=energy =manufacturing
Around 1997,
heliostat costs were
estimated at $250 to $300 per m^2. Adjusting for inflation, that's to $460
to $560 in 2022 dollars.
In 2022,
this report estimated the
cost of 2 modern heliostat designs at $127 and $96 per m^2. So, heliostat
real costs came down >4x in 25 years.
what changed
Some reasons
heliostats got cheaper are:
- Heliostat
fields used to use open-loop control, which required stable foundations and
precise positioning. Newer ones use closed-loop control, with mirror angles
detected by cameras or something.
- Heliostats all used to be connected
with wires for power and data. There's now a trend towards ones with a small
PV panel to power the drive system and a wireless datalink.
- People
found cheaper ways to manufacture the supports and drive systems. You can
see the report linked above for some examples.
With such low cost targets, a lot
of common manufacturing methods can't be used. For example, an obvious
approach to supporting a mirror is to stamp a plastic-coated steel sheet
into a ridged pattern, then bond it to the back of the mirror. But that's
too expensive.
On some of those heliostat designs, you can see this
funky drive with a roller chain going around a metal semicircle, connected
at the ends. I guess that works, but if you were actually doing mass
production, I feel like an injection-molded (overmolded) nylon semicircular
gear would be better...?
Another common actuator in low-cost
heliostat designs is a long narrow leadscrew that only pulls.
heliostat applications
$100/m^2 is not very expensive compared to construction. Windows in
buildings are >$200/m^2. Houses often cost >$1000/m^2 of floor area to
build. With helistat prices having decreased, perhaps it's worth considering
potential applications for them now.
lighting
Currently, people
use solar panels to turn sunlight into electricity, move that electricity
long distances, and use it to run LED lights. That has a net efficiency of
perhaps 8% and requires a lot of equipment. It's more efficient to use
sunlight for lighting directly.
Heliostats can focus sunlight onto
skylights or windows. A few buildings have used such systems, but they're
rare. You have to put the heliostats somewhere they get light, and people
prefer big windows to small windows with extra sunlight pointed at them, so
heliostats have largely been used to technically comply with
regulations about not entirely blocking sunlight to nearby places.
There's also a town in a valley
that built a few heliostats to give it sunlight for more of the year.
water heating
Even if
heliostats have gotten cheaper, if you don't need very high temperatures, my
understanding is that parabolic troughs are still cheaper. But using those
for water heating is a real thing.
As of 2017,
global solar hot water thermal capacity is 472 GW.
process heat
Rather than
generating electricity, high-temperature heat from concentrated sunlight can
be used for endothermic industrial processes directly. The obvious options
are methane steam reforming and cement production. There are a few
startups pursuing
concentrated solar for both those things, using heliostats and towers. That
seems cheap enough for society to do if it wants to, but not cheap enough to
be competitive with burning cheap natural gas.
PV solar
Once upon a time,
solar panels were expensive and 2-axis tracking was worthwhile for them.
Today, solar panels are cheap, and single-axis tracking is worthwhile but
2-axis is not.
solar-thermal
Installed
solar cost per peak watt is ~$1 for big utility projects and ~$3 for
residential solar. Average/peak power (capacity factor) varies with location
and tracking, but 25% is typical with 1-axis tracking. So, you multiply
those costs by 4 or so, more if it's rooftop solar with no tracking. And you
can't control generation times, so maybe that electricity is worth less than
average. (These are approximate numbers, of course.)
If you suppose
$100/m^2 for helistats, 250 W/m^2 average solar irradiance considering the
tracking, and 40% conversion efficiency, that's $1/W for the heliostats, but
that's acceptable. That leaves $3/W for generation to match PV solar, and we
can spend significantly more than that if there's storage.
Well,
actual solar-thermal projects have been considerably worse than that. The
Ivanpah
Debacle was $6/W peak output, with a 22% capacity factor, no energy
storage, and 19% efficiency. So, >$27/W overall and no advantages over PV
solar. And that's at a relatively good location for solar-thermal; most
places are worse.
Maybe investors should've gone
with, say, NREL people instead of those BrightSource idiots? Well, the
Crescent Dunes project went even worse. Let's see, the CEO
was...this guy?
MBA, international studies, former Navy...yeah, that makes sense.
Solar-thermal has a lower capacity factor than PV. It uses 2-axis tracking,
which helps somewhat, but it's worse when there are clouds or low sunlight
levels.
Some things that aren't economically viable for solar-thermal
power include:
- Using
steam. The steam turbines and such are too expensive. Coal plants aren't
really competitive these days, and intermittent use would make things much
worse.
- Using nitrate molten salts for heat storage. The salts are
cheap, but some corrosion mitigation is needed and that makes things too
expensive.
- Using carbonate or chloride molten salts for heat storage.
They're more corrosive than the nitrate salts and require higher-temperature
materials, so they're more expensive despite less material being needed.
- Using packed-bed thermal energy storage, with gas flowing through
something like sand. It's
too expensive.
- Thermal energy storage that stores heat by releasing
CO2 from minerals and later absorbing it. Too expensive, and not very
efficient either.
- Not using thermal energy storage. Without it,
solar-thermal has no real advantages over PV solar, and the
electricity-generating equipment isn't used enough.
Solar-thermal power can be viable. Just avoid the above things. It's a fun little challenge.
decoration
A cost-optimized
mass-produced heliostat is about as cheap as a moving thing that size can be
made. $1000 for 10 m^2? Do you know how much stupid-looking large moving
sculptures that rich people or companies might get cost? More than $1000,
that's how much.
If part of the mirror was painted over, the
reflected light would form an image. A painted heliostat could be used to
have a company's logo wander around a shaded wall during the day.