4 minutes ago, dpmiller said:

sensor pocket position

Where did you find best? And what hysterisis are you using?

[dpmiller]

48 minutes ago, JohnMo said:

Where did you find best? And what hysterisis are you using?

 

I've two pockets but they aren't really far enough apart. One just above the coil, about 2/5 up, the other halfway. I'm using the upper one.

Hysteresis is 1.5c plus an overshoot of 0.5 (overshoot give the inverter a chance to ramp down rather than turning off)

Tank setpoint varies by time but is between 50 and 54, but the top of the tank tends to end up 3c or so above setpoint

[JamesPa]

5 minutes ago, dpmiller said:

 

I've two pockets but they aren't really far enough apart. One just above the coil, about 2/5 up, the other halfway. I'm using the upper one.

Hysteresis is 1.5c plus an overshoot of 0.5 (overshoot give the inverter a chance to ramp down rather than turning off)

Tank setpoint varies by time but is between 50 and 54, but the top of the tank tends to end up 3c or so above setpoint

This is very interesting,  Do you have it set up so the HP tops it up whenever it needs it and stops heating the house, what flow temp are your running the HP at, and whats the size of the 'heat pump coil' (or whatever transfers energy from the HP to the thermal store). 

 

You seem to be pretty unique in using a TS with a HP.  Its installed by an MCS contractor of course (just joking, but do say if it was).  Was it just a regulat TS off the shelf or something special.  Basically, whats the setup?

Edited April 30, 2023 by JamesPa

[dpmiller]

1 minute ago, JamesPa said:

This is very interesting,  Do you have it set up so the HP tops it up whenever it needs it and stops heating the house, what flow temp are your running the HP at, and whats the size of the 'heat pump coil' (or whatever transfers energy from the HP to the thermal store). 

 

You seem to be pretty unique in using a TS with a HP.  Its installed by an MCS contractor of course (just joking, but do say if it was).  Was it just a regulat TS off the shelf or something special.  Basically, whats the setup?

Yes, DHW priority like any other ASHP

flow temp during DHW is controlled by the ASHP using it's own deltaT curve

It's a World Heat TS with a standard 3m2 coil and two immersions both on PV via an iBoost

but...

All heating flow also passes through the TS so the TS acts as a buffer volume. Heating flow (UFH *and* rads) is about 42c, the tank is never cold...

And this means I can heat the whole house off the immersion heater *or* the boiler stove, as a failsafe

[JamesPa]

10 hours ago, dpmiller said:

Yes, DHW priority like any other ASHP

flow temp during DHW is controlled by the ASHP using it's own deltaT curve

It's a World Heat TS with a standard 3m2 coil and two immersions both on PV via an iBoost

but...

All heating flow also passes through the TS so the TS acts as a buffer volume. Heating flow (UFH *and* rads) is about 42c, the tank is never cold...

And this means I can heat the whole house off the immersion heater *or* the boiler stove, as a failsafe

Thanks.  So if I understand it correctly (which almost certainly I don't!) its plumbed for the CH as a (4 port?) stratified buffer tank (two taps at top two at bottom, or all towards the bottom?) which runs (without WC ?) at FT=42 and with the DHW heat exchanger at (or tapped off from) the top of the tank?

[dpmiller]

TS has a 3m2 ASHP coil in the bottom third and a DHW coil in the upper third, roughly.

Flow from the ASHP passes through the bottom coil at all times except if cooling is required, when it's bypassed using a 3-port valve. I'd originally done this to enable blending of HP and buffer volume in parallel but that's never actually needed. 2-port zone valves divert the flow out of the coil to the manifolds and if neither manifold is calling, an auto bypass valve sends all flow to return

 

The boiler stove is plumbed direct to the TS (through a laddomat-style anticondensation valve.

[JamesPa]

2 hours ago, dpmiller said:

TS has a 3m2 ASHP coil in the bottom third and a DHW coil in the upper third, roughly.

Flow from the ASHP passes through the bottom coil at all times except if cooling is required, when it's bypassed using a 3-port valve. I'd originally done this to enable blending of HP and buffer volume in parallel but that's never actually needed. 2-port zone valves divert the flow out of the coil to the manifolds and if neither manifold is calling, an auto bypass valve sends all flow to return

 

The boiler stove is plumbed direct to the TS (through a laddomat-style anticondensation valve.

Makes sense.  Am I to deduce that the flow temp of the CH never goes below the tank temp, or is it the case that when it does you divert out of the coil.

