How to predict heat pump size from your EPC

[sharpener]

FWIW I received this comment from @mk1_manyesterday:

 

"Air2heat highly rate Panasonic units especially the new R290 versions."

 

They operate an MCS umbrella scheme and are in Shropshire.

 

 

[JohnMo]

On 18/04/2024 at 21:07, sharpener said:

Michael Podesta's blog

Think your main reference accepted the rule of thumb wasn't that good and produced a spreadsheet.

 

https://protonsforbreakfast.wordpress.com/2023/04/02/another-heat-pump-spreadsheet-beyond-the-rule-of-thumb/

[JohnMo]

15 minutes ago, sharpener said:

Air2heat highly rate Panasonic units especially the new R290 versions

Not that keen as they seem to force you have a hydraulic split unit equipped for two zones and DHW with freeze valves, multiple indoor pumps and mixer valves etc.  So the whole thing is tied in to one manufacturer (outdoor unit and an indoor unit with or without a small hot cylinder - like it or not.  Easy for an installer, no flexibility for end user.

[sharpener]

1 hour ago, JohnMo said:

Easy for an installer...

 

That's probably why they "highly rate" them! Also as an umbrella provider it de-risks the relationship with the actual installer.

[JamesPa]

3 hours ago, MrPotts said:

Sorry, I meant 12,500kWh

A few questions

• is the house warm enough?
• Whats your heating pattern?
• is it a gas boiler, how old
• What does your EPC say your consumption is
• In the spreadsheet calculation, what are the assumptions about the fabric that you cant verify, particularly air changes per hour?
• Where in the country are you/whats the local design temperature

Here are some facts about my house for you to compare.

 

I use 18-20MWh/year and have monitored my smart meter readings for 2 years.  I have a gas boiler which does space heating and DHW, and a gas hob.  Never does the gas consumption go above 8.5kW (and this only happens for a couple of hours if I have turned the heating off at night, or if Im using all the gas rings) and most of the time its around 4kW.  The house is warm and heated 24x7 in the very cold season, about 6am-11pm in the shoulder season.  In the cold patch in late 2022, when it was -2 day and night for about 5 days, the consumption (I was watching it) was 7.5kW.  Most of the time the smart meter reads 4kW. 

 

Its clear therefore that the actual demand at the design temp of -2 is about 7.5kW-8kW and I need a heat pump which does this but will modulate comfortably down to 4kW or preferably a bit less (unlike my hopelessly oversized (28kW) gas boiler which cycles a lot).  To prove this to myself, this year I turned the gas boiler down to as low as it will go, 8.5kW, and it was never cold.  Granted it was a fairly mild year, but note the measurements the previous year.

 

The MCS spreadsheet gets to 10.5kW.  Thats with default air change values and my best estimates of the construction (its a 1930s house with various fabric upgrades and some extension, so a very mixed fabric).  However I can get it to 7.5kW if I assume 0.5 air changes per hour as opposed to the default of about 2.5. 

 

Two MCS surveyors, each after a 3 hr survey, have told me I need 16kW, which is obvious nonsense.  They ignored the fabric upgrades that I made a special point of telling them about, double counted room to room losses and of course assumed ACH=2.5 per hour.

 

Assuming

a) I can get planning consent (don't ask) and

b) I can persuade an MCS installer to cooperate

 

I will be installing an 8kW ASHP, most likely the Vaillant badged as 7kW but actually good for over 8kW at 45C flow temp.

 

Scaling my measurements to your 12MWh suggests you do indeed need about 5.5kW.  You also need a heat pump that will comfortably modulate down to 3kW at say 10C OAT, preferably 2.5kW.  Obviously this is subject to there not being surprises in the answers to the question above, and is not a recommendation just an observation.  I wont be surprised if MCS installers recommend 10kW, which will most likely perform badly as it will probably be oversized.

 

 

Edited April 20 by JamesPa

[MrPotts]

@JamesPa

We have a 4bed detached house built in 1996. The annual gas consumption figure came from our smart meter.

