Hi all

I'm a mechanical engineer with some building experience, who's trying to get into eco building, namely passive house.

I'm trying to do some heat loss calculations (ventilation & fabric) for a large 2 storey Georgian (internal footprint 160m2) triple brick skin house for current (leaky) in Suffolk, and post insulated scenario.

Wondering if someone can help me out? I have posted some general questions on another post, but thought this one deserved a seperate post.

How do you do heat loss calculations (fabric) for a solid stone floor

I came across this formula in on the forums: U = 0.05 + 1.65(P/A) - 0.6(P/A)², which is for a solid floor. It gave me 0.55W/m²K
It doesn't take into account the floor material or soil temperature which seems odd.
The floor in this house is solid stone.
Should I be doing the above calc, or a standard u-value calc involving the conductivity & thickness of the stone as I would do for a wall or window, and just add some factor for direct contact with the soil?

Many thanks

This touches on a really important but unresolved question - just how does subsoil behave thermally (heat flow straight downward)?

Complicated by the fact that much (or most) of the loss is in fact by conventional conductive loss in curving path outward under the perimeter and back up to outside air. That being the sum of numerous different-length paths starting from different points on the slab to different points on the ground outside.

It's the ratio between those two loss types, differing for ea building, that the P/A ratio attempts to address.

Sprocket, you seem to think that this would be a doddle with finite element analysis? Possibly in a 2D slice thro a typical building cross section (like Therm), but in fact it's a v 3D model, with the plan shape of the slab perimeter making a big difference.

This wd be mainly dealing with the perimeter conduction loss part of the problem - the straight-down loss still remaining a mystery, for which people (and e.g. Therm) make all sorts of approximations and assumptions. A serious study is badly needed.

First thing, is that long time-lag is v dominant. We're talking annual cycles of slow heat flow, in fact multi-years till an equilibrium is reached. The rule of thumb is that a 'new' temperature wave propagates at 1m per month. That raises novel, advantageous possibilities - like putting summer heat in over here, and it reaches where it's needed (e.g. an uninsulated floor slab) 6 months later. Pouring heat into uninsulated ground has been proved to work (prototype houses in UK-like weather zones exist) though our boffins immediately prove it will all leak away - but apparently doesn't. The role of ground water flow (if any) thro the ground also needs study - it can whip all the heat away, alternatively (e.g. Foster/Arup's Reichstag scheme) be the heat-transport mechanism. Water flow may be crosswise, or vertical (rise and fall of water table).

Posted By: JT101I'm a mechanical engineer with some building experience, who's trying to get into eco building, namely passive house

....

How do you do heat loss calculations (fabric) for a solid stone floor

I'm not an expert but as I understand it there is a standard - BS / I.S. EN ISO 13370: 2007 (Thermal performance of buildings. Heat transfer via the ground. Calculation methods) - for how to do such calculations. And of course if you are interested in Passivhaus, you do it the PHPP way.

This is an interesting question, and would it be similar to a concrete floor?

Solid floors in contact with earth are complex as has been previouly stated.

There is a useful table in the CIBSE Guides for uninsulated floors which you read the length and breadth and it gives you the U-Value. This has been collated from empirical data and is pretty reliable.

We used to use this before floor insulation became the norm under building regs. We still can use it for large warehouses where the edge effects are negligible over the entire footprint - the average U-value can meet current regs without insulation.

I'll post a copy or a link when I'm in the office tomorrow.

Thanks to all for the very informative comments. This has settled it in my mind. I shall use the formula as I stated , and just add insulation for post work calculations.

I was tempted to insulate the floor in the same manner as the Kiwis which is, I believe, to put vertical insulation in on the circumference of the building down a couple of metres and let the ground under the building reach equilibrium with it. Couldn't convince the BCO though.

Paul
That is probably because they can't work out the heat losses from a slab either.

The way I see it is as a ratio of absorption area (the floor in your house) and then the emitter area (the other 5 sides, or as many as there are). The ground can be thought of as a semi-infinite heat sink and adjust for rainfall.
If you work on the worst case, generally February ground temp, as your heating load diminished the ground temp is also increasing, so you would not be far out. Gets around flux variation problems.

So if we assume a 1 metre cube, then the absorption area is 1, the emitter is 5, so every face in contact the the ground 5m^2, is loosing energy, so 1/5th of the energy in. If we assume 2 of these cubes next to each other then the absorber area is 2, the emitter area is 8, so 1/8th. Once you have a list of different lengths you can plot the data. Then divide the abortion area by the emitter area and you end up with a curve of diminishing exponential growth. This is similar to the Cooling Law. By punching in some real temperatures and times (remember that Watts are units of energy divided by time and Wh are units of energy) you will end up with your losses for any size and shape of slab.

Edit: May be worth looking at heat sink formula to get a better idea of what is happening. Q(k)=-kA dT/dx

Chart below shows that as the slab gets longer the losses will get less as any material has the ability to insulate.

Thing that bothers me is:

The 'U' value is as FT says, an average of all the heat flow paths from points on the floor to the outside air. Dominated I s'pose by the flow paths that start from the strip of floor adjacent to the perimiter wall, as these have the shortest path to outside air.

So a floor with thick perimiter walls standing on it should have longer heat flow paths to outside, and lower floor U, than a house with thin walls, no?

Our place the walls are about 600-1000mm thick from inside air to outside air, by the time we include the redundant chimney breasts and IWI.

Not sure how/if the formula accounts for different wall thicknesses.

Is academic now as we just lifted solid floor and relaid with insulation under.

The wall thickness wd count as slab area, and you take external wall perimeter, because the base of the walls at slab level are pretty much same temp (and same exposure to inside air) as the slab. But it's an inexact estimate so such finesse is inappropriate. Maybe the 'perimeter' line should be half way thro the wall thickness.

What about modelling half a slice though the floor (from the centre to the perimeter) and then working out the U-Value of the floor as if it were a circle with the same surface area as the floor? Now I know that this also won't be accurate as the effects at corners and if the floor is a long rectangle, heat loss will vary wildly depending on which path it is taking.

But given all of that, how far off reality would it really be? In theory, a 3d finite element model should be able to model it realistically (as a computer could), but would this 2d slice method work well enough. Perhaps someone with 3d modelling software and a spare afternoon could do it and let us know? I would happily run the 2d versions and calcs.