I'd call that a heat store. To me, heat sinks just dissipate heat from
transistors and other things.
That depends on how fast you need to withdraw the heat, which increases
with the house thermal conductance and indoor-outdoor temperature
Dry soil has about 30 Btu/F-ft^3 of thermal capacitance and about
R1 per foot of resistance,
That's 36Fx2000ft^2x8'd = 17.3 million Btu. Is that enough? Too much?
A 48'x48'x8' house with 96 ft^2 of R4 windows with 96/4 = 24 Btu/h-F of
thermal conductance and 3744 ft^2 of R40 walls and ceiling with 3744/40
= 94 and 30 cfm of air leaks with about 30 Btu/h-F totaling 148 would
only need 1.8(0-(-3))148 = 800 Btu/h to stay 0 C indoors on an average
-3C January day in Toronto, so the basement might keep it from freezing
for 17.3M/(24hx800) = 901 days :-)
That's a good idea. If cloudy days are like coin flips, storing heat for
5 days can make a house 97% solar-heated. If it's 70 F for 12 hours and
60 for the other 12, it needs 24h(65F-27F)148 = 135K Btu/day, or 675K
Btu for 5 days. With lots of surface, a 48'x48'x8" deep floor with 12
$5 42"x48'x6" deep water-filled polyethylene film greenhouse air ducts
laid flat on the ground among 144 hollow concrete blocks under 4'x4'
plywood floor slabs with 1008 ft^3 of water and C = 1008x62.33 = 68.2K
Btu/F can provide that with no heat pump if the floor is 70+675K/68.2K
= 81 F on an average day.
My 1981 NRCC Solarium Workbook (one copy that was not burned by the
Canadian government :-) says 2785 Wh/m^2 (883 Btu/ft^2) of sun falls on
a south wall and 1321 (419) falls on east and west walls on an average
January day in Toronto. If equal windows on 3 sides transmit 50%, the
house gains 0.5x32ft^2(883+2x419) = 27.5K Btu/day. We can get the
remaining 135K-27.5K = 107K from A ft^2 of R1 $/ft^2 corrugated
polycarbonate Dynaglas south wall "solar siding" with 90% solar
transmission over a 100 F air gap if 0.9x883A -6h(100-27)A/R1 = 107.5K,
ie A = 301 ft^2. With $ R2 Therma-Glas Plus twinwall polycarbonate with
80% transmission, A = 220 ft^2. A solar attic could also work.
A $5 1000 cfm car radiator with its 2 fans in series (20-watts total)
could move solar warmed air down under the floor through a duct near the
slab center with a separate siding cavity air return duct. We could heat
the house by allowing floor air to flow up naturally through the same duct
and back into the floor near the perimeter, using a 2-watt motorized damper.
The car radiator could also heat water for showers with a $0 1"x300'
pressurized black plastic pipe coil in a 4'x8'x3'-tall 140 F box with
a folded 10'x14' EPDM rubber roofing liner.
Nathan Hurst measured 1000 Btu/h-F in water heat gain in Melbourne (see
That requires about 1000 cfm of matching air movement. He ran his fans
at 16 W total with PWM controllers. Nathan and I ran my car radiator and
2 fans in series from a 20 W PV panel at the PA Renewable Energy Fest
last September. They use 36 W from a 12 V battery.
Longer blades tend to be more efficient. I like Lasko's $0 2155A 20"
window fan, which can move 2470 cfm with 90 watts, ie 27 cfm/W..
With large ducts and low airspeeds (<400 lfm) and few turns, we can get
close to zero pressure. Grainger's $49 3C690 48" industrial ceiling fan
moves 21K cfm with 86 watts, ie 244 cfm/W. Their $21 4C761 60" fan moves
46K cfm with 105 watts, ie 438 cfm/W.
PhD Rich Komp (author of "Practical Photovoltaics") solar-heats his
Maine house with a similar "hypocaust" hollow mass floor (a concrete
slab over lots of hollow blocks) and a small PV-powered fan.