In my topic Modelling the "Chunk" Heating of a Passive Slab, I discussed how I used a heat flow model to predict how my MBC WarmSlab heated by UFH + Willis heater would perform. What I wanted to do in this post is to provide a “6 years on” retrospective of how the house and slab have performed as built based on actual data that I’ve logged during this period, and to provide some general conclusions.
In this, I assumed 15 mm UFH pipework, but we actually used 16mm PEX-Al-PEX pipework with an internal diameter of ~13mm. At a nominal flow rate of 1 m/s, say, my three pipe loops in parallel have an aggregate flow rate of 0.4l/s or 1.4 m³/hr. At this flow, a 3kW (2.88 kW measured) heater will raise this stream temperature by 1.7 °C. However, when I commissioned the system, I found setting the Gunfoss manifold pump at a high setting (roughly equivalent to this flow rate) gave a very noticeable circulation noise in the adjacent toilet, so I tried the pump on its lower settings and found that the flow was almost inaudible on lowest one with in to return delta at the manifold still only about 5°C, so I stayed with this. The actual as measured delta for two loops of 4.9°C and the third slightly shorter loop of 4.1°C (close enough not to bother balancing the flows out). This corresponds to an actual flow nearer to 0.4 m/s or 0.56 m³/hr by volume. When scaled to adjust for this lower flow rate, the actual measured temperature profiles are pretty close to those modelled.
I measured the actual Willis heater’s heat input as 2.88kW. In analysing the actual slab heating rates, I found that this raises the overall slab temperature by some 0.45 °C / hour after the initial start up. Plugging typical specific heat and density figures for the concrete, this is empirically equivalent to heating 25 tonne of concrete (Cmass = Ewillis/ΔT/SIconcrete = 2.88*3600/0.45/0.9 kg), or 10.6 m³ concrete by volume (23000/2400 m³).
In the case where the Willis provides heating for the full 7 hour off-peak window (just over 20 kWh), at the end of this heating period the flow input to the slab is +9 °C above the initial slab temperature and the flow return is +4.4 °C. The temperature of the concrete immediately in contact with the pipe will follow this same gradient, with this temperature excess decaying radially away from the pipe centres. By the end of this heating window at the slab surface, there is barely a noticeable difference in the measured temperature of the floor above the out and return UFH pipe runs (perhaps 1°C). These temperatures and gradients are also comfortably within the reinforced concrete’s design parameters. As soon as the Willis is turned off, the internal temperature gradients start to flatten and any unevenness redistributed across the slab; the rebar reinforcing has a thermal conductivity 60 × that of concrete and this accelerates this, so that within an hour or so of the heating turning off, the overall slab is left about 3.1 °C warmer than at the heating start time (actually about 10% less than this, as the slab has already started to dump heat into airspace).
In my original modelling topic, I mentioned that my passive slab has ~73m² of concrete 0.1m thick (~ 17½ tonne of concrete with another ~10 tonne of perimeter beams, cross bracing and steel rebar, with the UFH runs laid in 3 × ~100m long standard “doubled back” spirals (common to most UFH designs) on ~150mm centres and roughly 50 mm below the slab surface. (Actually only 75% of the slab is covered by the UFH runs, because of the need to avoid proximity to ring beams, partition walls, areas under fitted cupboard areas, etc..) Nonetheless, this empirical 25 tonne figure is still consistent with the total volumetric 27½ total estimate if we assume that the rebar is effective at spreading heat through the wider slab over this multiple hour timescale.
In conclusion, based on this modelling and observation:
First recall our context: our house is near passive in class with a lot of internal specific heat capacity. We only need about 1kW overall heater input in the coldest winter months to maintain overall heat balance, e.g. either by a resistive heater such as a Willis or an ASHP.
IMO, there are two extreme approaches to house heating: (i) “agile” tracking of occupancy patterns so the living spaces are only heated when and where occupied; (ii) a 24×7 constant comfortable temperature everywhere within the living space. Our warm slab design is very much optimised for this second case, and our slab supplier did a good job in designing an UFH layout to match the slab characteristics to this
- The slab is covered in “doubled back” spirals with each loop using up a full 100m roll spaced on roughly 150 - 200 centres (and avoiding partition walls and cupboarded areas) so that each heats roughly 15 - 20 m² slab. In our case three loops were enough, and there was no advantage in trying to squeeze in a fourth.
- Our 3 loops will happily take up 3 kW heat input. Circulation speeds between ⅓ - 1 m/s seem to work well, with the only real difference being the slower the flow speed, the higher the delta between in and return temperatures.
The slab does just as its trade name suggests: it can be treated as a huge low temperature thermal store, but because of its extremely high thermal inertia, one that is not rapidly responsive to heat input. In our case, a heat input of 3 kW input will only raise the slab temperature by 1°C over a couple of hours, and radiating 1kW will drop the slab by only 1°C over roughly six hours.
In a true passive class house, one key to heating economy is the high level of thermal insulation coupled with a substantial internal heat capacity. Trying to drive such a house in an agile manner is a fruitless exercise, so forget the traditional having room-specific thermostat control; forget having traditional time-of-day heat profiles. It is far easier to treat all ground-floor rooms as a single thermal zone to be kept at a roughly constant temperature.
- In my view, using a resistive heating approach (such as a Wills heater) as well as an ASHP can both work well. In this second case something like the 5kW Panasonic Aquarea ASHP would be a good fit as it uses a modulated inverter compressor so it can heat the slab directly without needing a buffer tank. The choice is a trade-off between running costs vs. installation costs. In our case, switching from a Willis to this type of ASHP would save me about £600 p.a, in electricity cost, so I would really need to do the install for a net £ 3-4K to make the investment case feasible. However I would like to defer this discussion to a separate thread because there are other issues that such an approach would need to address.
Edited by TerryE