How to predict heat pump size from your EPC

[sharpener]

Here is a thought-provoking result:

 

Estimated Heat Pump size in kW = 0.4 x (Total EPC annual energy requirement in MWh)

 

For the reasoning behind the formula see this post in Michael Podesta's blog.

 

It would be interesting to know whether anyone on here has results which agree with it, or otherwise!

[JohnMo]

Doesn't section 8 not have the heat loss rate in Watts for each month. So wouldn't it be the figure for the worst month (Jan in my case), then allow enough for dhw heating.

 

Mine said 2980W, Jeremy spreadsheet pretty similar, heat meter said the same. Using a 6kW ASHP, which is really a little large, but works fine.

[sharpener]

2 hours ago, JohnMo said:

Doesn't section 8 not have the heat loss rate in Watts for each month.

 

Mine issued 22/5/23 doesn't, in fact it does not have numbered sections at all. It just has (separate) estimates for total annual heat requirement in kWh like this:

 

Estimated energy needed in this property is:
• XXXXX kWh per year for heating
• YYYY kWh per year for hot water

 

I haven't been able to find an example online that looks exactly like mine. There are several links to this format which appears to date from 2012, however somewhere I read it  was discontinued in 2017 so not a lot of help.

 

Your method is fine @JohnMo so long as your EPC shows such a high level of detail, what date was the EPC survey done? Can you find a similar published example or could you post an anonymised extract so I can see what the relevant bit looks like?

 

 

Edited April 18 by sharpener

[JohnMo]

9 hours ago, sharpener said:

There are several links to this format which appears to date from 2012, however somewhere

I see you only have the certificate, there is another set of sheets, where the certificate data is pulled from. It has your U values, airtightness and all the calculations for the house.

 

Just looking at the formula and applying to my figures, it ends up saying I need a 5kW heat pump, using just the house heating figures, so really no need to add an allowance for DHW to that as already a little large. You could definitely end up with a heat pump one larger than needed using 0.4.

 

Using my figures - Multiplication using 0.25 for heating only then adding a DHW allowance would be very close to perfect for my house.

[ProDave]

My house had a full as built SAP done to get the as built EPC.  The energy input that predicted in spite of having all the correct data is about 3 times as much as it actually uses.  Jeremy's spreadsheet gives a far more accurate prediction.

[SteamyTea]

12 hours ago, sharpener said:

Estimated Heat Pump size in kW = 0.4 x (Total EPC annual energy requirement in MWh)

Well as I am one of six places, which all have an EPC, lets us see what comes out

 

No 1 0.4 x 6,788 = 2,715 kWh.year^-1 

 

No 5 0.4 x 6,222 = 2,488 kWh.year^-1 

 

No 6 0.4 x 6,371 = 2,548 kWh.year^-1 

 

Now as some of the EPCs are the old ones, I can try and estimate from primary energy usage for the missing data.

 

No 1 Primary Energy 368 kWh.m².Year^-1 

 

No 2 Primary Energy 289kWh.m².Year^-1 

 

No 3 Primary Energy 320 kWh.m².Year^-1 

 

No 4 Primary Energy 183 kWh.m².Year^-1 

 

No 5 Primary Energy 364 kWh.m².Year^-1 

 

No 6 Primary Energy 378 kWh.m².Year^-1 

 

So using the average conversation factor of 17.5 x Primary Energy I can estimate the missing data.

 

No 2 0.4 x 5,047 = 2,018 kWh.year^-1 

 

No 3 0.4 x 5,589 = 2,236 kWh.year^-1 

 

No 4 0.4 x 3,196 = 1,278 kWh.year^-1 

 

Taking the median (middle value to account for odd results) of those 6 and converting to MWh gets us to 5.9

 

Now I have no idea how my neighbours live, but I do know my space and DHW usage.  The mean winter period for this year, so far is 11.5 kWh.day^-1. So if the heating was on permanently that would be 1.68 MWh.year^-1 which implies a heat pump size of 1 kW. But it isn't.  For 8 months of the year my usage is 5 kWh.day^-1, so space heating and hot water is 2.4 MWh.year^-1, that implies a heat pump size of 1 kW.

