Heat loss calculations

[Tim Lamb]

OK So I read the d-i-y Wiki article on heat loss calculations. Given my known mathematical limitations, I was soon confused:-) How does this relate to EPCs?

What I am attempting is choosing retrofit insulation for an existing flat roofed agricultural building undergoing conversion to an industrial use toilet block. The existing structure is 100mm solid concrete block *blown* together around 1970.

Current thinking is 150mm Rockwool between rafters and 75mm PIR foam pinned to the walls with screwed battens to take Versapanel (cement bonded particle board) rather than plasterboard. Not cheap but much cheaper than aquapanel! Heating by air source heat pump discussed elsewhere. Carpenters on site are keen on *Hybris* insulation by Actis. No fire resistance rating and won't do my walls.

The existing floor is poor quality 6" tiling on solid concrete. Could be ripped out and replaced with a degree of effort.

Ideally I need to check that the proposed insulation will meet current buildings regs. but with some thought towards meeting expected upgrades in future years.

Strictly, the building is outside the requirement for an EPC as the floor space is less than 50m2. However, I don't want local planning quibbling about such matters! Current minimum EPC is band E, but will change to band B from April

2030.

I have a *heat loss* calculator from the building regs. site. and the building dimensions. How do I proceed:-)

[Theo]

Building regs have a Standard Assessment Procedure, or SAP, for calculating heat requirements for buildings. That's what comes up with a particular kWh per year figure. To generate a SAP you need to input the makeup of every external surface (wall, roof, floor, etc) to work out the heat loss across that surface.

EPCs are the poor-man's relation to the SAP. Because lots of figures are unknown (eg what's under the floor, what your cavities are actually insulated with), and it would be too time consuming to do a full analysis, the EPC is a 'reduced data SAP' (RdSAP), which essentially means they just guess on various things. For example 'cavity walls insulated = tick' without inputting the exact materials of the walls.

Both of those give you a letter score for performance. The letters are based on 'SAP points' which is a 0-100 scale for impact. That's based on the bigger picture of energy demand - for example a more efficient boiler improves your SAP even though the heat losses are exactly the same.

I don't know what building regs will expect, but just to note the Stroma FSAP 2012 software is available so you can do your own EPCs:

this thread might be useful: Theo

[Tim Lamb]

OK Theo. I'll have a look this evening. The BR calculator simply wants areas and relevant U values and seems to give a *met* not met result.

Thanks

[Andrew]

Are the thermal building regs the same for 'commercial' buildings as residential ?.

Rather than 75 mm celotex on the wall, you could start with a layer of 50mm temporarily held in place with double-sided tape then fix to the walls with 30mm battens which can be conveniently made by cutting 95mm CLS timber into three. The width of the circular saw blade gives you a 30mm batten. Fix these through the 50mm first layer with suitable frame anchors (or ?concrete screws). Infill between the battens with 30mm celotex (I used Quinntherm which was dimensionally more accurate across its width than actual celotex). This gives you a nice flat surface which your inner layer of Fermacell or whatever is screwed onto.

The floor is the big problem. Maybe attack this first. If you repour a slab, use a perimeter upstand of insulation to create your perimeter insulation. The 100 page Irish Part L document has lots of useful diagrams.

[Tim Lamb]

Heated, occupied commercial currently requires an EPC of E. For a mixed use heated toilet, it seemed easiest to go with part L as I will be funding the heat!

The spreadsheet I found (for a domestic extension) calls for a flat roof U value of 0.18 which seems a bit extreme.

I don't think there is any saving on d-i-y battens and sawing C16 stuff finds the odd knot.

The fixing planned is *stand off* screws which allow the batten position to be adjusted once the screw is seated.

Indeed. The reason my head is jammed up with Lamdas and Rs etc. is to determine if the floor has to be redone.

[Andy Burns]

If it was new-build domestic it'd have to be 0.13

[Tim Lamb]

So ignoring chipboard, plasterboard and weatherproofing... 150mm of PIR foam?

[Andy Burns]

120mm for 0.18 U

[Andrew]

The three new houses being built near me seem to be using a lot of ?5 inch 'celotex', and the skip contained a lot of triangular offcuts of 12.5mm PB with 50mm expanded poly glued onto it. THis I presume is as a result of insulating the dormer cheeks (2.5 story houses).

[Tim Lamb]

Likely. I used 100mm PIR foam inserted between the uprights topped with

25mm PIR backed plaster board without upsetting BC:-)

My painful brain is recovering and my entered U value elements apparently meet the requirements of the Irish BR L1B (2010 edition) for a domestic extension!

However I have been unable to find a free version giving an actual overall U value.

There is a hint of one in the d-i-y wiki but not actually found:-(

Any suggestions?

[Andy Burns]

If you have multiple elements to "tot up", it helps to convert the individual U values to R values (take reciprocal of each) then add the R values and convert back to U (reciproical again).

[Tim Lamb]

Indeed. I finally got that far:-)

There are calculators available on manufacturers sites or paid versions or even restricted capability versions:-(

I need a spreadsheet that will allow me to enter the calculated U values for the floor, walls, roof together with their areas and come up with an overall U value for the building.

>

[The Natural Philosopher]

its trivial

for each surface take the area, and the R (=1/U) value, and multiply them.

Then tot up all the products, and tot up all the areas and divide the total of all the products by the total area.

Take the reciprocal.

That is the average U value of the building

[Tim Lamb]

Ok Thanks. So why didn't Google tell me that?

[Robin]

I've never had cause to delve properly into the requirements but the calculation TNP gives looks like the weighted method I thought ran for /extensions/ (L2B 4.7-4.8). I thought you were looking at a /major renovation/ so you were into 5.8 et seq.

But it's building control so assume the wind here is North-Northwest :(

[Tim Lamb]

I have yet to do the sums but accept that Andy and TNP have pointed up one of many shortcomings in my secondary school education.

I blame Jonnie Roscoe! For four years he taught me maths. Clever bloke in his demob stripey suits but not interested in teaching basics. I can tell you about dipping brass screws in lanolin for boat building, how to estimate maximum speed for a sailing craft from the waterline length and sail area, some of the methods used in constructing his scale model working steam locomotives but very little 2+2 basics.

[Andy Burns]

I wouldn't beat yourself up about it, yes with a bit of thought, you can do sums in metre squared kelvins per watt, but does anyone actually think in such weird units?

Then for building control purposes, extra levels of detail apply, so it's no longer just a sheet of PIR with R-value of blah.0, and a layer of plasterboard with R-value of 0.blah, but it includes screws od a certain diameter with x and y spacing of blah and R-value of 0.000blah

[Tim Lamb]

I have been attempting to relate my new overall U value to an actual EPC rating without success:-(

In my judgement, this building does not require an EPC as it is not intended for occupation. However it will need some background heating for frost protection at least. (do public toilets have heating?)

However, this is part of a much larger *change of use* plan so I am scratching for selling points.

[Andrew]

??

I still have the windows 98 version and it still works under windows 10 (but I have the 32 bit Pro version).

[Andrew]

Basically, a U values is the sum of the inverse of the R values of the individual components of the 'thermal element', i.e.U value = 1/R1 + 1/R2 + 1/R3 + ... 1/Rn

Better description here -