Thanks for the reply Biff, I am mystified about the inner layers and no I do not have a theoretical explanation. I do have some ideas about what is happening in the models but nothing I can prove, at least not yet. Unfortunately I only use the software, I have not yet grasped the heavy physics behind it, maybe one day!

I suppose the next step is for me to substitute something else into the middle layers of the modelled material to see how that affects the results. Paper sound like a good bet.

I am also quite happy to provide anyone with detailed information about which materials and other parameters I have used. It would be very useful to try this with different [Part L validated] software. Anyone out there who fancies a crack at it with something like IES?

Here is the result of the latest research simulation where I have included results for substituting the inner layers of multifoil with paper.

Hi Paul, I think it has to be a jpeg to work. This was rather a complex procedure! I imported the excel image into word, then converted the document to a pdf file. You can then import a pdf image into photoshop and save as a jpeg. Perhaps there is an easier way but I always seem to do things the hard way!

Doh sorry

The jpeg is of a chart on a full A4 page in a word document - perhaps that makes a difference? I just attached it using the browse function

I have been interested by this thread both because I am about to improve my loft insulation, and because of the physics
involved.

I found these links useful in understanding the current scientific basis for the performance of foil-based insulation.

http://www.timsa.org.uk/Multifoils%20statement%20May%202007.pdf
http://www.bre.co.uk/page.jsp?id=677
http://www.webdynamics.co.uk/media/LABC_technical_guidance_note.pdf
http://www.webdynamics.co.uk/media/BBA_cert.pdf

The data in the BBA certificate for Thinsulex ® are reasonable given the construction. The only thing I found surprising
was the surface emissivity of 0.4, which is higher than I would have expected. (No mention of mice!)

I think the “U-value” methodology for computing heat losses is part of the problem. This method tries to come up with a
single “thermal resistance” for an insulator that is independent of its surroundings. This works properly only when either
radiation is not a significant mechanism of transfer, or when the emissivities of the surroundings are always more-or-less
the same and temperatures are not too extreme. The real world is further complicated by free- and forced-convection, of
course.

For example, take a 25mm air-gap of perfectly still air (ignore convection). If you measure the “thermal resistance” of
this layer by the guarded hot-plate method, with plate emissivities of 1 (perfect radiators/absorbers) then for a hot plate
at 20°C and a cold plate at 0°C, you would measure a U-value of about 0.16m2K/W. If, however, the plates had emissivities
of 0.1, you would measure a value of around 0.79m2K/W. So the U-value of an air-gap depends on its surroundings, which is
inconvenient. The test does not measure a true thermal resistance – it is an “effective resistance” based on a combination
of conduction and radiation, and the “hot-box” method has similar characteristics. BR443:2006 would give a U-value of
between 0.34 and 0.50m2K/W for an air gap with “low emissivity” surfaces, depending on orientation. Presumably this is a
compromise based on assumptions about the surrounding emissivities and convection effects. All this suggests that the
margin of error is quite large if emissivities vary greatly.

For completely opaque insulation (in which even a very thin layer will block radiation, particularly in the 1 to 100um
wave-band), the guarded hot-plate method will give a true thermal resistance that is independent of the emissivities of the
plates. So this type of insulation is not very sensitive to its surroundings. (Fibrous insulation such as Rockwool is not
completely opaque by this definition, and will show some radiation effect – but I don't know how much).

This means that analysis beyond the simple application of U-values may be necessary when air-gaps and low-emissivity
surfaces form an important part of the insulation. At the moment, the work-around seems to be to test a complete structure
including all the air-gaps and surfaces. This is fine provided the installation does not deviate from the test article,
but if it does, the change in performance may not be easy to compute accurately.

But even Sparrow hasn't shown why the inner layers of foil are any better than, say, sheets of black paper.

But surely you can't write the software until after you understand the physics?

Sure, so the software adequately models the stone wall. The known physics of stone walls was, I presume, included in the software sufficiently well for the thing to work.

