pnw solar houses (was re: ground-moderated inlet air)
11 mar 2000
larry brasfield wrote:
>...i am designing a (nearly) passive solar house... getting to
>100% here in the pacific northwest is quite a challenge.
where will it be? december is the worst-case month for 100%-solar
house heating in eugene, or, with a 30-year average 41.1 f temp and
460 btu/ft^2 of sun per day on a south wall, ie 460/(70-41.1) = 15.9
btu/degree-day, vs portland with 15.8, seattle with 14.2, and olympia
with 390/(70-38.2) = 12.3 (much tougher than phila with 26, el paso
with 50, or brownsville with 90.)
on an average december day in olympia, 270 btu/ft^2 of sun falls on
a horizontal surface and 110 falls on a north wall, with 170 and 180
on east and west walls. nrel says an e-w horizontal-axis concentrator
can collect at most 190 btu/ft^2 per day, less than an hour a day of
direct beam sun, on the average.
>here are the principal thermal features:
>- octogonal, 2 stories, 16 f facets, 2400 f^2 total
something like a 40' diameter cylinder with 1,236 ft^2/floor and 8
16x16' walls? (there's an old hexagonal building nearby. these are
apparently rare, only three in the world, according to the newspaper...
there's also an old octagonal schoolhouse. maybe these are more common.)
>- earth bermed on 5.4+4.6 (upper+lower) facets
>- 615 f^2 earth roof, 620 f^2 (plan view) tile roof
or a domed plastic film roof with a shallow pool underneath?
>- windows on 3+1.7 facets s,se,sw, .6 facets n
i don't understand this facet (area?) numbering,
and how and how much the berming helps thermally.
>- clerestory windows facing s and se, 85 f^2 total
>- double paned, low-e, air filled glazing facing s
>- triple paned, low-e^2, air filled glazing facing n
what are their u-values and solar transmittances?
>- reinforced concrete walls as internal thermal mass
about 3x256ft^2, 8" thick, with an 8" 1,200ft^2 floorslab, totaling
1968x8/12x25 = 32.8k btu/f of thermal capacitance with a series
resistance of about r0.2x4" = r0.8 for the concrete plus r0.67
for the inside air film, about r1.5 together?
>- plenty of insulation underneath, around and over
>- tight construction and heat recovery ventilator
say 0.2 natural air changes per hour, ie 64 cfm?
and a 100 cfm hrv with 60% efficiency?
and 300 kwh/month of electrical use?
>- most solar gain taken on moderate mass concrete floor
direct gain. (ackackack.)
>- solar floor and walls plumbed for heat flow control
hydronic walls too...
>- greenhouse/airlock on most used, informal entry
>- seasonally deployable airlock wall on formal entry
these don't save much energy in houses, vs department stores, although
they can make a nice whoosh. (ever tried to get out of a big pipe organ
air chest with a 5" h20 blower? :-) if 5 people each enter and leave a
2-story 70 f house 10 times per day and hold a 3'x6' door open 10 seconds,
and it's 40 f outdoors, and during that time 16.6(3'x8')sqrt(16'(70-40))
= 6,500 cfm flows, reheating the air takes about 8,100 btu, a fraction
of typical daily house heating. very small airlocks might avoid that.
>- movable insulation/shading with automatic controls
what brand of movable insulation? how much per square foot? measuring
r8 altogether, including air leaks around the edges? you describe a
shutter that folds horizontally and parks above the window. is there
a commercial product like that these days? shurcliff's 1980 book only
lists about four automatic systems.
>- masonry wood stove with plumbing for backup heat
later, maybe... :-)
>- small lp fueled dhw tank for lazy/sick backup heat
sounds great, if hardly used. then again, some say a "solar house"
is "one with no other form of heat." these are more honestly solar
heated than psic's 30% "solar houses" or those built to fit northern
hemisphere lots with big picture windows facing north.
>> four people x 15 cfm/person (ashrae) only adds about
>> 60 btu/h-f to a house...
>i figure to make this work with .35 ach, which for
>this house is about 50% more air than your figure.
that's 0.35x2400ft^2x8'/60 = 112 cfm.
>i believe i'm working against a code requirement.
i wonder which one, and what it says.