 

So I would not use EPCs as a guide to heat pump sizing as it could easily be 70% out, and in a small house impossible to buy.

 

Rather than guess at usage from proxy data, just spend a few quid and measure it.  If a HP installation is going to cost £3000, spending £100 on an energy and temperature logger really will take the mystery out of it, and if you have Smart meters installed, all you need is a dual temperature logger.

 

 

I think I have made an error somewhere, shall try and find it before I go out.

Actually I think I am right, I did not take into account that 4 of the houses are end terrace/semi detached and mine is a mid terrace, so less exposed wall area by quite a margin.  My usage is similar to the other mid terrace, so an adjust upwards should be applied.

Edited April 19 by SteamyTea

[ADLIan]

Remember SAP is not intended to be a deign tool for the space and water heating systems.

[sharpener]

2 minutes ago, SteamyTea said:

No 1 0.4 x 6,788 = 2,715 kWh.year^-1

 

No, I think you have got the dimensions wrong, but I am not sure I follow yr calcs, the product is the HP size in kW not kWh.m².Year^-1.

 

Don't know what yr 1.75 is.

 

Should kWh.m².Year^-1 be m^-2?

 

12 hours ago, sharpener said:

Estimated Heat Pump size in kW = 0.4 x (Total EPC annual energy requirement in MWh)

 

[SteamyTea]

1 minute ago, sharpener said:

No, I think you have got the dimensions wrong, but I am not sure I follow yr calcs, the product is the HP size in kW not kWh.m².Year^-1

I thought I corrected for that later on.

1 minute ago, sharpener said:

Don't know what yr 1.75 is.

That is to estimate the usage from the missing EPC data, it has the kWh.m^-2.year^-1 data on them all, but only the total annual usage on 3 of them.

3 minutes ago, sharpener said:

Should kWh.m².Year^-1 be m^-2?

Yes, blame it on gungy eyes this morning after a days gardening yesterday.

[sharpener]

42 minutes ago, ProDave said:

My house had a full as built SAP done to get the as built EPC. 

 

Well clearly if you have got all the underlying data then it is of course better to use that in Jeremy's Spreadsheet or Heatpunk or Heat Engineer.

 

I have been impressed by Michael Podesta's use of degree-days to calculate HP size from gas consumption. Essentially this method just estimates gas consumption from the EPC figures which can then be used to imply a certain HP size.

 

The old GINO adage always applies, Garbage In Nonsense Out.

[sharpener]

3 minutes ago, SteamyTea said:

That is to estimate the usage from the missing EPC data, it has the kWh.m^-2.year^-1 data on them all, but only the total annual usage on 3 of them.

 

The three that you have to extrapolate in this seem to be rather different from the first three so I would be a bit sceptical about using them but I suppose taking the median helps.

 

I take it these are perhaps smallish ?terraced houses (or maybe bungalows) of identical floor area?

 

In which case an HP of 0.4 x 5.9 = 2.4 kW sounds plausible, whereas your 1 does not.

[SteamyTea]

26 minutes ago, sharpener said:

take it these are perhaps smallish ?terraced houses (or maybe bungalows) of identical floor area?

All identical except 4 are end terraced.

My usage, like my immediate neighbour, is lower as we are single occupancy and out a fair bit at work, and both as tight as  Cornish can be. And we are both the only mid terraced.

 

If 6MWh annual usage is close to real, then that implied a 2.4 kW heat pump, which realistically becomes a 4 kW heat pump as I think that is just about the smallest available (might be 3.8kW).

 

I think the main point is that using an EPC is probably not a reliable way of sizing (the multiply by 0.4 has been touted around for a long time).

