Has anyone tried comparing actual ASHP performance with their heat loss calcs?

[DInwood]

I'm trying to reconcile the actual power consumed by our ASHP with what I expected: I'm finding it is taking more than twice the power I calculated, a combination of lower than specified CoP and higher than predicted heating energy output.  I'll explain my reasoning below, but my interim conclusions are either

1. the ASHP manufacturer's reported heat output AND their claimed CoPs are seriously overstated, OR
2. I completely misunderstand the relationship between MCS heat loss calculations and reality.

It may of course be a combination of both and I look forward to hearing what other people's experiences are of this.

This is not simply a complaint that our CoP is too low (this is a fact, but the heating installer is on the case and making gradual improvements.  However he seems to think our best ever day of 3.5 was good [despite manufacturer claimed 4.4 for those temps] but that ignores the issues I discuss here).

For the avoidance of doubt, I’m discussing real measured evidence.  I’m not talking about the system producing too much warmth – the house is about the right temperature, but I’m taking into account what is actually measured.  I’m also discussing a “first world problem”, I know that compared to many we are using not a lot of energy.  This is very much an intellectual exercise from a very analytical person with too much time on his hands (and who spent too much on building a low energy home)!

So, the background:

We bought a 1970's bungalow just before covid and refurbished/extended it into a 4 bedroom, 150m² house intended to be energy neutral.  It is well insulated, airtight (with MVHR), heated by ASHP and has plenty of PV.  We asked the architect to get as close to Passivehouse as we could without incurring the costs of certification.  We moved in in April '22 and when the heating was first needed in November, I was shocked at how quickly the kWh into the HP were racking up (regularly consuming >10kWh/day with outside temps of only 7C).  I love a good spreadsheet so I quickly set out to try to understand what was going on.  I was "lucky" that I had just started playing with Home Assistant and was able to produce good logs of actual daily average temperatures & energy use.

My starting point was to model expected heat demand.  I had the MCS design compliance pack provided by the heating engineer and it wasn't difficult to reproduce this in a spreadsheet using actual u values & dimensions (from the architect's drawings and suppliers).  This gave me fabric heat loss figures almost exactly the same as the MCS pack.  However I found a big divergence in their ventilation losses as they had not taken into account the MVHR.  Instead I used the heating engineer's ventilation figures and the MVHR's claimed efficiency (verified by looking at the actual airflow temperatures).  I've ignored draughts as we got a good air-tightness test result (1.1m³/h/m² at 50Pa and we've since sealed a leak that accounted for some of that).  This allows me to work out the total building heat losses per degree difference between inside & outside temperatures.  At that time I didn't have room by room actual temps so I used a house average taken from the MVHR to be the temp of the air extracted.  The actual outside temperature was taken from the ASHP.

As a result of these calculations, I get a total house heat loss of 88W/C (and just under 12W/m²).

Now for the difficult bit.  In the MCS pack they estimate total annual heating demand using "degree days", an official annual figure which is in my view a shortcut to reality and a recipe for waiting for a year before you find out there's a problem!  What is really going on is that the house is losing heat, some of which comes from background sources and the rest supplied by the heating system.  These background sources are well known and some are listed & estimated in the MCS pack.  These are: hotwater tank losses, heat given off by occupants, heat given off by electrical appliances and solar gain.  HW tank losses are easy (from the manufacturer, pro-rated down as we keep our tank quite a bit cooler than they assume).  Body losses are taken at 80W per person (and I'm afraid I really do have an estimate of daily occupancy since I started this - I know, a sad case).  Electrical appliance heat output is going to have to be a guesstimate, although I have a good record of daily power consumption (& can take off the only large external consumption which is the car charger), some of this isn't going to produce heat.  However I make a good stab at estimating this & can share my formula, which I think allows for the dishwasher & washing machine's discharge of hot water.  That leaves the solar gain, which on our north facing site is pretty low in winter.  Some sun does come in however and I have made small allowance for this from calculations based on the PV power generated daily (again, happy to share my thinking).  We were away over the new year & my logs showed the house cooling down slowly (0.1C per day per degree difference between inside & outside temp).  In an idle moment I looked at what I could learn from this and was surprised at the inferred thermal mass.  It tells me that even small day-to-day internal temperature changes can consume (or release) a lot of heat.  I now take this into account.