 

A few questions

• is the house warm enough? Yes 20-21c.
• Whats your heating pattern? Heating on 7-10am and 4-10pm
• is it a gas boiler, how old. 9 year old 24kw gas boiler.
• What does your EPC say your consumption is. Not got an EPC, we have lived here for 25years.
• In the spreadsheet calculation, what are the assumptions about the fabric that you cant verify, particularly air changes per hour? I just entered room sizes, radiator sizes and building fabric details, all other assumptions were left at defaults.
• Where in the country are you/whats the local design temperature. North Hertfordshire, design temp was -2.3c

 

[JamesPa]

1 hour ago, MrPotts said:

@JamesPa

We have a 4bed detached house built in 1996. The annual gas consumption figure came from our smart meter.

 

A few questions

 

• is the house warm enough? Yes 20-21c.
• Whats your heating pattern? Heating on 7-10am and 4-10pm
• is it a gas boiler, how old. 9 year old 24kw gas boiler.
• What does your EPC say your consumption is. Not got an EPC, we have lived here for 25years.
• In the spreadsheet calculation, what are the assumptions about the fabric that you cant verify, particularly air changes per hour? I just entered room sizes, radiator sizes and building fabric details, all other assumptions were left at defaults.
• Where in the country are you/whats the local design temperature. North Hertfordshire, design temp was -2.3c

 

All good stuff and no surprises.  One more question, by how much does it cool during the night time off period (Im trying to get an estimate of the average house temp to correct for that). 

 

For example if it cools by say 5C then the average house temperature is very, very roughly 9*20.5 +15*(20.5-2.55) = 19C.  So any figure calculated from annual consumption should be uplifted by (20.5--2.3)(19--2.3), about 7%.  That's assuming you plan to heat 24x7 with the ASHP when the OAT is at -2.3.

 

If you can download your half hourly readings then you can look back at the consumption to be even more sure.  I averaged over 3 hours to iron out, but with your short heating periods the averaging time would arguably need to be longer. 

 

The reason to do these calculations, IMHO, is to sense check what prospective suppliers tell you.  If you can get close to the same answer using two or more independent methods, then its probably near enough right.  With a 1996, ie fairly modern, house the default assumptions should be fairly close, unless you have done fabric upgrades (eg thicker loft insulation or better windows). Have you?.

Edited April 20 by JamesPa

[MrPotts]

10 minutes ago, JamesPa said:

All good stuff and no surprises.  One more question, by how much does it cool during the night time off period (Im trying to get an estimate of the average house temp to correct for that). 

 

For example if it cools by say 5C then the average house temperature is very, very roughly 9*20.5 +15*(20.5-2.55) = 19C.  So any figure calculated from annual consumption should be uplifted by (20.5--2.3)(19--2.3), about 7%.  That's assuming you plan to heat 24x7 with the ASHP when the OAT is at -2.3.

 

If you can download your half hourly readings then you can look back at the consumption to be even more sure.  I averaged over 3 hours to iron out, but with your short heating periods the averaging time would arguably need to be longer. 

 

The reason to do these calculations, IMHO, is to sense check what prospective suppliers tell you.  If you can get close to the same answer using two or more independent methods, then its probably near enough right.  With a 1996, ie fairly modern, house the default assumptions should be fairly close, unless you have done fabric upgrades (eg thicker loft insulation or better windows). Have you?.

Hi James

 

All that I have done is increase the loft insulation from 100mm to 300mm.  The house temperature usually drops to about 17c in winter, about 19c in summer.

[MrPotts]

4 hours ago, JohnMo said:

Think your main reference accepted the rule of thumb wasn't that good and produced a spreadsheet.

 

https://protonsforbreakfast.wordpress.com/2023/04/02/another-heat-pump-spreadsheet-beyond-the-rule-of-thumb/

I have had a look at the epreadsheet but it doesn't work for me.  If I change the default temperatures I get the #ref error.

[JamesPa]

27 minutes ago, MrPotts said:

Hi James

 

All that I have done is increase the loft insulation from 100mm to 300mm.  The house temperature usually drops to about 17c in winter, about 19c in summer.

The extra loft insulation is going to be worth 1kW or possibly more (somewhere you should be able to change the assumption about loft insulation!), which reduces your spreadsheet calculation to perhaps 6.6kW. 

 

Its about a 3.5C drop from 20.5 to 17.  The average house temp in winter is thus very roughly 9*20.5 +15*(20.5-1.75) = 19.4C.  The correction factor is therefore 20.5--2.3)(19.4--2.3), about 5% suggesting your design temperature load is about 5.8kW.  You will need bigger than this if you plan to heat part time when its at its coldest.   5.8kW and 6.5kW are not that far apart!