OK as a first order check, but probably that is all.

 

Much better to spend an hour a week reading the meters.

[JohnMo]

52 minutes ago, sharpener said:

have been impressed by Michael Podesta's use of degree-days to calculate HP size from gas consumption. Essentially this method just estimates gas consumption from the EPC figures which can then be used to imply a certain HP size

I originally used degrees days for heat calcs, but the better you insulated and made airtight the further you need to manipulate degrees days to work. I stop needing heat once the average outside temp get to about 10 degs.

 

Rules of thumb generally as reliable as using your thumb as a straight edge.

 

Just calculate the overall house loss and allow for DHW - done.  Use room heat loss calcs for radiator sizing.

[sharpener]

53 minutes ago, SteamyTea said:

(the multiply by 0.4 has been touted around for a long time)

 

Do you have a reference for this? It was new to me.

 

54 minutes ago, SteamyTea said:

OK as a first order check, but probably that is all.

 

Yes.

 

54 minutes ago, SteamyTea said:

Much better to spend an hour a week reading the meters.

 

If for example you are looking at houses to buy they will have to show you (or you can look up) the EPC and from this you might then get a fair idea of HP size. As you also could using the Heat Geek cheat sheet. And then you could work out where it might be sited. (The installers who came yesterday are proposing a big Stiebel Eltron mounted over a passageway on wall brackets).

 

But you cannot retrospectively ask for a year of weekly meter readings to be done, at most you might get quarterly gas bills and of course in that case the result would be more accurate than using the EPC.

[JohnMo]

3 minutes ago, sharpener said:

cannot retrospectively ask for a year of weekly meter readings to be done, at most you might get quarterly gas bills and of course in that case the result would be more accurate than using the EPC.

Maybe maybe not - if a gas boiler it could be hugely oversized short cycling like mad and consuming a disproportionately high amount of gas. More shi** in shi** out.

[SteamyTea]

48 minutes ago, sharpener said:

Do you have a reference for this? It was new to me.

Not without looking. It was mentioned when I was at university 20 years ago by one of the lectures, and I think that link in the first post covers it.

[sharpener]

56 minutes ago, SteamyTea said:

It was mentioned when I was at university 20 years ago

 

I don't think we even had EPCs 20 years ago, this article says they were introduced on 1 August 2007.

 

59 minutes ago, SteamyTea said:

I think that link in the first post covers it

 

Michael Podesta's blog on predicting HP size from gas consumption dates from 2022, and the extension of the idea to EPCs first appeared yesterday AFAIK.

 

[JamesPa]

16 hours ago, sharpener said:

Here is a thought-provoking result:

 

Estimated Heat Pump size in kW = 0.4 x (Total EPC annual energy requirement in MWh)

 

For the reasoning behind the formula see this post in Michael Podesta's blog.

 

It would be interesting to know whether anyone on here has results which agree with it, or otherwise!

I would trust this about as much as I trust the MCS GIGO spreadsheets - ie not at all.  

 

Personally I think that, for a retrofit which has had any 'history' of fabric changes since it was first built (ie most retrofits earlier than perhaps the 1980s), the only sane way is to measure it, whether from actual annual consumption (as opposed to the EPC estimated annual consumption) or more sophisticated smart meter data if you have it, or if you really want to go to town one of the companies that put in a calibrated heater.

 

I do accept @JohnMo s comment about hugely oversized short cycling gas boilers.  Most gas boilers are of course well oversized, so I would certainly not assume  the >100% efficiency that is claimed for a condensing boiler unless there is evidence to support it.  But a measurement of actual consumption is still better, surely, than pure guesswork by an MCS surveyor, whose only real concern is to ensure that the heat pump is well bigger than it needs to be in order to avoid call-outs and guarantee that he cant be challenged on the grant rules.  At least with actual consumption you are unlikely to be more than about 20% out and, unless the gas boiler is perfectly set up and condensing properly (which most aren't judging from observation locally), this will be erring on oversizing not undersizing.

