I don't think an 'interseasonal store' type slab has ever been practicable. When I was planning my build I did consider incorporating a 20,000 litre water tank inside the house, extremely well insulated and using a variable ventilation system to direct the heat losses either to the outside or inside of the house depending on the season. Inspired by things like https://jenni.ch/publications-448.html?file=files/jenni/inhalte/pdf/Publikationen/Energie_fuers_Leben_english.pdf&cid=2910

I agree that the floor slab is an important part of the 'thermal mass' of a house, and that UFH is a good way of incorporating a heat emitter in a building.

I don't think there's any question of 'charging up' the slab on a seasonal basis. You simply maintain it at a constant temperature during the year.

I presently heat up the 'thermal mass' of our house overnight using direct electric heaters, and that plus solar gains the following day are sufficient to keep the house comfortable. Control is pretty manual at present, relying on forecasts of how cold and/or sunny it will be for the next 24 hours.

If I was building now, I would put UFH in the floor slab and use a heat pump with some suitable tariff to keep the house warm. But it's not obvious to me what the best transition is from my current system.

Posted By: neelpeelI've often pondered about using a large tank, but I kept coming to the conclusion that a zero maintenance large slab was the better long term solution.

You need to consider the quantity of energy (heat) that you want to store.

energy stored = mass of store x specific heat of store material x temperature difference

energy required = heat loss rate x time

The big question is what temperature difference you can achieve/tolerate if you use the slab to store heat. With a water tank you potentially have about 80 K or so.

Posted By: djhInspired by things like [jenni.ch]

Yes, there are quite a few using that method in Switzerland, either for individual houses, or small apartment blocks, like this one featured on YouTube: https://youtube.com/watch?persist_app=1&app=m&v=D0wPnlQBVUM

Posted By: djhThe big question is what temperature difference you can achieve/tolerate if you use the slab to store heat. With a water tank you potentially have about 80 K or so.

I agree. The tank in that YouTube video can reach 95°C - certainly rather more than you'd want your slab.

Then the specific heat capacity of concrete is around 20% of that of water. you'd need 5x the volume of concrete to store the same quantity of heat at 95°C. Then to cap the maximum to around 30°C, you'd need to multiply that by around another 3 times - so 15 x the volume of water (actually more than that, due to the extra heat losses due to the greater perimeter).

When you get to those sorts of volumes, a 'better' alternative, if you're clear of the water table, might be to heat the ground, and to use a GSHP, rather than ASHP. Though that would need 2 separate pipe loops - one to pull the heat from the garden (or under the drive, which is likely to be hotter) to the store, another to pull the heat from the store to the house in winter.

just some very rough numbers to get an idea of the magnitude of this endeavour. I'm taking this as an inter-day store, rather than inter-season (you won't store enough to achieve inter-season I'm thinking)...

40m2 floor prob gives around 110m2 total of floor/roof/walls. Say Uvalue of 0.12W/m2K gives 13W/K.
Say another 10m2 windows at Uvalue 1.2W/m2K gives 12W/K.

So working with 25W/K heat loss. Taking zero degC outside, heat loss = 500W

The heat-degree-days info for your area, would let you work out how much energy you would need to keep the extension warm for a year (try this link to the met office HDD map)

https://climatedataportal.metoffice.gov.uk/datasets/TheMetOffice::annual-heating-degree-days-projections-12km/explore

Might have this clac'd wrongly, but if your HDD is 2200, and your heat loss is 25W/K, then you'd "lose"...
2200x24hrsx25 = 1320kWh each year.
Replacing that, your ASHP would draw say 330kWh. Taking average of 22p/kWh (assume 27p and 7p), then you heating cost annually for the extension would be = £66/year.

By setting up a system that pushed heat to the low level heavy mass at night, you might save 3/4 of that (night tariff only), let's say saving £45/year, with the risk that you might not be able to control the heat emission to the rooms very well, and indeed might be over charging the slab, which I suspect would be quite likely on a daily basis where our outdoor temps often leap 10-15 degK over night.

If you put 50tonnes of crushed stone under the slab (0.5m thk), the stone alone would cost you £1500, never mind excavation, filling, additional insulation, additional heat pipes, design, controls, etc, etc, and the carbon footprint all that would entail. That's a 33 year payback just for the stone purchase. Prob never recover the carbon.

I've likely screwed up the numbers somehwere, but the above might explain why not many examples exist??