 

Based on this my guess, and it is only an (educated) guess, is that you would get away with a 6kW unit or an 8kW unit.  The latter, unless it has a particularly good modulation depth, will likely cycle a bit at mild temperatures (when you are likely to be running at 2.5kW or thereabouts), but you are probably in practice going to continue turning it off at night anyway just for comfort, which will tend to counter that because you need to deliver more heat during the day to compensate for the energy lost from the fabric at night.  I would be very wary of any suggestion from an installer that you need bigger than 8kW, and if you went for 6kW then you may need a bit of supplementary heating in an extended very cold spell, which happens rarely in Hertfordshire.  Note sticker capacity and actual capacity at design conditions may differ, either way!

 

Quite a few people argue, contrary to the practice of most installers, that mild undersizing is better than oversizing.  From an efficiency perspective this is true and, if you are prepared to put up with occasionally needing a bit of supplementary heating it is likely a good way to go.  Others will argue that you should always oversize if in doubt.   Its a trade off of efficiency vs the inconvenience of occasionally needing supplementary heating.  

 

Note that some heat pumps have poor modulation capability (ration of max to min output).  This includes some 'big names'.  Some manufacturers  produce fewer distinct hardware variants than the number of different models they advertise, basically 'range rating' them by clamping the max output in the firmware.  The downside of this is that, in  a 'clamped' variant, the min output is still the same as it was when the same hardware was not 'clamped', so the modulation ratio is degraded.  The more transparent manufacturers don't do this and are more 'honest' about the number of distinct hardware variants they actually produce.

 

Hope that helps.

Edited April 20 by JamesPa

[MrPotts]

17 minutes ago, JamesPa said:

The extra loft insulation is going to be worth 1kW or possibly more (somewhere you should be able to change the assumption about loft insulation!), which reduces your spreadsheet calculation to perhaps 6.6kW. 

 

Its about a 3.5C drop from 20.5 to 17.  The average house temp in winter is thus very roughly 9*20.5 +15*(20.5-1.75) = 19.4C.  The correction factor is therefore 20.5--2.3)(19.4--2.3), about 5% suggesting your design temperature load is about 5.8kW.  You will need bigger than this if you plan to heat part time when its at its coldest.   5.8kW and 6.5kW are not that far apart!

 

Based on this my guess, and it is only an (educated) guess, is that you would get away with a 6kW unit or an 8kW unit.  The latter, unless it has a particularly good modulation depth, will likely cycle a bit at mild temperatures (when you are likely to be running at 2.5kW or thereabouts), but you are probably in practice going to continue turning it off at night anyway just for comfort, which will tend to counter that because you need to deliver more heat during the day to compensate for the energy lost from the fabric at night.  I would be very wary of any suggestion from an installer that you need bigger than 8kW, and if you went for 6kW then you may need a bit of supplementary heating in an extended very cold spell, which happens rarely in Hertfordshire.  Note sticker capacity and actual capacity at design conditions may differ, either way!

 

Quite a few people argue, contrary to the practice of most installers, that mild undersizing is better than oversizing.  From an efficiency perspective this is true and, if you are prepared to put up with occasionally needing a bit of supplementary heating it is likely a good way to go.  Others will argue that you should always oversize if in doubt.   Its a trade off of efficiency vs the inconvenience of occasionally needing supplementary heating.  

 

Note that some heat pumps have poor modulation capability (ration of max to min output).  This includes some 'big names'.  Some manufacturers  produce fewer distinct hardware variants than the number of different models they advertise, basically 'range rating' them by clamping the max output in the firmware.  The downside of this is that, in  a 'clamped' variant, the min output is still the same as it was when the same hardware was not 'clamped', so the modulation ratio is degraded.  The more transparent manufacturers don't do this and are more 'honest' about the number of distinct hardware variants they actually produce.

 

Hope that helps.

Thank you for this detailed information, it is extremely useful and certainly gives me confidence in the Heatpunk results.  I have already selected the correct depth of  loft insulation within Heatpunk.

 

Looks like most of my radiators will need replacing if i decide to go ahead.

 

The Heatpunk recommended heat pump is an Vaillant aroTHERM plus 7.  The heat pump load is.......

 

The Outside Design Temperature for postcode AL5 is -2.3°C

At the Outside Design Temperature and a Flow Temperature of 45 °C the output power of the heat pump is 8563 W.