So I've now got a speadsheet that can estimate the net heat input that should be needed to explain the actual inside temperature each day.  I've got the actual electrical power taken by the ASHP (a meter on the consumer unit that I have to record every night - it's a scandal that I haven't automated this!).  I have to take off the energy used to heat water (I'm pretty sure I've got a good way of calculating this as this was the only thing the ASHP was doing from April until November), so I know fairly accurately what power the ASHP is taking each day to produce heat.  I know what I was expecting from the above net heat requirements & using the ASHP CoP tables (which allow for the ambient temp outside & the flow temp to the UFH).  What I'm seeing is an electrical power usage that is 2.2 times higher (last week's average).

Is my logic wrong or is the ASHP lying?

 

 

In exploring the details I can look at the ASHP's own estimate of heat output which it does by measuring flow & return temperatures and the flow rate.  They stress it's an estimate, but if my net demand calcs are right it must be a 50% over-estimate.  They also give the electrical energy consumed (which is also slightly wrong compared to the panel meter) and from that I can see the CoP as they would calculate it: last week it averaged 3.0, compared to their tables for these temperatures of 4.1.  This is beginning to feel like Dieselgate: manufacturers producing performance data that has no relationship with reality.  Or I'm mad in doing these sums.  Which is it do you think?

 

 

My installation:-

157m² 4 bed house N facing, very close to the Torridge estuary (so winter average temp is warmer than Devon generally).  The old structure had insulation & airtightness added underfloor (u-value 0.16) & external insulation added & cavities refilled (u-value 0.2).  New flooring was slab on insulation u-value 0.14, new walls 0.15, flat roof 0.12.  There is lots of Norrsken triple glazing, average u-value 0.79 & some roof lights 0.6.  The weighted average is 0.23.

We have: Ecodan PUZ-WM85VAA monobloc ASHP, Zehnder Q350 MVHR, Omni UFH, 6.4kWh of PV mounted on frames on otherwise flat roof.

My calculations:

A spreadsheet that works out heat losses in watts per degree C difference between average internal and external temp as discussed.

A logic chain like this:

Gross heat requirement = 88W x (outside - inside temp)

- Internal heat generated = bodies x 80W + tank loss + f₁(house power consumption) + f₂(PV power generated) + thermal mass change

= net heat requirement

Electrical power expected = net heat requirement / ecodan claimed CoP at today's ext temp.

 

 

I’ve been thinking hard about error sources that might explain a day to day variability in my data and the huge discrepancy discussed above.  I did wonder whether wind might be a factor but the only article I found made me dismiss this (an Irish study which showed that as long as the outer vapour membrane as intact the heat losses were not dependant on wind speed).  What else might I be missing?

I’m grateful to anyone who has made it this far and wants to weigh in!

 

 

 

Edited January 22, 2023 by DInwood
sp

[ReedRichards]

There are too many words for me to take it all in but I did a search on "weather" and got no results so I gather you don't explicitly say if you are using Weather Compensation or not.  Failure to do so would reduce your COP in milder weather.  What is the (maximum) output water flow for heating that your heat pump is targeting?  

[billt]

Difficult to read a massive block of small text, but if you want some real world data I recommend the Open Energy monitor site. https://heatpumpmonitor.org/

 

By and large these are well designed and installed installations so the performance should be good.

[DInwood]

The Ecodan doesn't use that word, it is however set to "curve" which adjusts the flow temp according to the ambient temp.  I think that is the same thing.  However, as I implied and with no apology for the number of words used, this an anorak's discussion of theoretical vs actual performance and we are talking about huge discrepancies, not the fine tuning of CoP in my view.

[JamesPa]

Not with ASHP (yet) but here are my load figures (heat demand at -2C)

 

As calculated by and MCS registered company using the MCS 'method': 14kW

As calculated using the same general methodology, but adjusting the U values where I know I have improved the insulation) and told the MCS assessor - but he ignored it):11kW

As measured during the extended cold spell in December, when it was -2+/- a degree or so for several days : 7.5kW

 

Says it all to me.

