So, we all seem to agree that we need better models and empirical data.

Should the first step then be for all prospective ASHP owners to set their condensing boiler flow temperature to 57 deg.C, for example? Also to leave heating on at comfort and setback temperatures to run as a heat pump would when this winter hits?

If they have weather compensation already, then all the better.

Also boiler outputs can be capped in the settings, rather than letting them run up to maximum output.

This self-proof of radiator sizing and heat input, combined with a blower test and better guidance on surveying should get clients and installers on the same page pre-installation. 

Variations could then be formalised to cover all parties rather than a continual battle of a#$€ covering and finger-pointing.

Thoughts / variations......etc?

Totally agree, @jamespa.

Any installers viewing the spreadsheet calcs as gospel need to be re-educated. The MCS process that doesn't mention corroborating the spreadsheet calcs where possible needs to be changed. However, any opinions voiced that the spreadsheet calcs are rubbish so don't use them also need to be challenged, and that is precisely what I've seen you do many a time here.

@cathoderay's labelling of the spreadsheet calcs as "whatiffery" is akin to throwing the baby out with the bathwater. Those calcs are a reasonable starting point, so should not be abandoned or dismissed as a failure. They should, however, be corroborated or further honed wherever possible by other, as you say, empirical means.

As for the multivariate analysis you suggest, you may well be right. More importantly, I would imagine in the longer term the spreadsheet being replaced by a web application. That way, the model behind it could use that analysis and could offer assessors the ability to enter far more empirical data were it exists, only making assumptions where necessary. Something like that would, I suspect, provide a much more accurate result.

Posted by: @majordennisbloodnok

↑

In fairness, @jamespa hasn't said whatiffery fails. He went on to say...

In fairness, I didn't say he said whatiffery fails, it is me who said whatiffery fails. I am deliberately plugging this point because as frequent readers on this forum know only too well (a) I think we rely far too much on models, and not enough on empirical data, across the board, not just in heat pump sizing, and (b) particularly for older buildings, of which there are many, the spreadsheet based heat loss calculators are very possibly worse than nothing. An extreme position, perhaps, but it needs to be aired and debated. As @jamespa and others have noted, the spreadsheet calculators are not used for the customers' benefit, but to protect the installers. In effect they are a pseudo-scientific re-work of the old fossil fuel boiler idea that bigger is better. I could even say that the spreadsheets are the today's iteration of the installers' monsters of old, hidden behind that old refuge of the unscrupulous, cavernous numbers meaningless to man.

Posted by: @jamespa

↑

I would argue that experimental data is superior to model in most if not all cases.

Yes, we do agree, only I might change that to 'I would argue that good experimental data is superior to a model in all cases.'

It's about the primacy, which comes first. As long as the experimental data is sound (which is why I added 'good', and you covered the same point earlier in your post), then it is the best reflection of reality. That is why the data comes first (the primacy), and only then do I fit a regression line (which I accept is a form of model, but there is no whatiffery, unless and until I extrapolate, and yes, that is whatiffery). This is in contrast to starting with a model, and then trying to make predictions about the real world, which is exactly what the spreadsheet based heat loss calculators do.    

Posted by: @majordennisbloodnok

↑

However, any opinions voiced that the spreadsheet calcs are rubbish so don't use them also need to be challenged, and that is precisely what I've seen you do many a time here.

In fairness I think (well I hope) I have said that they can be rubbish (by which I mean can lead to design choices which have material and unnecessary implications for the householder). That is self-evidently the case on at least some occasions, as demonstrated by my example (and, it seems, the example presented by @cathoderay).  Furthermore I have heard more than one installer declare that they cannot and will not take into account fabric upgrades that they cant verify which suggests very strongly that there is a systematic (and material) misapplication of the method not just occasional random failures.

If I have in fact implied that they are always rubbish I apologise, but I will stick to my view that, badly done, as they clearly sometimes they are, they can be rubbish.  That's why I advocate using empirical calculations to sense check as a minimum protection.

To an extent we are arguing over semantics here, the key thing, as @larry has observed, is that we all seem to agree that we need better models and empirical data.  The problem is to convince MCS.

Posted by: @jamespa

↑

If I have in fact implied that they are always rubbish I apologise, but I will stick to my view that, badly done, as they clearly sometimes they are, they can be rubbish.  That's why I advocate using empirical calculations to sense check as a minimum protection.