 

Obviously for a new build you cant do this, but for a new build you know what the fabric is.  For a retrofit the MCS guys will (at least in my personal experience) make some random assumptions which may be a long way from the actual truth.  EPCs are even less rigorous than MCS GIGO, at least as  far as I understand them.

 

Edited April 19 by JamesPa

[MrPotts]

If I use the rule of thumb on Michael de Podesta's blog then I need a ASHP of 4.28kW, I have also gone through the process of working out a heat pump size on Heatpunk and that states 7.69kW.  Which is right, I have no idea!

[JamesPa]

5 hours ago, MrPotts said:

If I use the rule of thumb on Michael de Podesta's blog then I need a ASHP of 4.28kW, I have also gone through the process of working out a heat pump size on Heatpunk and that states 7.69kW.  Which is right, I have no idea!

Do you know how much energy you currently use to heat your house?  How does it compare with the EPC value?  If you do that will give you a clue.  Have you got any way of monitoring it real time ( I watch my smart meter and download half hourly meter readings). If you can do this then you should be abke pretty much to nail it.

 

Also how much of the 7.7kW is down to unknowns (the biggest possibly being ach). Again potentially bit more data ( this time on the uncertainty in the spreadsheet answer).

 

Realistically all you probably need to know is whether to buy an 8kW or a 5-6kW pump.  If your house really is 4.3kW, then most of the time it will likely be consuming 2.2, so you need the pump to work smoothly at that load.  An 8kW one won't, a decent 5-6kW one should.  On other hand if your house is really 7.5kW it will be consuming 4kW most of the time, which a decent 8kW model will modulate down to.

 

Sometimes (not always) combining data may allow the answer to be deduced to a level which is, with a high degree of certainty, good enough.

 

 

Edited April 19 by JamesPa

[PhilT]

for those who have or had gas - kWh usage / 2900 seems to be a good sense check (can't remember what efficiency was assumed - 90% I think). Mine would have been 21000/2900 = 7kWh which is spot on based on my highest recorded peak usage on the coldest winter day/night so far. I have an 11.2kW heat pump but I'm still getting COP of 4.2 average for Oct-March so not a disaster by any means.

[MrPotts]

10 hours ago, JamesPa said:

Do you know how much energy you currently use to heat your house?  How does it compare with the EPC value?  If you do that will give you a clue.  Have you got any way of monitoring it real time ( I watch my smart meter and download half hourly meter readings). If you can do this then you should be abke pretty much to nail it.

 

Also how much of the 7.7kW is down to unknowns (the biggest possibly being ach). Again potentially bit more data ( this time on the uncertainty in the spreadsheet answer).

 

Realistically all you probably need to know is whether to buy an 8kW or a 5-6kW pump.  If your house really is 4.3kW, then most of the time it will likely be consuming 2.2, so you need the pump to work smoothly at that load.  An 8kW one won't, a decent 5-6kW one should.  On other hand if your house is really 7.5kW it will be consuming 4kW most of the time, which a decent 8kW model will modulate down to.

 

Sometimes (not always) combining data may allow the answer to be deduced to a level which is, with a high degree of certainty, good enough.

 

 

I use just under 12.5kWh of gas in a year. That also includes hot water and cooking on a gas hob.

Edited April 20 by MrPotts

[JohnMo]

1 hour ago, MrPotts said:

just under 12.5kWh of gas in a year

Are you really sure, MWh?

[MrPotts]

7 minutes ago, JohnMo said:

Are you really sure, MWh?

Sorry, I meant 12,500kWh

[JoeBano]

I’ve just got a hob on gas (mistake on kitchen planning) I use around 25 kWh a month if that helps. Cost £1.92 and £7.90 standing charge 😭

Edited April 20 by JoeBano