 

 

 

 

Edited January 22, 2023 by JamesPa

[RobLe]

While the electricity used by a heatpump is easy to measure, heat is not.  I recommend an external MID certified electrical meter on the heatpump (they’re only £20 for single phase) as a minimum, and consider external heat monitoring (more difficult).

 

The ‘insulation performance gap’ is worth googling; maybe there are details wrong?

 

While mvhr can be very efficient with ‘sensible heat’, it does not recover any of the latent heat of evaporation of water - so I believe every kg of water that evaporates indoors takes 0.68kWh of heat to get rid of, irrespective of the mvhr(anybody confirm?).

 

Once your home is that well insulated, user differences start to have more impact I think.  Curtains, doors, windows, how much solar gain, activity ?

 

Now is the time to borrow a thermal camera.  Get a little training/find someone who’s used it, look for any hot details outside, and/or cold inside.

 

 

Edited January 22, 2023 by RobLe

[JohnMo]

Nearly lost the will to live, reading the first post, and still didn't find any details needed.

 

Do you have a buffer or volumiser or neither?

 

What are your flow temps? Typical will do.

 

How many zones, TRVs/ thermostats

 

How do you actually operate the system?

 

Keep it short please.

 

[DInwood]

44 minutes ago, JamesPa said:

Not with ASHP (yet) but here are my load figures (heat demand at -2C)

 

As calculated by and MCS registered company using the MCS 'method': 14kW

As calculated using the same general methodology, but adjusting the U values where I know I have improved the insulation) and told the MCS assessor - but he ignored it):11kW

As measured during the extended cold spell in December, when it was -2+/- a degree or so for several days : 7.5kW

 

Says it all to me.

 

 

 

 

Thanks for this: I think you are saying your actual heat use is significantly *below* what you calculate using the MCS methodology.  In this, are you allowing for internal sources of heat as I described?

For comparison, on our coldest day so far (avg ext temp -0.7C) we used 23kWh of electrical power for heating (measured at the consumer unit supply to the Ecodan) and the Ecodan itself estimated 39kWh of heat output.  My method showed a net heat demand on that day of 34kWh & 10kWh power in. So one of the closest matches on heating but still poor electrical performance. 

I have no evidence that the house is haemorraging heat to the extent my analysis implies.  With no heating it cools down really slowly - consitently with the theoretical u-values & ventilation losses.  The architect who supervised the builder is pretty adamant the performance can't be that far off spec.

[DInwood]

1 hour ago, billt said:

Difficult to read a massive block of small text, but if you want some real world data I recommend the Open Energy monitor site. https://heatpumpmonitor.org/

 

By and large these are well designed and installed installations so the performance should be good.

Thanks.  I've just had a very quick glance at this and whilst it shows lots of data on ASHP performance I can't immediately see data addressing my issue of corresponding expected heat demand.

[DInwood]

1 hour ago, JohnMo said:

Nearly lost the will to live, reading the first post, and still didn't find any details needed.

 

Do you have a buffer or volumiser or neither?

 

What are your flow temps? Typical will do.

 

How many zones, TRVs/ thermostats

 

How do you actually operate the system?

 

Keep it short please.

 

If you've lost the will to live imagine what my wife has to put up with!  This post is really about the expected heat demand of a not-quite-passivehouse retrofit vs reality and whether the heat pump data can be trusted.

To your questions: buffer, flow typically high 20s to mid 30C. 12 zones, heatmiser room stats controlling 12 zones of UFH.

Operation: we have tried the all day constant temperature approach with very low flow temps but are currently experimenting with morning & evening operation with slightly higher flow temps (typ 30-35C)

[IanR]

2 hours ago, DInwood said:

I'm finding it is taking more than twice the power I calculated, a combination of lower than specified CoP and higher than predicted heating energy output.

 

I wouldn't first jump to the conclusion the ASHP Manufacture is cheating the test conditions under which its published (S)CoP figures are based.

 

I'd first look into where your energy losses may be higher than predicted, and the calc used for that prediction.

 

2 hours ago, DInwood said:

As a result of these calculations, I get a total house heat loss of 88W/C (and just under 12W/m²).