Exactly the opposite, @jamespa. I have never seen you state or imply that. Everything I have said was meant in support of, not as criticism of, your approach and viewpoint. No apology needed; you keep on doing precisely what you're doing.

Posted by: @larry

↑

a#$€ covering and finger-pointing

Juxtaposing those two could lead to interesting results...

Seriously though, I think the following, which are not a million miles at all from your suggestions, might be worth pursuing:

(a) a general acceptance that spreadsheet based methods are what I call whatiffery, or in plainer English, they are fallible, and get more so the older and more irregular the building, which means a large proportion of the UK housing stock. They involve a lot of work (several hours) only to  produce too often a result that is just plain wrong. Just knowing that is important.

(b) it should be normal practice, whenever it is possible to do so, to use empirical data. It should become the norm. I accept there are practical difficulties, the main one being you have to have the system running in equilibrium to keep the analysis simple, but it may be that only a short period of running in equilibrium is needed. My n=1 example from the last week posted above, based on mild weather, managed to get pretty close the the heat loss based on data during cold weather, the Jan 2024 data. Since I have the data, I can do different analyses to see what works and what doesn't, as of course can others, and indeed I hope they will. I think @jamespa has already recorded some data from his system, and derived a data based heat loss.

The thing we haven't really touched on is how to get from a more accurate heat loss assessment to actual heat pump sizing. My current best assessment such as it is is that there is yet more chicanery at work here, principally in the comedy data produced by the heat pump manufacturers. For example, my house appears to have a heat loss of ~8kW, yet I need to have a nominal 14kW unit to heat it. How can that be?

Well, first off the 14kW disappears when the OAT approaches zero. On paper (in the manufacturers' small print engineering data) at these OATs I should get about 11kW. But I actually get around 8kW at -2 degrees OAT, which is either enough or just under enough to keep the IAT stable at design temp (another reason to believe the 8kW heat loss figure). I suspect all this has much to do with what in medicine is called efficacy and effectiveness, the former being what can be achieved by zealots in perfect conditions, the latter is what happens in the real world, where Gilberts and Biddies don't always do what they're told (which incidentally they are perfectly entitled to do). The point is, the real world has a curious habit of not matching up very well to the theoretical world. For heat pumps, it may simply be a case of obliging manufacturers to report real world output (ha ha very funny) or more realistically applying a sensible empirically determined de-rating figure. It might be as simple as knowing that a Midea heat pump with a nominal output of X will in the real world only put out 50-60% of X. It shouldn't be that hard to identify the relevant de-rating figure for other brands and models. Once known, they are easy to apply.         

And then finally yet another can of worms: radiator sizes, based on a room's heat loss. How do we go about getting that right, given the conflicting whole house heat loss results, and the fact the empirical method only gets whole house heat loss, not room by room heat loss? 

Posted by: @cathoderay

↑

The thing we haven't really touched on is how to get from a more accurate heat loss assessment to actual heat pump sizing. My current best assessment such as it is is that there is yet more chicanery at work here, principally in the comedy data produced by the heat pump manufacturers. For example, my house appears to have a heat loss of ~8kW, yet I need to have a nominal 14kW unit to heat it. How can that be?

You need to start with the detailed performance data not the sticker output.  The '14kW' Midea unit isnt 14kW at the flow temperature and OAT under consideration

If we assume that the detailed performance data is correct, then we need to consider system degradations.  A buffer tank/llh/PHE will most likely require that the flow temperature at the heat pump is higher than at the rads, by ~5C (more accurately the radiator deltaT in the case of a buffer/llh, or about 5C in the case of a sufficiently well plumbed PHE, but perhaps quite a bit less if its very well plumbed).  So for starters you have to look at the heat pump capacity at that elevated flow temperature not the design temperature for the radiators.

You then need to account for any loss in external pipework and, if it isnt included in the calculation, add 10% to account for 2hrs per day 'out' to heat DHW.

That should be good provided that the system is operated 24*7, the radiators properly sized and well balanced.  We need more evidence to decide if it is good

.

Posted by: @cathoderay

↑

And then finally yet another can of worms: radiator sizes, based on a room's heat loss. How do we go about getting that right, given the conflicting whole house heat loss results, and the fact the empirical method only gets whole house heat loss, not room by room heat loss? 