But what is the known physics of the internal layers of foil and how has this been included in the software?
Maybe you could ask your students (or perhaps even students over in the physics department) 'What do the internal layers of foil in a multifoil insulation do?'
And then post the replies :)

But that presupposes that the person who wrote the software does know how the physics of the internal foils works. Who is this guy? If the material behaves in a way that is fundamentally different to the way stone behaves then using the same method is going to result in garbage.

Fascinating article no doubt, but it doesn't seem to include the words mutifloil, foil, reflectivity, reflectance or even shiny. I can't see how a piece of software designed to model heat flow in stone would work for mutifoil unless the author knew something about how the properties of multifoil affect the system. Maybe they do and maybe it does. All I'm asking is how do the properties of the internal layers of foil affect the system? Would it, for example, make any difference if the reflectance of the internal foil layers was changed from the 95% to say 5%? And if so, why?

The number of foils needed in a multi-foil stack depends upon the conduction of heat between the foils and their emissivity. In space, where there is a vacuum between the foils, the U-value of a stack with low emissivity foils is roughly inversely proportional to the number of foils – hence the desire to use many-foiled stacks for really good insulation. On earth, conduction between the foils makes life more complicated.
I have done some sample calculations for the following situation. There is a multi-foil stack of constant total thickness 30mm. The (true) thermal conductivity between the foils is 0.03W/mK, and the insulation is presumed to be transparent. The stack has 25mm air-gaps either side, facing surfaces with emissivity of 1. The temperature of the hot surface was 20C and the cold one, 19C. The data below show how the effective U-value of this system changes with a) the emissivity of the foils and b) the number foils in the stack. Convection, boundary layers etc. are ignored in this calculation.
Case 1. Foil emissivities 0.4
Number of foils. U-value (W/m2K).
6 ___________________0.72
5 ___________________0.74
4 ___________________0.78
3 ___________________0.84
2 ___________________0.98
Case 2. Foil emissivities 0.1
Number of foils. U-value (W/m2K).
6 ___________________0.45
5 ___________________0.45
4 ___________________0.46
3 ___________________0.47
2 ___________________0.49
You can see that multiple foils improve the performance a bit if the emissivity of the foils is 0.4, but would have very little effect if the emissivity can be reduced to 0.1. In case 1, if the 4 inner layers of a 6 foil stack were replaced by “paper” with an emissivity of 0.9, the U-value would be 0.83 W/m2K, equivalent to a 3 foil stack. Note that if the insulation between the foils were totally opaque, inner foils would have no effect at all (ignoring their tiny thermal resistance).
I think dynamic effects (usually) can be safely ignored when it comes to choosing the insulation itself, provided water retention is minimal. Dynamic effects are controlled by thermal capacities, latent heat (if evaporation is relevant) and heat flows. Add-on insulation typically affects heat flows strongly but has very little thermal capacity compared to the rest of a building. Therefore, two insulation systems having the same effect on heat-flow (crudely represented by their U-values) should be equivalent as far as the dynamics of the overall building are concerned. Water retention in the insulation might be relevant, as this may increase its thermal and latent heat capacity to the point where it is not negligible any more, and also increase its U-value.

My previous post should have said "In space, where there is a vacuum between the foils, the U-value of a stack with low emissivity foils is roughly INVERSELY proportional to the number of foils...". Sorry.

Corrected in the original now. Thanks for pointing out the "edit" button Mike George!

Thanks Sparrow. Tell me if I've interpreted what you say correctly. In space (or the evacuated jackets used in cryogenic apllications) multifoils are the business. In buildings insulation, where the foil layers are separated by insulating layers of fluffy stuff, the internal layers of foil do b****r all.

Gosh we seem to be making progress!

>I suspect the fluffy stuff is thin enough, and transparent enough at infra-red wavelengths to be effectively transparent, so

That gives rise to the thought, would it be better if the fluffy stuff was a dull black? I don't suppose adding some dye to the polywhatsit mix would add much to the cost.

Oh look what has just been posted on Period Property UK:

"We are converting 2 loft areas into rooms as part of whole property renovation and building control have insisted on TRi Iso rather than Kingspan"

http://periodpropertyshop.co.uk/phpBB2/viewtopic.php?p=88155#88155