To Zone or Not to Zone When Upgrading a Wet Heating System from Gas to Heat Pump for Maximum Climate Impact: A UK View

[DamonHD]

I had a paper published: https://www.mdpi.com/2071-1050/16/11/4710

 

TL;DR: when you are replacing a gas boiler for radiators with a heat pump such as an ASHP in a typical UK home and using weather compensation for the system, consider retaining TRVs in areas such as bedrooms, sunrooms or others with variable incidental gains, and also in low-occupancy rooms, as this should be a cheap and easy way to save more energy and improve comfort.  You will still need to ensure sufficient flow and volume for the heat pump.

[SteamyTea]

Read about a third of it.

Good feeling getting things published isn't it.

[DamonHD]

Yes!

[Dave Jones]

On 07/07/2024 at 12:47, DamonHD said:

I had a paper published: https://www.mdpi.com/2071-1050/16/11/4710

 

TL;DR: when you are replacing a gas boiler for radiators with a heat pump such as an ASHP in a typical UK home and using weather compensation for the system, consider retaining TRVs in areas such as bedrooms, sunrooms or others with variable incidental gains, and also in low-occupancy rooms, as this should be a cheap and easy way to save more energy and improve comfort.  You will still need to ensure sufficient flow and volume for the heat pump.

 

this does seem at odds with the heat geek guys advice. Creating differential temperatures in the same slab equals losses so greater input is needed than otherwise would be the case.

 

Can see if you are using different emitters, some UFH some rad etc fair enough. Bit like having 2 TRV's on 1 rad.

[sharpener]

Following this with interest. @DamonHD you mention radiator (over) sizing as a complicating factor but I did not see any conclusion about the magnitude or sign of this effect, can you enlarge on this point for me?

 

I would love to see the work extended to the savings (or otherwise) of whole-house night setback. @JamesPa has done some modelling of this which IIRC showed real but modest savings from the heat losses being somewhat lower. Again there is a "bad setback" effect if the HP has to work hard to re-attain the daytime temps in a reasonable time. But I am guessing this may be reduced or eliminated if it happens only inside a cheap electricity window.

 

We have a two storey long thin 180 sq m barn conversion with various unusual occupancy patterns. In particular the middle two bedrooms are used mainly as studies in the daytime, and the fourth only when we have guests. But the living room (which occupies half the ground floor) is only used in the evenings.

 

There are Honeywell wireless TRVs on nearly all the radiators and I am considering how best to incorporate these into the control strategy for the 12kW HP which is shortly to be installed.

 

Initial approach will be to simply use them for overheat prevention while I get the weather comp optimised first. After that it may depend on my patience for carrying out meaningful and detailed experiments!

 

[DamonHD]

A key point which only became clear to me from a throw-away remark from one of the peer reviewers: the bad setback effect is forced by having rads only just large enough for the no-setback state.  This suggests always at least slightly oversizing rads if you can.

Edited July 15 by DamonHD

[JohnMo]

48 minutes ago, DamonHD said:

A key point which only became clear to me from a throw-away remark from one of the peer reviewers: the bad setback effect is forced by having rads only just large enough for the no-setback state.  This suggests always at least slightly oversizing rads if you can.

So question is do you need a higher flow temp to allow you play catch-up, once out of setback? if so how is this managed? Also do you allow in the calculations the heat pump running a full or increased load for a few hours to to recover from a set back.

 

Obvious question is why would you bother, with the added complexity, cost to install other stuff and then have to fine tune for steady state, setback and recovery from steady state.  

[sharpener]

40 minutes ago, DamonHD said:

the bad setback effect is forced by having rads only just large enough for the no-setback state

 

Ah, so ideally you need rads which are big enough to cope also with the heat leaks to the setback rooms. Yes we are getting 6 rads changed for bigger ones, required uplift factor 1.9x. In one of the study/bedrooms this is not quite achievable bc of furniture but fitting foil and having it plumbed TBOE will help a bit.

 

Your paper concludes Using common open-loop weather compensation (radiator flow temperature driven by external temperature only) eliminates the bad setback effect, and indeed saves a little extra energy. This is in return for a small sag in temperature for A rooms, though likely well within tolerable bounds.

 

But this doesn't really eliminate the bad setback effect. If the "small sag in temperature" is "tolerable" then you could turn the setpoint/WC slope down by a corresponding amount all the time and make even greater savings!

