I remember reading somewhere [can't remember where] that using steady state calculations; the maximum heat loss through a ground bearing slab is predicted to be around 10% of the total for a building. Obviously this varies with the introduction of insulation. Is the 10% figure correct? If so how much of a saving is made by the introduction of insulation and how quickly does the saving diminish as u-values are reduced? At what point do steady state calculations predict that the otimum level of insulation will be reached?

Anyone care to have a stab at some ball park figures?

Optimum: the level [either finacial or environmental cost] at which increasing the thickness of insulation becomes unviable

Steady state: calculated according to the methodology used for u-values, IE by reference to conductivity of materials as tested in a hot plate under controlled conditions.

I'm not a traditionalist either.

Insulation will slow the passage of heat but will not set up temperature differentials per say. Thus heat lost through the under slab insulation will warm up the sub strata and once this has happened the insulation is doing nothing ( this statement is not entirely true but very nearly ). Once warm under the insulation then it cant do anything much of use? can it We are back to steady state theories again?

The above argument does not apply to walls or roofs where the heat is lost to the air surrounding the building and is quickly transported away. There is a substantial temperature differential across a wall but under the house there is not rather a gradual temperature gradient and much slower heat flows result.

This only applies to ground bearing solid floors.

The thing is that the stuff underneath will get a lot warmer too indeed in steady state it will get to very nearly the same temperature as the water in the pipes.

Get out of that!

Posted By: bayouboyunderfloor heating in a solid build up. There the temperature differential is much greater and you are losing heat directly

Posted By: tonythe stuff underneath will get a lot warmer too indeed in steady state it will get to very nearly the same temperature as the water in the pipes

Not unless the stuff underneath is separated from the outside air by a long path length of adjoining 'stuff', either by downstand insulation, e.g. cavity insulation carried down to top of strip found. Failing that, the stuff underneath will be sucking heat most efficiently out of the pipes and wasting it, ultimately to outside air.

At 1m per month travel-rate of the temp wavefront through the ground, multiplied by what, 6months? of heating season, that's a path length of 6m, before the travel reverses, as the slab heating stops and the ground surface becomes warmer than the slab. That's 3.5m of downstand insulation below slab top (heat travelling down, under the bottom of the insulation and up outside of it to just short of ground surface)! That's why it looks easier to forget downstand for this application, instead do a 6m wide horizontal skirt below topsoil level, around the building's perimeter (or 'umbrella' as the American call it - because they incorporate a membrane along with the insulation, to keep the subsoil dry, preventing rain/snow melt from percolating down, taking the stored heat away with it).

Has there been any monitoring under an uninsulated slab? It would be good to know if this had been proven.

Only other things that come to mind are:

Is Fostertom's comment about the ground temperature reversing right? I was not expecting the ground temp to change at the depth below the slab much during the year. Ambient ground temperature below the topsoil level here in Orkney is about 7C all year round. The temperature in the house is unlikely to drop below 19 most of the year, so I am not sure this reversal can take place.

Second item is about whether the ground is indeed a solid mass. With ground water movement I assume there must be a lateral transport of water. I expect it isn't much as the velocity is likely to be low, but given water has a much greater specific heat capacity than air I would have thought that this might be an issue.

Finally: Well done for posting this topic up. Checking on 'common wisdom' is always refreshing, and occasionally illuminating.

Some papers with research on ground under slab on grade, see page on floor properties.

Very interesting links Jeff, found the typeface hard to read on the second one though.

In the first paper the authors have taken the temperature of the ground at one metre below the slab to be the average seasonal ground temperature. This is WRONG in my opinion and will lead to higher heat losses than actual ones. There are lots of reasons that it is not right.

If we cant get the models right what hope is there for getting it right in the real world?

Temperatures under the house will not be the same as average temps measured away from buildings.

Therefore the heat losses through the floor will be less than the model predicts.