Posted By: GreenPaddyThe heat-degree-days info for your area, would let you work out how much energy you would need to keep the extension warm for a year (try this link to the met office HDD map)

Ooh, thanks for that! Had no idea that data set was available.

@GreenPaddy
Not put off the concept of slab UFH in general, but...the more I think about it, the less it makes sense to have a radiator / UFH mix in my situation.
If I was starting from scratch for the house I think I would go for a slab with good thermal mass and UFH.
But given that the UFH only be for a small proportion of the house (just over 20%) then it would need to at a higher temperature and therefore could lead to overheating - and would be disproportionately expensive.

The ASHP works well with bigger emitters, which was leading me down the UFH path (as well as other UFH benefits of space saving and cosy feet), but maybe cleverly building in an oversized radiator in the new space makes more sense.

I already crank up the heating by 1.5deg overnight to utilise cheap rate and offset some of the morning usage until the PV kicks in. Works pretty well.

The new space will be insulated to building regs, so might need much less heating than we're used to. So not very big radiators at all, even at low flow temperatures.

GreenPaddy calculated 500W, which a couple of minimal £40 / 600mm radiators will supply at low flow temperature. They'll look pretty lonely in a 40m2 space! "Oversized" rads are only for uninsulated rooms with multi kW heat losses.



On heat storage: slabs and fills are good for daily heat storage but useless for interseasonal stores. If you are into numbers you can modify the daily Thermal Admittance method to work out how much a slab will store in a year. It depends on the heat capacity, but also the conductivity of all the layers which heat has to flow inward and outward through.

Basically if the slab is thick enough to store seasonal amounts of heat, then it gets so thick that the heat cant flow in/out fast enough.


EG, heat can flow around 100mm into most solid floor materials during a 12h-12h day-night cycle, there are thermal admittance websites you can play with this.

That stores about 0.1m x 2000kg/m3 x 1kJ/kg x 0.6 ÷ 3600 = 0.03 kWh/m2/K, so enough heat to gain and lose a few degree per day in a well insulated house, that's ok. Less if we use the walls as storage as well.

But if we wanted to store a 6month-6month summer-winter cycle of heat, we'd need much more fill (like 30 metres). But heat can only flow about 2 metres deep in that time, so we could not charge up the store fast enough just by heating it from the top.

Alternative solutions are to run layers of pipes every few metres deep throughout the store. Or use a very flat/wide store, like a garden or field (eg with a GSHP). Or heat the store to 100+ degrees. Or store lots of hot water (higher heat capacity, can be pumped). All are a bit more costly!

One thing I maybe didn't explain...the new space will be open plan to the existing living area (wall being removed). Also, the living area has an open staircase to the upstairs study, so the entire space/zone will be maybe 80-90m2.

So yes, the new area will be better insulated, but my plan was to oversize any new radiators/UFH to bring down the flow temps for the whole space and hence reduce the heat curve. UFH makes some sense in this regard, but I hadn't really done any costings as yet.
Also, regarding the potential overheating, there will some thermal gain in the space so I suppose a concern is that the UFH won't have the speed of response required to deal with this.

There is a book that reviews tank based systems across Europe written 20 ish years ago

Solar Heating Systems for Houses: A Design Handbook for Solar Combisystems

From memory they all needed additional energy input during the winter

The famous Drake Landing solar village in Alberta used pumps and pipes to store and retrieve heat at 60degC down hundreds of boreholes under a playing field. It's a good solution, but not the slab OP was considering. It worked well for a good while but unfortunately broke down and was too complex to repair, was finally converted to gas fired heating last year.

If you want to reduce heat coming into the building there are several possibilities:
(1) make the windows smaller
(2) use solar glass
(3) increase the shading
(3a) brise soleil
(3b) external shutters
(3c) external planting

Increased thermal mass generally helps moderate temperature peaks

the slab would normally be 125 to 150 thk, though a smaller room could get away with 100mm. Struct Eng'rs usually spec A142 steel mesh, but I use A193, as it's more robust to walk on for adding UFH pipes and pouring the conc. That should be flat enough to glue wood flooring directly to, if the surface is floated. Have done this lots of times on projects over the years. Be aware, you do need to allow the misture content of the conc to drop, before applying the floor fininsh, though you can apply a liquid DPM, to retard the misture liberation, and lay the floor sooner.

Half a day to DPM/Insulate/mesh, couple of hours for fit UFH pipes, then a couple of hours to pour conc, with a follow up float of the surface. Quick, simple, get on with the rest of the project.