 

To respond to what @JohnMo has just posted, we have already got the wireless TRV setup so it is a question of optimising how to use it. I certainly wouldn't be fitting it from new when installing the heat pump. But it does allow flexible zoning by time and temperature and this will I think be useful.

 

As a specific example I envisage using the UFH to heat the living room 18/24 to a setback temp of say 16C, there are also rads which will quickly bring it up to 20C at 1700 using off-peak heat from the thermal store. The ceiling is very leaky so this will provide the two bedrooms above with background heat as well. Only when they are in use for guests will they be heated to 18C, I cannot think that heating them all the time will save money.

 

Verifying all this experimentally will be very complex when I also pose the question

Will Octopus Cosy with its 3 periods at 10p/kWh be better than Octopus Intelligent Go with a single 6 hours at 7p plus whatever is achievable by gaming their algorithm by playing with the EV charging times?

[joth]

Congratulations on the publication.

The main comment I have is same as I made on LinkedIn a few weeks ago: without also modelling the effect that increased short-cycling on ASHP COP, I don't think a firm conclusion can be drawn either way on this zoning debate. 

[DamonHD]

I have my research plan / confirmation report open right now, and I am putting "the effect that increased short-cycling on ASHP COP" in it right now...  B^>

[DamonHD]

@sharpener"You could keep turning it down" is a bit of a straw man.  😀  My point here is mainly about tolerance (dynamic and static) of deviations from a selected preferred temperature.  Yes, indeed, if you are too hot all the time then move your setpoint down.  However, not all users of the same home have the same preferred setpoint or tolerance just to complicate issues, and each of us have our tolerance affected by things such as external weather, time of day, recent activity levels, ...

 

I note that my daughter wants a set point far higher than I do, at a point where I am beginning to feel uncomfortably warm sometimes.  And she is dressed sensibly!  So our rooms have different set points: hard to acheive without zoning.

Edited July 15 by DamonHD

[DamonHD]

3 hours ago, JohnMo said:

So question is do you need a higher flow temp to allow you play catch-up, once out of setback? if so how is this managed? Also do you allow in the calculations the heat pump running a full or increased load for a few hours to to recover from a set back.

 

Obvious question is why would you bother, with the added complexity, cost to install other stuff and then have to fine tune for steady state, setback and recovery from steady state.  

My model is purposely as simple as I could make it and is steady state, eg no thermal capacity, losses other than conductive, incidental gains, etc...

[JohnMo]

1 hour ago, DamonHD said:

My model is purposely as simple as I could make it and is steady state, eg no thermal capacity, losses other than conductive, incidental gains, etc...

How do you achieve a fully considered conclusion?

[sharpener]

1 hour ago, DamonHD said:

@sharpener"You could keep turning it down" is a bit of a straw man.  😀  My point here is mainly about tolerance (dynamic and static) of deviations from a selected preferred temperature.  Yes, indeed, if you are too hot all the time then move your setpoint down.  However, not all users of the same home have the same preferred setpoint or tolerance just to complicate issues, and each of us have our tolerance affected by things such as external weather, time of day, recent activity levels, ...

 

Well IMHO the scenario where you allow the temp in the A (not set back) rooms to sag proves nothing. Sensible people will (for any specific room) already have chosen the lowest setpoint they are comfortable with, bearing in mind likely activity and clothing level. This by definition means they will detect sag and so not tolerate it. Certainly in the detached house example I would not be happy with a sag of 1.9K

 

I would argue that the only rigorous comparison is where you have "stiff" temp control of the A rooms. Your model shows that in this scenario the "bad setback" highlighted by Heat Geek does in fact occur.

 

The key point is the relationship between CoP and flow temp, you say "Raising the flow temperature reduces the CoP (coefficient of performance) of the heat pump (with data points from a real device) by a greater factor than the heat demand reduction caused by the TRVs, thus the electricity demand of the heat pump goes up while the two rooms are set back."

 

I would be interested (since you do not say) what the "real device" was, whether it has R32 or R290 or some other refrigerant, and how sensitive the result is to choice of heat pump manufacturer, can you comment on this?

 

That is all for yr steady-state analysis. With whole house night set back the likely potential savings are of the order of 8% as discussed a bit here. But as @JohnMo says, having to play catchup moves the HP onto an even worse point on its CoP diagram than required just to combat the "bad setback", this will be a bigger effect but as Michael Podesta  has demonstated the overall result is, "Well, it depends, but it’s marginal, either way".

