Posted by: @ianmk13

↑

My understanding is that the relationship between heat loss and the heat transfer coefficient is considered to be linear for our circumstances.  Although my gas consumption is broadly consistent with the house heat loss as determined by the usual U-value spreadsheet methods, I have been concerned that I cannot get my condensing boiler to run without producing plumes of steam, even with return water temperatures below 50°C.  My measurements suggest a heat loss for my house approximately two-thirds of that determined using heat loss spreadsheets.

What have I missed?

What is your concern triggering the question.

I am impressed at trying to determine individual rooms, I was content to determine whole house by measurement and use this to 'calibrate' the individual rooms in fabric spreadsheet 

FWIW my house loss determined by gas consumption is 7kW vs 10.5kW based on the spreadsheet method - I put the difference down to ACH.  So Im not surprised by what you have found.

In terms of 'what have you missed', thermal mass matters for changes in OAT as well as for changes in IAT and you are ignoring this (I think).  I once did a 'best fit' exercise playing with two thermal masses, one affected by IAT and another affected by OAT.  The OAT effect was about one quarter the size of the IAT effect, but this was one house only and I suspect it may be heavily dependent on construction.   When I did my whole house measurements, where no account was taken of thermal mass, I found that 12 hr offset gave best fit - a sort of crude correction for thermal mass. 

Posted by: @jamespa

↑

What is your concern triggering the question.

I'm simply trying to determine  a more accurate heat loss figure so as not to oversize an ASHP. I didn't want to rely on my gas consumption as a 'sanity check' for the reasons stated.

Posted by: @jamespa

↑

I am impressed at trying to determine individual rooms

It would be impracticable to use my mmeasurement method otherwise.  Too many valriables. There are enough measurement uncertainties as it is.

With regard to thermal mass, I haven't ignored it as such.  I've done my best to ensure stable conditions and all I can do is hope for minimal changes during my measurement period. I believe that a degree or two of 'measurement error' will be swamped by the much larger IAT-OAT value, so I feel confident in ignoring that.  When I started I made the mistake of increasing my room temperature for testing then realised that was bleeding energy into adjacent rooms. Where you are considering thermal mass, I think in terms of thermal inertia.  The external walls will have some effect but my test method removes internal mass from the equation. I have tried correlating gas consumption against external temperature over several 24 hour periods but drew a blank. Other than obvious major effects I couldn't see any time offset I could apply to my measured data.  If the time constant for my home is greater than, say, six hours then I'm not going to get any convincing data from my endeavours. All I will have is my 'gut feel'.

Posted by: @ianmk13

↑

I have tried correlating gas consumption against external temperature over several 24 hour periods but drew a blank. Other than obvious major effects I couldn't see any time offset I could apply to my measured data.  If the time constant for my home is greater than, say, six hours then I'm not going to get any convincing data from my endeavours. All I will have is my 'gut feel'.

Firstly, your method may well work, Im jolly impressed and please continue to post to inform us all!

In response to your comment above

a) I obviously don't know the time it takes your house to settle down but mine well exceeds a day (I know this from close monitoring of a week about 2 years ago when it was -2 all day and night for several days running).  I have a 1930s house originally with solid walls, that has two modest modern extensions and piecemeal fabric upgrades (loft insulation and partial and very variable internal wall insulation).

b) the scatter in my case over 'several 24 hour periods' was certainly far too large for a determination.  I ended up using 2 years of data, although one year would have sufficed.  

in case it helps here is the data I persuaded my installer (and one other) to use to 'calibrate' the GIGO spreadsheet method.  This is the aggregate from two 'seasons'.  Points on the consumption vs degree day chart represent a day, but of course several days with similar figures may be superimposed.  Note the quite considerable scatter, easily sufficient that just taking a few days really wouldn't have 'done'. 

Bear in mind that this measured data for my house shows a loss of 7kW, two 3 hr surveys (one of which I paid £300 for!) came out with 16kW, and a fabric spreadsheet completed by me but with the 'default' MCS assumptions came out as 10.5kW.

I hope that might give you some points of comparison.

@jamespa 

My 200m²property was built in 1990. I have increased the loft insulation, replaced the original double-glazed patio doors with lower U-value double-glazed bi-folds and taken the opportunity to add extra internal insulation on the northern kitchen wall during a re-fit.  My own fabric spreadsheet with default MCS assumptions came out around 11.5kW but, as you will know, this is greatly influenced by air change values. I paid a modest amount for a desk-based Heat Engineer assessment (based on information I supplied) which came out with a similar figure but this also takes historic gas consumption into account, I believe. My own gas consumption versus degree-day analysis provided similar results but I am highly sceptical of my condensing boiler's efficiency.  My suspicion is that the heat loss is more like 8kW and, despite the lack of clarity and repeatability in my measurements, I think they generally substantiate my view.

One concern I have with my measurement method that I have not mentioned is that the energy data from the Shelly power meter and the temperature measurement device controlling the fan heater are both updated once per minute (polled by HomeAssistant). I don't expect any artefacts as a result of asynchronicity of data but I do believe that this may add significant noise or scatter.  Typically, the heater could be on for 10 minutes then off for six. Noisy data is more difficult to process for shorter time periods (e.g. if the OAT isn't behaving nicely).

Regarding the thermal mass or thermal inertia issue, I have just tried plotting my half-hourly gas meter reading alongside OAT for the past couple of days.  By eye, if I had to identify a correlation, I'd say that the gas consumption lags OAT by around 2 hours.

Posted by: @ianmk13

↑

but I am highly sceptical of my condensing boiler's efficiency. 

Thats definitely a reasonable suspicion.  The initial results fom my heat pump suggest my condensing boiler (which was operated at FT50/55 so should be condensing) might transfer to the house <90% of the declared calorific value of the fuel used, whereas in principle this figure could be as high as 110% (gas calorifc value doesn't include latent heat of condensation).  I need more data to confirm and of course I don't know how accurate the 'heat out' measurements are from my heat pump (I have seen it stated that Vaillant heat output is an understatement).  I cant comment on the lag other than it seems short, but as you are keeping the IAT constant maybe not actually.  

@jamespa 

I know my boiler flow and return temperatures.  If I knew the flow rate then I assume I could determine the heat energy transferred to my home.  I believe it's simple fixed head/speed pump.  I may be able to monitor when it's running from the 'call-for-heat' control or similar (although I know there are occasionsl situations when the boiler doesn't necessarily fire-up). Presumably I could find the flow from the specification sheet (if I could find it).

That's correct.  You probably will need to measure flow rate though,  spec sheet will likely just be a pump curve.