Artificial vs. natural illumination for space habitats

"delt0r" wrote in message
oups.com...

But i'm not sure about concerntration it 70x --thats not trival optics
and you may have just throw'n away all the simplicity.

We view this much the same way. And the peer-reviewed designs very
frequently valued simplicity over efficiency.

To me, concentrations of 10x or less would still go a long way toward
reducing window area.

--


Regards,
Mike Combs
----------------------------------------------------------------------
By all that you hold dear on this good Earth
I bid you stand, Men of the West!
Aragorn

Mike Combs wrote:
"Joe Strout" wrote in message
...

I know you're working toward a point here, and it's probably rude of me
to interject -- but to this particular question, the answer is not
difficult: the original light energy is in a different place, and
probably of a different spectral quality, than what and where you want
it. Compare to this question: why convert the rotational power of the
steam turbine at the power plant to electricity, pipe it to someone's
home, and then convert it into rotation in your blender, when rotation
is what you wanted in the first place? You could instead have long
rotating axles (connected by universal joints) transmitting the rotation
directly to your blender. But it's much more practical to use
electricity as the intermediary; it's easy to transmit and reshape into
whatever particular form of power you need. This is true despite the
efficiency losses.

I take your point, but I'll never consider getting sunlight inside of a
windowed habitat as being nearly as impractical as your axel to the blender.

That may be so, but this leaves some things out, like the Chevron
shields needed on top of any glass panels,

Bear in mind that the chevron shields are for one particular habitat design:
the Stanford Torus. The Bernal Sphere and Crystal Palace deal with this
problem via other (possibly better) methods. And once we get up to the
Island 3 scale, the thickness of the structure itself eliminates the need
for separate shielding (chevron or otherwise).

and the radiator mass needed
to reject all that extra heat,

Undeniably true. But I think one thing this issue hinges on is which is
more expensive per sq meter: simple panels of aluminum tubing, or panels of
PV arrays? The natural illumination option needs to reject more heat,
meaning somewhat greater area heat radiators, but the artificial
illumination option requires a lot more electricity than the habitat would
otherwise need just for routine electrical utilities, which means /much/
bigger solar arrays.

as well as severe constraints it imposes
on overall colony geometry, which can have serious impacts on your mass
budget -- for example, it would be hard-to-impossible to make much use
of natural light in a multi-deck design like Kalpana One.

I would agree, but have a personal dislike for multi-deck designs. They
pursue a different design goal, which is "provide as efficiently as possible
living area for X people". I still have a fondness for O'Neill's original
design goal, which was "recreate as closely as possible the most attractive
parts of the Earth's surface".

Yes, I found that section useful. I'm not convinced it's a good idea
either, though. I'm really not trying to be contrary, but it seems to
me that if your inhabitants really want windows, they probably expect a
decent image (you know, normal-sized sun during the day, ordinary stars
and maybe a glimpse of the Earth by night). If they're willing to forgo
having windows they can actually look out of, then I think it's probably
not worth bringing in natural light at all.

Yes, 70x sounds a bit extreme to me. I'm not sure what concentration factor
is assumed in the Bernal Sphere design, but it doesn't look to be beyond
10x. And that strikes me as about right. I'd agree that making the windows
substantially less than 50% of your overall hull is probably to be
recommended, but I think you can go to extremes here.

If we use some of the sunlight passing through the windows for power
generation, let's let the physics determine how 'extreme' we should go.
Months ago I read descriptions of cooled quartz windows passing
10,000X sunlight (sorry, can't seem to find it again). Let's assume a
more modest 1000X concentration of sunlight. Also assume, as described
elsewhere in the thread, a cylindrical colony, 8km long and 4km in
diameter, with the interior illuminated at 100W/m^2, and ½ the energy
of the sunlight (the 'useless' wavelengths) captured at the windows.
Let there be 1000 such windows scattered around the middle of one
endcap of the colony. A fluid, opaque to the appropriate wavelengths,
would circulate between quartz layers to capture the energy, cool the
window, and provide power to the colony (heat engines are still more
efficient than PV cells). At night, the fluid would change so as to
absorb _all_ the light; no reason to waste the power-generating
capacity of the windows.

Each such window would be 4.3m in diameter, much easier to maintain or
to repair if ruptured than huge expanses of floor-windows. The total
area of the windows would be about 1/9th of 1% of that endcap's area.

I think we need to forget about windows on O'Neill-style colonies that
allow views of the starry sky and of the Sun swinging by in an
Earthlike fashion. O'Neill may have proposed it to encourage the idea
that space development didn't mean being trapped in little, tin-can
space stations or having to terraform planets but I don't think it's
practical to force an actual design to mimic _every_ aspect of life on
Earth's surface. I would design the colony as a large version of many
houses built in desert climates, with a lush garden in the center and
walls all around. If you want to see the Sun and the stars, go down
into the colony 'basement' to a nice, dark chamber and watch through a
floor-window. Sunlight should enter the colony in the safest, most
efficient manner. Using ½ the floor area for relatively fragile acres
of windows doesn't seem like the way to go.

In article ,
"Mike Combs" wrote:

and the radiator mass needed
to reject all that extra heat,

Undeniably true. But I think one thing this issue hinges on is which is
more expensive per sq meter: simple panels of aluminum tubing, or panels of
PV arrays?