 

The only getout I can see is to do it on cheap rate electricity and this will I hope give a net positive result as the savings are partly at normal rate but the boost phase will all be cheap.

 

[HughF]

Having to play catchup only affects COP if you have load compensation adjusting your LWT upwards to speed the recovery... If the recovery method is as simple as 'room is below setpoint, turn on heat pump' then the LWT will sit at the same point on the WC curve as if the HP was already running. In fact, it might end up running at a greater COP because it isn't cycling, instead running into a cold start scenario.

[sharpener]

In the article linked to above Michael Podesta shows that without a boost phase you might save 12% but the recovery time is impossibly long.

 

 

He suggests that you need 6kW boost power (where the steady-state is 3kW) to get a recovery time of 3 hrs and this reduces the savings to 9%. Also if the CoP falls from 3.0 to 2.5 then the savings are more than wiped out. Clearly if the OAT is such that you do not have 100% overcapacity to play with you cannot do this, and in the limit you dare not have any setback at all or you can never recover. Otherwise you will have to choose which rooms you can heat at any one time.

 

 

[dpmiller]

control on return temp is another option...

[sharpener]

2 hours ago, dpmiller said:

control on return temp is another option...

 

How would this help?

[dpmiller]

6 hours ago, sharpener said:

 

How would this help?

it's my perception that it's sensing the system demand more accurately,  giving a bit more boost on heatup.

[DamonHD]

On 15/07/2024 at 15:47, sharpener said:

I would be interested (since you do not say) what the "real device" was, whether it has R32 or R290 or some other refrigerant, and how sensitive the result is to choice of heat pump manufacturer, can you comment on this?

 

I am using the numbers from the original Heat Geek page / pump, a Mitsubishi Ecodan.

[JamesPa]

On 15/07/2024 at 15:56, HughF said:

Having to play catchup only affects COP if you have load compensation adjusting your LWT upwards to speed the recovery... If the recovery method is as simple as 'room is below setpoint, turn on heat pump' then the LWT will sit at the same point on the WC curve as if the HP was already running. In fact, it might end up running at a greater COP because it isn't cycling, instead running into a cold start scenario.

One of the things I realised when playing around with the equations for recovery from whole house setback, is that it happens faster than you might think even if the lwt is not increased above the weather compensation baseline.  This is because (without changing the lwt):

 

a. the delta t between emitter and room is higher than the design (because the room is colder) thus the emitter emits more energy at the same lwt

 

b. the delta t between room and exterior is lower than design (because the room is colder), thus the room loses less energy to the outside.

 

These effects combine and the result is an asymptotic convergence towards the design temperature, with the ramp up starting quickly but of course slowing down.  The equation for this is in the form T = Td - ae^(-kt) where a and k are constants T is room temperature, t is time and Td the design temperature.  This is very different (and much more favourable) to a simple linear convergence model based on the (false) assumption that the power output of the emitter doesn't change. 

 

Of course the consequence of an  asymptotic recovery is that it never actually reaches the design temperature unless the emitters are at least microscopically oversized.  I'm struggling to decide what reasonable assumption to make to continue the analysis.   Also I haven't had the time to think about this in the past few weeks, and must confess I have lost motivation a little now that my local planning authority has formalised it's policy of requiring heat pumps to have a noise pressure of only 20dBA at the most affected neighbour which is essentially impossible.

 

I do intend return to this at some point but perhaps @DamonHD will beat me to it!

Edited July 28 by JamesPa

[JohnMo]

9 minutes ago, JamesPa said:

the delta t between emitter and room is higher than the design (because the room is colder) thus the emitter emits more energy at the same lwt

This is true but you need to look at what occurs at the heat pump also. The dT between flow and return will increase. The ASHP will try to maintain flow target and dT.

 

If the target dT is 5, and you start removing more heat to play catch up, the dT could move to 6 or 7, the ASHP trying to manage dT will allow the leaving water temp to fall below target. The ASHP ramps up power to full power to try and get to target flow temperature.

 

Question - Can it actually deliver the required power to catch up? Possibly on anything other than design outside temp. But likely to run a full load until it gets back to equilibrium. Will you take a hit in the wallet - most probably.