[dpmiller]

the base of the tank sits around or just below flow temperature.

[sharpener]

18 hours ago, sharpener said:

 

If it were to improve the heat transfer by 2x then this would double the output the HP would run at for the same delta T from LWT to DHW. I wish I could devise a simple experiment though.

 

Have posted this as a query on another forum. Most of the replies are off topic/unhelpful as per usual but one has proposed an experiment. I will report back to both fora and hopefully before too long (once I have reclaimed a length of 22mm Cu pipe from being a dinghy trailer axle. 1/2 in galv steel pipe (which I have) will be a better fit for the new wheel OH has just bought).

 

BTW I have just viewed this Mixergy video and at 1'50" Pete Armstrong the CEO says heat pump conversion retains all the features and benefits. But I see from the Mixergy User Guide that the heat pump conversion kit almost completely destratifies the tank (presumably a result of pumping it all round throught the external PHE). So it is hard to believe that the principle of heating only the water you want to use etc is maintained with a Mixergy/HP combo, the fancy user interface app becomes largely pointless and you are no worse off fitting the pump to a conventional tank, with or without the PHE.

 

Edited May 1, 2023 by sharpener
Added Mixergy para

[JamesPa]

3 hours ago, sharpener said:

 

Have posted this as a query on another forum. Most of the replies are off topic/unhelpful as per usual but one has proposed an experiment. I will report back to both fora and hopefully before too long (once I have reclaimed a length of 22mm Cu pipe from being a dinghy trailer axle. 1/2 in galv steel pipe (which I have) will be a better fit for the new wheel OH has just bought).

 

BTW I have just viewed this Mixergy video and at 1'50" Pete Armstrong the CEO says heat pump conversion retains all the features and benefits. But I see from the Mixergy User Guide that the heat pump conversion kit almost completely destratifies the tank (presumably a result of pumping it all round throught the external PHE). So it is hard to believe that the principle of heating only the water you want to use etc is maintained with a Mixergy/HP combo, the fancy user interface app becomes largely pointless and you are no worse off fitting the pump to a conventional tank, with or without the PHE.

 

I read it the same way.  The interest for me was the principle not the application to Mixergy tanks.

[sharpener]

36 minutes ago, JamesPa said:

Have posted this as a query on another forum. Most of the replies are off topic/unhelpful as per usual but one has proposed an experiment. I will report back to both fora and hopefully before too long

 

Here is the experiment as proposed by jrmichler here

Or you could run a simple test. Get a piece of copper tube about the diameter of the tubing in the heat transfer coil, and solder a cap on one end. Drop the sensor from an indoor/outdoor thermometer into the tube and fill it with water. Fill a bucket with hot water. Insert the tube into the water, and measure the rate of temperature rise in the tube. Do this while holding the tube still, then while swishing it back and forth. This test measures the rate of heat transfer of the total system. It also gives you a good idea of how fast you would need to circulate the water in order to get the water filled tube to heat up 3 to 4 times faster.

 

OK so the result is useful. Without any agitation the initial rate of rise inside the tube is about 6C/min and with agitation it is about 10. Without agitation the temp is still rising slowly after 10 mins and with agitation it reaches a plateau at 7. However while the swishing motion is effective at mixing the water in the bucket it is not clear to me if it is effective at mixing the water in the comparatively narrow tube.

 

Conclusion: the heat transfer is improved by a factor of at least 1.5 and it might be more if the regime on the inside of the tube were continuously flowing as in the real-life situation.

 

This would make the difference between keeping a 6kW HP from short-cycling and a 10 kW ditto so I think a worthwhile improvement for the cost of the pump.

 

@JamesPa you wrote <I read it the same way.>

 

Not sure what your "it" is, the most important things for me were that many of the claimed advantages for the expensive Mixergy tank disappear when you connect a heat pump to it and their CEO is less than truthful about this on youtube.

[dpmiller]

^ don't discount how strong and turbulent the convection currents will be inside the tank with the heat rising off the coils...

[sharpener]

32 minutes ago, dpmiller said:

^ don't discount how strong and turbulent the convection currents will be inside the tank with the heat rising off the coils...

 

I think if they were either strong or turbulent in the first place then you would never get stratification.

[SteamyTea]

8 minutes ago, sharpener said:

stratification

Not the best way to describe a temperature gradient.

Think it took about an hour for my cylinder to 'settle down' after heating.