Yes indeed.  The obvious crude estimate is to scale the fabric spreadsheet method according to the empirical measurement.  So if the fabric spreadsheet method comes to 10kW and the empirical method 8kW, then multiply the figures in the fabric spreadsheet method by 8/10.  If you can, where you do decide an upgrade is necessary, err on the high side because you can turn rads down but not up. 

These guys are developing technology to make the measurements (and control the flow) on an installed system, on a room by room basis.  This could well make a 'suck it and see' approach good in many cases ie first install the heat pump and run it at whatever flow temp is necessary to heat the house.  Do this for a season collecting the results from the Adia technology.  Then, and only then, work out if there is a business case to upgrade the rads.  I would guess that the business case will be weak, or have a long payback time, in a pretty large number of cases - or, put another way, we are doing vastly more rad upgrades than is justified financially, and as a result inhibiting the roll out of ASHPs by making the 'conversion' more expensive than it needs to be!

OK, I'm going to go out on a limb and likely make myself look stupid. No matter; I'm old enough now to know I already look stupid, so just being honest is not a big deal....

It seems to me we currently have been discussing two approaches:

• The "official" heat loss calculation approach of measuring room sizes, making assumptions about fabric materials, adding the data into a spreadsheet and ending up with a guesstimate.
• An empirical approach of looking at data of energy historically injected into the property (or carrying out a controlled experiment to gather the data in real time) to try to work out the same guesstimate.

I'm wondering if a third approach of measuring the heat actually lost over a set time might be both feasible and practical, so I'm going to lay out the steps I'd foresee and see what you all think.

1. In preparation for a site visit, the assessing company ensures the house being assessed is heated to a stable single internal temperature in all rooms. If necessary, supplementary heating (fan heaters, for instance) are provided for the assessment day to deal with any cold spots without emitters.
2. An assessor arrives on site as with the existing heat loss calculation process. They turn up with tape measure or laser thingy in hand.
3. Digital temperature recorders (that the assessor brings with him/her as part of their professional toolkit) are positioned in the middle of every room in the house and the temperatures checked to verify the stable single temperature requested in point 1 is actually in place. Internal doors are closed to isolate each room.
4. Another digital temperature recorder is placed outside, shaded from any sun and protected from any precipitation, to measure the outside temperature throughout.
5. The heating is turned off and left off for a set period of time; I'd guess 4-6 hours.
6. Internal and external dimensions are taken so the volume of each room and the house in total can be calculated.

Given the difference between internal and external temperatures is known, the volume of the rooms and total building are known and the rate the internal temperature drops is being measured, wouldn't that provide a reasonably accurate idea of the thermal energy being lost over time?

From a practicality point, the assessment wouldn't take much longer than the current spreadsheet one. In fact, if two properties needed to be assessed in the same general area, the assessor could conceivably do both in parallel since measurements over a slightly longer time period would be preferable. The measuring equipment would be an extra investment but digital temperature loggers are not expensive things so the investment would not be significant.

Seems to me an approach like that would then be able to show heat loss per room as well as the house overall and therefore make it easier to size emitters for each room appropriately as well as understanding the likely size of heat pump needed. I suppose in theory such a heat loss calculation process could form the basis of a next generation EPC instead and then be used by installers instead of requiring them to do the calcs themselves; at least at that point we'd perhaps get an EPC we could rely on.

So, am I talking out of the wrong orifice or could this theoretically work? My suspicion is the former on the assumption that if it were that simple someone'd have started doing it already, but I have to ask since that is, as I say, an assumption.

Posted by: @majordennisbloodnok

↑

OK, I'm going to go out on a limb and likely make myself look stupid. No matter; I'm old enough now to know I already look stupid, so just being honest is not a big deal....

It seems to me we currently have been discussing two approaches:

• The "official" heat loss calculation approach of measuring room sizes, making assumptions about fabric materials, adding the data into a spreadsheet and ending up with a guesstimate.
• An empirical approach of looking at data of energy historically injected into the property (or carrying out a controlled experiment to gather the data in real time) to try to work out the same guesstimate.

I'm wondering if a third approach of measuring the heat actually lost over a set time might be both feasible and practical, so I'm going to lay out the steps I'd foresee and see what you all think.