 

Do you need a buffer to if messing with some room on set backs, maybe. Will it hit your wallet with an uplift of flow temp to have a buffer - probably.

 

I ran my heat pump last year deliberately on this basis, with UFH my set back was 0.5 degs, because heat loss was slow (well insulated and thick screed). I did this way to make use of cheap rate electric. Looking back I used loads of cheap electric. This year it will be straight WC no set backs, the heat pump left to tick away on WC.

 

Last year in summer I ran the ASHP in cooling mode it ran for 3 hrs straight in the morning and the same after DHW in the afternoon, it spent the rest of the time off. Used about 4 to 6kWh a day (free from PV).

 

Made some changes now have had the heat pump running continuously for 10 days, setting is basically if the return temp drops below 19.6 degs, the heat pump compressor starts. It runs 10 to 15mins and the depending on sun and outside temperature the compressor could stay off 30 mins or 3 hrs. Using about 2 kWh a day. Getting a cooler more comfortable house using a between 1/2 and a 1/3 of the electric.

 

Slow and steady well setup is nearly always running at low loads.

[JamesPa]

On 28/07/2024 at 21:42, JohnMo said:

If the target dT is 5, and you start removing more heat to play catch up, the dT could move to 6 or 7, the ASHP trying to manage dT will allow the leaving water temp to fall below target. The ASHP ramps up power to full power to try and get to target flow temperature.

I'm not sure why it would do the first unless it had run out of power headroom (in which case then yes I agree, but most of the time we are operating below max power (because the oat is above the design value) so there is likely to be headroom in most practical circumstances). 

 

On 28/07/2024 at 21:42, JohnMo said:

Possibly on anything other than design outside temp. But likely to run a full load until it gets back to equilibrium. Will you take a hit in the wallet - most probably.

Mitsubishi publishes cop at various loads, it doesn't vary much for any given lwt/oat combo, so I can't see that running at max or thereabouts power materially harms efficiency, provided that lwt isn't increased.  Of course you do supply more energy, but thats just making up for energy not supplied during setback.  Obviously if oat is lower or let higher then this will cause a penalty.   The whole point of the analysis I started on but didn't yet finish is to quantify this.  There are some truly wild claims for the energy saving through whole house setback which prima face violate the laws of thermodynamics.  Equally there are statements floating around that there is no saving.  Experimental measurement is almost impossible without a controlled lab environment.  Crude models, both theoretical and numerical, suggest modest savings order 5% are likely but not certain (depends on precise conditions), I was hoping to do (and made a start on) a next level theoretical model, with a hope that it might be possible at least to establish some quantitative parameters for when setback becomes counter productive.  But as I say above I got distracted.

 

On 28/07/2024 at 21:42, JohnMo said:

Do you need a buffer to if messing with some room on set backs, maybe. Will it hit your wallet with an uplift of flow temp to have a buffer - probably.

I assume this is directed at op.  I have been thinking about whole house setback where the issue doesn't arise (but I agree with you in the case of room by room setback).

 

On 28/07/2024 at 21:42, JohnMo said:

I ran my heat pump last year deliberately on this basis, with UFH my set back was 0.5 degs, because heat loss was slow (well insulated and thick screed). I did this way to make use of cheap rate electric. Looking back I used loads of cheap electric. This year it will be straight WC no set backs, the heat pump left to tick away on WC.

Interesting, did the cheap electric not work out or is the aim to compare the two modes before settling on one?

 

 

 

[JohnMo]

18 minutes ago, JamesPa said:

Interesting did the cheap electric not work out or is the aim to compare the two modes before settling on one?

With battery, some extra PV and E7, when compared to gas heating, and no battery, I was a lot cheaper. But the kWh used was way more than it should have been.

 

So would like to run as it should be, low and slow, single zone, zero controls, straight WC. See how it goes.

[JamesPa]

8 hours ago, JohnMo said:

With battery, some extra PV and E7, when compared to gas heating, and no battery, I was a lot cheaper. But the kWh used was way more than it should have been.

 

So would like to run as it should be, low and slow, single zone, zero controls, straight WC. See how it goes.

Im guessing cheap electric Is mostly nighttime when it's cold outside.  That presumably accounts for high usage, plus of course the need for higher lwt.  Double whammy, although if it's cheap it's probably also green, so arguably not bad!  I look forward to hearing about the comparison.