The heat loss from the building at the Outside Design Temperature is expected to be 7689 W.

The total area of the building is 157.98 m2, giving an average heat loss of 49 W/m2.

 

Thanks for all your help.

The Outside Design Temperature for postcode AL5 1HG is -2.3°C

At the Outside Design Temperature and a Flow Temperature of 45 °C the output power of the heat pump is 8563 W.

The heat loss from the building at the Outside Design Temperature is expected to be 7689 W.

The total area of the building is 157.98 m², giving an average heat loss of 49 W/m².

[sharpener]

20 minutes ago, MrPotts said:

Looks like most of my radiators will need replacing if i decide to go ahead.

 

Heatpunk gives 7763W as the output at 53C flow and -2.3 OAT. So by choosing 53 not 45 as the flow temp it would still more than match the heat loss, and you might well need fewer radiator upgrades and maybe saving more in capital than you would lose in running costs.

 

You will probably not find an MCS installer willing to do this though.

 

When I lived in Hitchin there was snow on the ground for some weeks in I think the winter of 1968 - 69, but it was most unusual, I don't suppose you will see -2.3 very often nowadays.

 

[JohnMo]

4 hours ago, MrPotts said:

I have had a look at the epreadsheet but it doesn't work for me.  If I change the default temperatures I get the #ref error.

That's because you changed the temperature in the wrong cell, you change the temperatures in cells C8, C9, and C10 only. You change the yellow and with red text cells as it says in the intro tab.

 

 

[JohnMo]

Put all my details into the spreadsheet for my gas boiler (last year's data) and the ASHP size comes out a little too small.  To get that I had to put the actual lowest temperature we got , which was -7. 

 

I know the normal degree days don't work very well once well insulated, as I don't need to heat when average temp goes above 10 deg, not the normal 15 degs. Messing about with the degree days I altered the degree days location to Heathrow instead of Aberdeen (800 degree days less), I get about the right size Heat Pump output.

 

Spreadsheet is ok for the more average house. A good enough ball park for a sense check.

[JamesPa]

4 hours ago, MrPotts said:

The Heatpunk recommended heat pump is an Vaillant aroTHERM plus 7.  The heat pump load is.......

Attached is the table of max outputs as a function of FT and OAT.  If you felt like being adventurous you might just get away with the 5kW model which is actually just over 6kW at your design parameters.  

 

Unfortunately I don't have min outputs to hand (perhaps @sharpener (or someone else) does. 

 

Postscript, I found the detailed specs on another forum here https://community.openenergymonitor.org/t/vaillant-arotherm-owners-thread/21891.  They are in Czech, but the graphs are easy to read. 

 

 

 

arotherm-plus-spec-DB outside.pdf

Edited April 20 by JamesPa

[MrPotts]

1 hour ago, JohnMo said:

That's because you changed the temperature in the wrong cell, you change the temperatures in cells C8, C9, and C10 only. You change the yellow and with red text cells as it says in the intro tab.

 

 

That is where I changed them. When I then selected which temperature for morning, afternoon etc that is when I got the #ref errors. I will have another go tomorrow.

Edited April 20 by MrPotts

[sharpener]

You can download the full set of tables from the Arotherm FB page here starting at p34.

 

5 hours ago, MrPotts said:

The heat loss from the building at the Outside Design Temperature is expected to be 7689 W.

 

So I can't see any way the 5kW model would satisfy this even with massive rads and a flow temp of 30.

[JohnMo]

7 minutes ago, MrPotts said:

That is where I changed them. When I then selected which temperature for morning, afternoon etc that is when I got the #ref errors. I will have another go tomorrow.

Just leave it all 20, keep it simple, your either going to batch charge the house or run on Weather Compensation. So 20 is close enough or your average temperature.

[JamesPa]

8 hours ago, sharpener said:

You can download the full set of tables from the Arotherm FB page here starting at p34.

Thats the same Czech document as on the OpenEnergyMonitor thread I posted earlier.  I have not seen this document in English (and presumably neither has @sharpener), I wonder why not.  Vaillant are a German company but perhaps they design their heat pumps in the Czech republic,  Alternatively perhaps they think that Czech heating engineers are more likely to understand it.

 

8 hours ago, sharpener said:

So I can't see any way the 5kW model would satisfy this even with massive rads and a flow temp of 30.

