We have a barn in the Dordogne France currently being converted. We are just about to complete the roof which has been renovated and insulated to a good standard (300-500mm wool with 20mm pavatex board.

The next stage is putting in concrete flooring and we are looking at the insulation options. The most strightforward route would be underfloor insulation, however for 200mm of foam as example this will be expensive (floor area is 232 meters square). I have looked at just insulating the perimiter inside to a depth of about a metre and maybe using 300mm polystyrene. This would be more cost effective as the perimiter is arround 64 metres, and we would keep the thermal mass of the floor.

Where I am stuck is that Ii have no idea of how to calculate the effectiveness of either option (or a combination of the two). Are there any software tools I could use to calculate the insulation value of perimeter insulation or maybe someone has experience of the effectiveness of using perimeter insulation.

Just for information the finished floor will be 100mm concrete, poower floated and will use low temperature UFH.

Insulating the perimeter inside to a depth of about a metre to me implies digging down 1M and then putting in the insulation. It is generally NOT a good idea to dig below the bottom of the foundations. Are you sure that you can dig down that much without compromising the integrity of the walls?

If you use the perimeter insulation I would expect the UFH to have a significantly slower response time than with under slab insulation.

The UFH will run at a low temperature, arround 20 degrees Celsius in the winter. The hope is that once up to Temperature the floor will not vary by much due to the mass.
When I looked at this previously it was suggested that some under slab insulation be used (40mm) to slow the transfer of heat down. This raises the question of how much insulation and what affect this would have.

I am guessing that you could go lower than the bottom of the wall if done in small sections similar to unerpinning. The insulation would then become structural and the excavation properly backfilled.

But possible?

The concrete is the finished floor so am hybrid limecrete floor is not feasible.
Though I agree with the risk of moisture being driven out and up though the walls, in our case the ground is dry, the water table being 6 or 7 meters down, there is a well nearby and the water lever never changes.

Insulation can't 'become structural' except perhaps glass foam boards throughly packed around with weak concrete - anything else will 'give', offering no resistance to the extrusion (if any) of ground below the founds. Don't on any account go below the bottom of the stones that serve as found - except at a good distance from same - I'd apply a 30o plane downward from the found stonesand not go beneath that. If you inadvertently go too deep, make it back up with weak conc. In soft, extrudable clay I wouldn't even go half as deep as the founds bottom.

You have EWI on 600mm stone resting on the ground at its founds, and a conc floor within that. Whatever you do under the floor, whether over all or perimeter, you leave a massive conductor - the wall thickness - completely unguarded and leaking heat down and outward back up to surrounding winter ground surface. The EWI can never overlap or join up with underffloor insulation.
The best you can do is a trench around the outside, down as far as poss but not below found bottom, and continue the EWI down to its bottom. If that can't go deep, then a wide horizontal skirt of insulation, beneath the ground surface as well or instead.
The aim is to create a long path length that heat must travel, down through uninsulated floor around its perimeter, outward under or through the foundation wall and on outward under any external horizontal skirt of insulation before it can turn upward to the cold external surface.
The path length that the hear must travel uses the ground as theinsulator - it may be a poor insulator but in sufficient 'thickness' i.e. path length it can be an excellent insulator.

We are using internal wall insulatiom, effectivly the permiter insulation will be a continuation of the IWI.

Am I correct in assuming that the U value will be the conductivity of the soil at a distance equal to the shortesr path from the inside to outside. With the perimeter insulation extending down 1meter the shortest path would be 2 meters?

Posted By: fostertomInsulation can't 'become structural'

Nonsense - my entire house sits on EPS. EPS has specific structural data associated with it (think EPS 70, EPS 300 etc) and that is used by the structural engineer to certify the foundations.

I agree with what you say about not digging too deep though.

Posted By: nigelmWhere I am stuck is that Ii have no idea of how to calculate the effectiveness of either option (or a combination of the two). Are there any software tools I could use to calculate the insulation value of perimeter insulation or maybe someone has experience of the effectiveness of using perimeter insulation.

Just for information the finished floor will be 100mm concrete, poower floated and will use low temperature UFH.

There's a page that seems to have reasonable information:
http://warmafloor.co.uk/support-centre/u-values/

It says: "The formula for calculation of a ground floor U-Value is given in CIBSE Guide A3 and is very complicated." It then goes on to give a simplified formula. There are various U-value calculators online, some of which claim to deal with floor calculations. It's something I left to my architect, engineer and PHPP.

Posted By: tonyNo, the U value of the floor will be considerably better than that, you are talking about the shortest path, U value will be the average of all the different paths.