 

The 88W/°C, for a 157m² house is just inside the PassivHaus target, and very similar to my own PHPP calcs.

 

2 hours ago, DInwood said:

the old structure had insulation & airtightness added underfloor (u-value 0.16) & external insulation added & cavities refilled (u-value 0.2).  New flooring was slab on insulation u-value 0.14, new walls 0.15, flat roof 0.12.  There is lots of Norrsken triple glazing, average u-value 0.79 & some roof lights 0.6. 

 

Summary:

Floor old = 0.16

Floors new = 0.14

Walls old = 0.2

Walls new = 0.15

Roof 0.12

Windows doors 0.79

Air Tightness 1.1m³/m².h @ 50Pa

 

For my own property, floor is 0.11, walls 0.11, roof 0.1, windows and doors 0.68 average. The Air tightness figure used in the PHPP calcs was 0.6m³/m².h @ 50Pa

 

My reality is the house uses slightly less energy than predicted, which I can explain by making more of the Solar Gain than calculated and achieving a better air tightness than the figure that had been used in the calcs.

 

I'd look into how you've come up with the 88W/°C, since pro-rata, your losses should be around 50% greater than mine, but you are calculating them at around the same.

[JohnMo]

7 minutes ago, DInwood said:

buffer, flow typically high 20s to mid 30C. 12 zones, heatmiser room stats controlling 12 zones of UFH.

So how is your buffer plumbed, is the normal 4 port, any idea of the size. Is it controlled by a thermostat?

 

 

[DInwood]

39 minutes ago, IanR said:

 

I wouldn't first jump to the conclusion the ASHP Manufacture is cheating the test conditions under which its published (S)CoP figures are based.

 

I'd first look into where your energy losses may be higher than predicted, and the calc used for that prediction.

 

 

The 88W/°C, for a 157m² house is just inside the PassivHaus target, and very similar to my own PHPP calcs.

 

 

Summary:

Floor old = 0.16

Floors new = 0.14

Walls old = 0.2

Walls new = 0.15

Roof 0.12

Windows doors 0.79

Air Tightness 1.1m³/m².h @ 50Pa

 

For my own property, floor is 0.11, walls 0.11, roof 0.1, windows and doors 0.68 average. The Air tightness figure used in the PHPP calcs was 0.6m³/m².h @ 50Pa

 

My reality is the house uses slightly less energy than predicted, which I can explain by making more of the Solar Gain than calculated and achieving a better air tightness than the figure that had been used in the calcs.

 

I'd look into how you've come up with the 88W/°C, since pro-rata, your losses should be around 50% greater than mine, but you are calculating them at around the same.

Thanks IanR.  I'm puzzled why your average u-value is so high - have you got acres of glazing?  My figure of 0.23 is the average weighted by area for each element and feels about right.  Plus, as I said, my spreadsheet gives a very similar result to the MCS pack for the fabric losses.  I'd be very interested to know what your actual heating energy input & outputs are.

It could be I'm underestimating ventilation losses, but even if I double these it only adds 5% to the rate loss (ie I get to 94W/C), not enough to account for the discrepancy.

Edited January 22, 2023 by DInwood

[DInwood]

20 minutes ago, JohnMo said:

So how is your buffer plumbed, is the normal 4 port, any idea of the size. Is it controlled by a thermostat?

 

 

Normal 4 port 26l no thermostat

Edited January 22, 2023 by DInwood

[IanR]

4 minutes ago, DInwood said:

Thanks IanR.  I'm puzzled why your average u-value is so high

 

"windows and doors 0.68 average", not the overall property average. I wouldn't be anywhere near 0.23 for what you are calling a weighted average. Hence me saying you should look into how you are calculating your losses, as they calculate the same as mine, whereas they should be approx. 50% greater than mine. 

[DInwood]

20 minutes ago, IanR said:

 

"windows and doors 0.68 average", not the overall property average. I wouldn't be anywhere near 0.23 for what you are calling a weighted average. Hence me saying you should look into how you are calculating your losses, as they calculate the same as mine, whereas they should be approx. 50% greater than mine. 

My apologies, bad reading skills.  The MCS calculated fabric load is 1778W & my s/s gives 1698W, so I don't think my error is in the uvalue calcs.