 

According to the document

 

• the 5kW model can do about 6.5kW at FT45/-2 and modulate down to 2.5kW at FT35/+12
• the 7kW model can do about 9kW at FT45/-2 and modulate down to 3.2kW at FT35/+12

I have chosen FT35/12 as a representative 'mild weather' scenario assuming weather comp

 

The document also seems to list a 6kW model which does about the same as the 5kW model, which may be why its not available in the UK sofaik.

• According to the spreadsheet OP's loss is 7698.  This, is likely to translate to about 4kW most of the time
• According to the 'total energy' method OPs loss is 5.8kW.  This is likely to translate to about 3kW most of the time.

That, in all probability, bounds the uncertainty.  Ideally therefore we want a machine that can do 8kW and can modulate down to 3kW.

 

If OP could verify the actual loss eg from smart meter measurements is closer to the 'total energy' figure than it is to the 'spreadsheet' figure then it might be possible to show that the 5kW model is adequate.  Otherwise I would agree with @sharpener that it is not a safe choice.

 

The 7kW model should, based on what we have heard, have adequate capacity.  At a min output of 3.2kW it might struggle to achieve the 'most of the time' requirement without cycling if, in fact, the house demand is only 5.89kW.  However as OP is likely to do an element of batch heating (based on current patterns) this should, with a bit of luck, compensate.  

 

Thus I would tend to agree with @sharpener that the 7kW model is likely the one to choose unless the actual demand can be shown to be closer to the 'total energy' figure.  There is a risk of cycling in the shoulder season, but mild batch heating, which appears likely to be the desired pattern anyway, will probably overcome this.  If I were OP I would look at whether I could bring any more evidence to bear on the question, but if I couldn't I wouldn't sweat about it and would run with the 7kW.  That's of course if OP can persuade anybody to fit it!

 

Fir what its worth here are the performance tables for the '8.5kW' R32 Mitsubishi Ecodan.  Given the uncertainties I would describe these as not markedly different to the Vaillant ones as regards capacity and modulation.

 

 

 

[sharpener]

40 minutes ago, JamesPa said:

I have not seen this document in English (and presumably neither has @sharpener), I wonder why not. 

 

No. Most Vaillant installers I have spoken to do not know they exist. Nor did the V regional tech mgr for the SW.

 

I have not found any explanation for why they are not available in either English or German (which I can read well enough). There is also a Planning your Installation doc of which the EN and DE versions are very different, it's all rather a mess, on the FB page there have been various discussions about this.

[sharpener]

1 hour ago, sharpener said:

There is also a Planning your Installation doc

 

https://www.vaillant.ua/downloads/proekti/0020222099-10-compressed-2406863.pdf?fbclid=IwZXh0bgNhZW0CMTAAAR2cgInGE7BtvCgpdxYDszZUgKRAQE8jm3AIb_4C0KwHKwd_pHVLc5itfYA_aem_AaxrDlhR_6iT7VEGsQhhKVj5KjwLaeJTS-GNzLG_Gy5tBQRLd1zxg8__lVP42DpLQzO2V26rFwsW0FhdKIEFWMSo

[JamesPa]

On 20/04/2024 at 16:53, JamesPa said:

On 20/04/2024 at 16:28, MrPotts said:

The house temperature usually drops to about 17c in winter, about 19c in summer.

Its about a 3.5C drop from 20.5 to 17.  The average house temp in winter is thus very roughly 9*20.5 +15*(20.5-1.75) = 19.4C.  The correction factor is therefore 20.5--2.3)(19.4--2.3), about 5% suggesting your design temperature load is about 5.8kW.  You will need bigger than this if you plan to heat part time when its at its coldest.   5.8kW and 6.5kW are not that far apart!

Thinking about it I realise that the calculation of the 'correction factor' for part time heating is wrong.  Since it is effectively being applied to the whole season consumption, it should be calculated with reference to the average OAT not the extreme OAT.  I don't know what the average OAT in the heating season is, but its probably around 7C https://www.statista.com/statistics/322658/monthly-average-daily-temperatures-in-the-united-kingdom-uk/. 

 

Thus the 'correction factor' is  20.5-7)(19.4-7), about 8% instead of 5%.  This means that the estimated load based on the annual consumption is  suggesting your design temperature load is about 6kW.  This doesn't make much deference but does increase the level of confidence  that the 7kW Vaillant should do the job (see calculations in this post)