Tony is neglecting the fact that below the insulated perimeter the area across which heat travels increases as the heat flow lines diverge. The result of this is that the temperature at the bottom edge of the perimeter insulation will be colder than you would calculate using Tony's assumptions. How much warmer it will get further in at this depth depends to a great extent on the width of the floor.
I spent some time looking at this issue in detail a while ago, but I'm not sure I have the mental wherewithal to revisit it now. Best I can say is to advise caution - it aint simple.

Posted By: nigelmWe are using internal wall insulatiom, effectivly the permiter insulation will be a continuation of the IWI.

I missed that - so that makes sense.

Posted By: djh

Posted By: fostertomInsulation can't 'become structural'

Nonsense - my entire house sits on EPS

If you expect to remove heavy subsoil, replace it with lightweight insulation and have it resist subsoil extruysion (from under the found), the insulation has to be
a) more or less incompressible
b) packed tight with no gaps to take up, and
c) downloaded to prevent it 'floating' upward under powerful slow ground-extrusion - that's why I said foamglas (or, yes, structural grade EPS or other) but 'packed tight' with weak concrete. If held down somehow by the slab, that would be safe.
It's easy under slab but barely practical vertically.
But it's only dangerous if going beyond the excavation depths mentioned above.

Don't you end up heating a much larger volume of soil then though? Is it particularly costly to dig down a metre?

Posted By: fostertoma big block of subsoil slowly 'fills up' with heat, to give a new equilibrium temp which is very stable, winter and summer.

A big leaky bucket can be filled but requires the tap to be kept running to keep it full. Far better IMHO to put some insulation below a slab with UFH in it. Fabric first! Insulate, insulate, insulate! etc etc

Posted By: djhA big leaky bucket can be filled but requires the tap to be kept running to keep it full

True - but how leaky is the bucket and therefore how much top-up is required?

The leakiness is the conduction that remains via a long path-length of ground as insulator - and that leak decreases as the path-length grows - changing fast in the first year, slowing to near-stable re-equilibriation in the second year but in theory continuing 'forever'.

Posted By: mike7For there to be any merit in Tom's view of how it works, things would need to be on a large scale, with no serious heat sinks - eg a moving water table - anywhere near

True that a water table rising and falling through the subsoil, and/or a horizontal water flow through it, would remove the 'filling up' heat as it arises - but 'anywhere near' needs qualification.

It can be quite near without upsetting the mechanism - i.e. as near as the under-edge of the 'coffer dam' of edge-downstand perimeter insulation. If that near, would just mean that the 'improving forever' bit of the model wouldn't happen. The fuzzy boundary between thermally 'filled up' ground, and the 'unfilled' stone-cold ground beneath would come to rest at the top of the water-cooled zone, thus defining the path length of dry ground as insulator. How long that path-length is, depends on the geometry that can be attained in any given situation.

Note that a water table that doesn't rise or fall (or very little) is actually an asset, not a problem, adding thermal mass to the subsoil without washing heat away, although to balance that, soil wetness would increase the conductivity of the subsoil.

Posted By: nigelmin our case the ground is dry, the water table being 6 or 7 meters down, there is a well nearby and the water lever never changes.

sounds very good indeed, to me.

Posted By: gravelldTom, do you have measured observations for this?

Fraid not - it's a theory and starting from anecdote that underground house builders found, even or especially when in uninsulated ground contact, that their heating input (to maintain target temp) fell and fell year on year. Google 'PAHS passive annualised'. The whole subject is sadly unquantified.

Posted By: gravelldWould also be interesting to learn the differences for soil

Too right - I'm sure that there's a mass of uncordinated info out there, borehole logs etc, which would correlate underground subsoil and water conditions with observable surface features - it wd be onerous to bore/survey a site over several seasons to be really sure of its charactereistics in a way where all this cd be reliably pre-calculated.

However, two projects, variants on this theme, have proved very succesful, as reported by clients.

This (AGS) I think is about absorbing summer solar heat, penetrating the glass line, into the floor slab, which then warms the interior during winter. To achieve the latter, the slab has to be up to say 5K above target interior temp by the end of summer, in order to emit during winter. That over-temp would exagerate the loss to ground, even though it depletes by winter's end leaving the interior perhaps 5K below target temp.

What I'm suggesting (and is more like PAHS which uses remote solar collectors, not the windows as collectors) is a slab that's steadily warmed by the maintained interior temp, summer and winter, not so much by deliberately maximised peaky summer solar input. It's not intended to fully warm the interior in winter, more to maintain a pretty steady surface temp summer and winter, but still necessitating small winter fuelled input to the room.