1. In preparation for a site visit, the assessing company ensures the house being assessed is heated to a stable single internal temperature in all rooms. If necessary, supplementary heating (fan heaters, for instance) are provided for the assessment day to deal with any cold spots without emitters.
2. An assessor arrives on site as with the existing heat loss calculation process. They turn up with tape measure or laser thingy in hand.
3. Digital temperature recorders (that the assessor brings with him/her as part of their professional toolkit) are positioned in the middle of every room in the house and the temperatures checked to verify the stable single temperature requested in point 1 is actually in place. Internal doors are closed to isolate each room.
4. Another digital temperature recorder is placed outside, shaded from any sun and protected from any precipitation, to measure the outside temperature throughout.
5. The heating is turned off and left off for a set period of time; I'd guess 4-6 hours.
6. Internal and external dimensions are taken so the volume of each room and the house in total can be calculated.

Given the difference between internal and external temperatures is known, the volume of the rooms and total building are known and the rate the internal temperature drops is being measured, wouldn't that provide a reasonably accurate idea of the thermal energy being lost over time?

From a practicality point, the assessment wouldn't take much longer than the current spreadsheet one. In fact, if two properties needed to be assessed in the same general area, the assessor could conceivably do both in parallel since measurements over a slightly longer time period would be preferable. The measuring equipment would be an extra investment but digital temperature loggers are not expensive things so the investment would not be significant.

Seems to me an approach like that would then be able to show heat loss per room as well as the house overall and therefore make it easier to size emitters for each room appropriately as well as understanding the likely size of heat pump needed. I suppose in theory such a heat loss calculation process could form the basis of a next generation EPC instead and then be used by installers instead of requiring them to do the calcs themselves; at least at that point we'd perhaps get an EPC we could rely on.

So, am I talking out of the wrong orifice or could this theoretically work? My suspicion is the former on the assumption that if it were that simple someone'd have started doing it already, but I have to ask since that is, as I say, an assumption.

 

Without a doubt a measurement broadly of the type you are suggesting could be made, and indeed there are companies that will do it, albeit over a longer period of time.

The 'whole house' unknown which you don't mention is the heat capacity of the house.  That is generally significant (in fact will dominate) relative to the heat loss over a few hours, so any measurement of heat loss must be carried out over a much longer time so that the fabric itself returns to its starting state and thus the heating or cooling of the fabric can be ignored.  Otherwise all you are really measuring is something unknown about changes in the temperature profile through the fabric. 

That makes sense, @jamespa. Perhaps this is the kind of process that could be more routinely scheduled to coincide with people taking a holiday or a long weekend away or something. I'm mindful that not everyone can afford the luxury of a break like that and that that's yet another thing that might disproportionately disadvantage those less well off, but my guess is that the benefit of more accurate heat loss estimates is worth pursuing nonetheless.

Posted by: @majordennisbloodnok

↑

needed

There is a company that specialise in an overnight test on this basis. I don't actually know if they've experience of feeding this into the official method in an approved manner.

Worth investigating.

@majordennisbloodnok - as you well know, the general view here is the only dumb question is the unasked one.  

I too have thought about what measurements can be taken from the outside, because after all that is where the heat goes, but I came up against what to measure, and, for the things I could measure, what did they tell me? I couldn't get a satisfactory answer...

The method you suggest is certainly ingenious, but to my mind it over-complicates things. The other thing that occurs to me is that it is a measurement of a dynamic state, a house that is cooling down, and that further complicates things, because the rate of cooling depends very much on the temperature difference between the building and outside.

In contrast, my suggested approach uses an equilibrium state, the IAT just needs to stay pretty much the same, and in that state the heat loss must be the same as the energy delivered, give or take. I don't even think the whole house needs to be at the same IAT, all that matters is that everything is stable. This is easy enough to achieve with a heat pump, given the way we run them, and a fossil fuel system could I think be adjusted to do something similar, at something similar enough, but that is not much use for historical data that may well have come from timed heating. The problem here is the same as in your proposal, we are now dealing with a dynamic state, one in which the heat supply and loss are in a state of flux. The key question here it seems to me is whether averages can be relied upon. Is 10kW for 12 hours followed by 12 hours off the same as 5kW for 24 hours? That bit should be true, both are 120kWh, but the problem is the heat loss will vary, as the IAT changes, in the timed scenario. The question then becomes, can we use the concept of an average heat loss? Is the heat loss from 12 hours at 20 degrees IAT followed by 12 hours at 16 degrees IAT is the same as the loss from 24 hours at 18 degrees IAT? I know in practice it doesn't happen like that, with step changes, this is just a hypothetical question to consider the principle. In practice you would calculate the actual average IAT over the period.