(i recently helped a woman who could not afford a lawyer argue
in court with a township official who said the 1990 boca code
required a homeowner to submit a pe analysis to repair a roof,
after a fire. he pointed to a boca section that said the code
official was allowed to hire a pe at township expense if he
wanted to (but not for residential-class property.)
the code official went on to say that the homeowner had to reapply
for an occupancy permit, citing a boca section that said the official
must either a) issue one or b) give reasons why not, if the owner
should choose to apply for one.
this poor stubborn republican middle-aged ex-marine repaired her roof
in spite of the township, having paid for a pe study and gotten the
permit they demanded, and the township now wants $600 per day in
fines since december, as well as their own solicitor's fees for this
their boca criminal prosecution and their upcoming civil suit based
on her same actions and their zoning code (in case they lose the boca
case), because she lived in her house while doing the work, and didn't
apply for another permit (with pe study) for siding replacement.
needless to say, she's suing them back for trespass and turning off
her electricity, inter alia.)
>on average, according to my calculations, the loss due to ventilation
>air in january... is about 30% of my average solar gain. with the hrv,
>i can get close to 2/3 of that back.
this might add about 64+2/3(112-64) cfm = 96 btu/h-f to
the overall conductance, which seems like a lot to me.
>the house has about a 5 day time constant due to its high thermal mass,
>but i am concerned about week long extreme cold spells...
cold spells may be sunnier. cloudy spells with average temperatures
may be more difficult. you might run a simple simulation using nrel's
nearest hourly tmy2 data.
could this be superinsulated? a modest 300 kwh/mo of electrical use
adds 34k btu/day of internal heat. with 60 cfm of air leaks the house
needs about 24h(70f-38.2f)60 = 46k btu/day, ie a higher electric bill,
even with lots of wall insulation and very few windows.
with a perfect air-air heat exchanger and 0.2 ach of air leaks to make
60 cfm of fresh air and 300 kwh/mo of internal heat, a superinsulated
house needs 34k = 24hx0.2v/55(70f-38.2f), ie a maximum volume v = 12.3k,
as in a house with 1,500 ft^2 of floorspace, eg a 32'x16' cylinder with
lots of insulation and no windows.
>...i'm really close to making 100% solar as long as
>i control the movable insulation in an optimal way.
if an r2 window is only r10 at night, with the shutter closed, and
its shutter is open for h hours per day, it needs to gain at least
h(70f-38.2f)1ft^2/r2+(24-h)(70f-38.2f)1ft^2/r10 btu/ft^2 of solar
heat to break even on an average day, compared to a house wall
with a much larger r-value.
if the south window shutters are open for, say 6 hours on an average
december day in olympia, and the se and sw shutters are open for 5,
with 4 for e and w and 3 for ne and nw and 2 for the north, they need
at least 280, 170, 156, 140 and 125 btu/ft^2 per day of sun just to
break even, compared to a superinsulated wall, which isn't much less
than the 350, 260, 175, 140 and 110 average sun. these windows gain
little net heat per square foot in december, especially the northern
ones. can hugeness and automatic shutters compensate for that?
if the s, se, sw, e and w walls are r2 windows with r8 shutters that
your computer moves on the schedule above, with sun power sensors, and
they admit the daily solar energy listed above, and the rest of the
40' 2-story cylinder surfaces had r-value r, the conductance between
the house and the outdoors would be about
u = 64 btu/h-f for air infiltration
+ 2400ft^2/r for the floor and ceiling
+ 3x256ft^2/r for the non-window walls
+ 6/24x128ft^2/r2 + 18/24x128ft^2/r10 for the south wall
+ 5/24x256ft^2/r2 + 19/24x256ft^2/r10 for the se and sw walls
+ 4/24x256ft^2/r2 + 20/24x256ft^2/r10 for the e and w walls...
= 179+3168/r btu/h-f.
the daily solar energy gain would be 0.8(350+2x260+2x175)x128ft^2
= 125k btu. with 300 kwh/mo of electrical use, on an average day
we need 159k = 24h(70f-38.2f)(179+3168/r), ie r108 non-window
exterior surfaces, eg 21" of styrofoam (!)
that sort of takes care of average days, but this house is also
bad news on cloudy days, with mostly r10 walls if the shutters are
always closed. thermal storage for cloudy days is a problem, since
the house is so poorly insulated.
its thermal conductance is 179+3168/108 = 208 btu/h-f, with rc
= 32.8kbtu/f/(208btu/h-f) = 157 hours. the electrical use makes the
cloudy day balance point temp 38.2f+34k/24h/208 = 45 f, so after 5
cloudy days in a row, the concrete reaches 45+(70-45)exp(-5x24/157)
= 56.6, with 38.2+(56.6-38.2)/(1.5/1968+1/208)/208 = 54 f room air.
how can we avoid room temperature droop over a few cloudy days with
no backup heat, as long as the thermal mass is close to room temp?
a solar attic? a compost furnace? some smaller and warmer mass with
a larger temp swing can eliminate droop...