From your comment "50% greater" I am clearly missing something significant. 

Edited January 22, 2023 by DInwood

[IanR]

8 minutes ago, DInwood said:

My apologies, bad reading skills.  The MCS calculated fabric load is 1778W & my s/s gives 1698W, so I don't think my error is in the uvalue calcs.

 

MCS completely over estimated my energy losses, PHPP was closer.

 

I'm surprised that you are not inquisitive about you having U values typically 50% higher (accept for glazing) than most PH houses in the UK, but your calcs are predicting energy losses equivalent to a typical PH house. Especially as your reality and calcs are not currently matching.

[JohnMo]

Having a buffer will knock about 1 of your expected CoP.  But doesn't account for all your excess electrical consumption.

 

Sorry for lots of questions.

How long is your heat pump running for is at least 10 mins

 

Are all the zones on at once, or at different times.

 

 

[JohnMo]

3 minutes ago, IanR said:

surprised that you are not inquisitive about you having U values typically 50% higher (accept for glazing) than most PH houses in the UK

Doesn't u value required tie in directly to the form factor. Two storey nearly a cube would need way less insulation than my house, long thin, and single storey with all sloping ceiling. My house to comply with PH heat loss would require u values all below 0.1.

[DInwood]

17 minutes ago, IanR said:

 

MCS completely over estimated my energy losses, PHPP was closer.

 

I'm surprised that you are not inquisitive about you having U values typically 50% higher (accept for glazing) than most PH houses in the UK, but your calcs are predicting energy losses equivalent to a typical PH house. Especially as your reality and calcs are not currently matching.

The MCS pack i was given was also wrong but only in its ventilation loss calcs.

Indeed this gets at exactly what I'm trying to understand. I accept I might be wrong and am very open to ideas as to why. If I'm underestimating expected energy demand, what am I doing wrong? Is it my method of estimating other sources of heat? My method was pretty good at predicting when the heating would be needed.  Is it ventilation load? Wind effects?

Edited January 22, 2023 by DInwood

[DInwood]

10 minutes ago, JohnMo said:

Doesn't u value required tie in directly to the form factor. Two storey nearly a cube would need way less insulation than my house, long thin, and single storey with all sloping ceiling. My house to comply with PH heat loss would require u values all below 0.1.

The area based MCS type calcs take this into account and we are pretty cubic

Edited January 22, 2023 by DInwood

[DInwood]

13 minutes ago, JohnMo said:

Having a buffer will knock about 1 of your expected CoP.  But doesn't account for all your excess electrical consumption.

 

Sorry for lots of questions.

How long is your heat pump running for is at least 10 mins

 

Are all the zones on at once, or at different times.

 

 

HP runs for good long periods and all zones have same time clock settings. Obvs some go off earlier than others though.

[JamesPa]

2 hours ago, DInwood said:

Thanks for this: I think you are saying your actual heat use is significantly *below* what you calculate using the MCS methodology. 

Yes

2 hours ago, DInwood said:

In this, are you allowing for internal sources of heat as I described?

No, neither the MCS calculation (upon which installers were relying to suggest a system design) nor mine (upon which I am more confident to rely, but modified significantly by the actual measurements) took into account internal heat sources, nor solar gain.

[DInwood]

11 minutes ago, JamesPa said:

Yes

No, neither the MCS calculation (upon which installers were relying to suggest a system design) nor mine (upon which I am more confident to rely, but modified significantly by the actual measurements) took into account internal heat sources, nor solar gain.

Aha. If I ignore internal heat sources the puzzle dwindles, with only a 10% excess (ashp stated output vs my calculated gross output). However it would then be horribly wrong in milder temps with heat demand weeks earlier. You really do need to allow for body heat etc.

[JamesPa]

1 hour ago, DInwood said:

Aha. If I ignore internal heat sources the puzzle dwindles, with only a 10% excess (ashp stated output vs my calculated gross output). However it would then be horribly wrong in milder temps with heat demand weeks earlier. You really do need to allow for body heat etc.

In a low consumption house definitely, in a leaky house less so.  MCS doesn't, to the best of my knowledge, allow for body heat etc, unless someone can correct me.