Well, it's the total cost that matters, not the cost per square meter.
Unfortunately that probably requires a more detailed analysis than we
can do here.

as well as severe constraints it imposes
on overall colony geometry, which can have serious impacts on your mass
budget -- for example, it would be hard-to-impossible to make much use
of natural light in a multi-deck design like Kalpana One.

I would agree, but have a personal dislike for multi-deck designs. They
pursue a different design goal, which is "provide as efficiently as possible
living area for X people". I still have a fondness for O'Neill's original
design goal, which was "recreate as closely as possible the most attractive
parts of the Earth's surface".

Yes, this is a very insightful point. To put it another way, O'Neill
was addressing the concern, "People won't want to live in a can." He
pretty much took it as assumed that people could afford to do so if they
wanted to. In my efforts, I'm generally trying to address the concern,
"people can't afford to build a can they can live in." I take it as
assumed that people will want to if they can afford it.

Of course, O'Neill didn't wastefully make his designs expensive, and I
don't wastefully make mine unattractive. But emphasizing one or the
other does lead to some different design decisions.

But getting back to the multi-deck designs, they don't necessarily have
to be less natural than, say, a torus. For example, I've toyed with
(but not yet studied in detail) disk geometries that would have a true
image of the sky on each deck (the image is actually focussed through
long windows on the side wall).

I also think that we might actually be better off with an entirely
artificial sky -- that is, a smooth white surface with a series of
high-power, high-resolution projectors trained on it, like at a
planetarium. This could show an Earthlike sky during the day (including
sunrises and sunsets), and at night, show you the real outside view.
Assuming the sky is sufficiently far away (say, 50 m or more), and the
projection is of sufficiently high quality, this would look as good or
better than real windows.

But I know, our gut tells us that this is a "fake" while windows and
mirrors are "real," and emotionally this is a significant difference.
It's hard to tell which one most people would prefer -- a realistic but
very unnatural sky, or an artificial but very natural-looking one. Some
data points can be gathered from places like the Blue Bijou in
Disneyland, where you can eat under a twilight Louisiana sky, even at
high noon. People know it's fake, but (in my experience) find it quite
pleasant -- and that's a roof only 3 m over your head instead of 50. In
fact, by the time I'm done with lunch and emerge into the real sunlight,
I always feel a bit of a shock, as I had subconsciously forgotten that
it wasn't actually dusk.

Best,
- Joe

Mike Combs wrote:


Bear in mind that the chevron shields are for one particular habitat
design:
the Stanford Torus. The Bernal Sphere and Crystal Palace deal with
this
problem via other (possibly better) methods. And once we get up to
the
Island 3 scale, the thickness of the structure itself eliminates the
need
for separate shielding (chevron or otherwise).

This may be the case for raidation shielding, but not for impact
shielding, where a multilayer design is far far superior, but makes
heat rejection problematic.

We also now know that we need to plan against terrorist attack - a
multi layer shield on an O'Niell cylinder could probably withstand some
forms of nuclear attack. (Any thoughts on this? - how effective is 5cm
of steel, followed by 50m of vacuum followed by 4 metres of moon rock
foam, in absorbing a nuclear blast?)


and the radiator mass needed
to reject all that extra heat,

Undeniably true. But I think one thing this issue hinges on is
which is
more expensive per sq meter: simple panels of aluminum tubing, or
panels of
PV arrays? The natural illumination option needs to reject more
heat,

Why do you assume this? The primary mirror can be coated with an
optical filter to absorb non wanted frequencies. What is then
cioncentrated and beamed in to the cylinder is close to 100% useful. In
comparison, no artifical lighting comes close to being this efficient.

meaning somewhat greater area heat radiators, but the artificial
illumination option requires a lot more electricity than the habitat would
otherwise need just for routine electrical utilities, which means /much/
bigger solar arrays.

But assuming a multi layer design, as proposed above, heat rejection is
a much bigger problem. It's not the area of the radiators, it getting
the fluids to the radiators via the hubs.

as well as severe constraints it imposes
on overall colony geometry, which can have serious impacts on your mass
budget -- for example, it would be hard-to-impossible to make much use
of natural light in a multi-deck design like Kalpana One.

I would agree, but have a personal dislike for multi-deck designs. They
pursue a different design goal, which is "provide as efficiently as possible
living area for X people". I still have a fondness for O'Neill's original
design goal, which was "recreate as closely as possible the most attractive
parts of the Earth's surface".


Yes, 70x sounds a bit extreme to me. I'm not sure what concentration
factor
is assumed in the Bernal Sphere design, but it doesn't look to be
beyond
10x. And that strikes me as about right. I'd agree that making the
windows
substantially less than 50% of your overall hull is probably to be
recommended, but I think you can go to extremes here.

You get most of the benefit by 10x. However, if you want an axial beam,
it helps to keep the glass size down. Hence I'm now more in favour of
circumferetial lighting.

"Alex Terrell" wrote in message
ups.com...

This may be the case for raidation shielding, but not for impact
shielding, where a multilayer design is far far superior, but makes
heat rejection problematic.

And I think probably for that reason is why multilayer designs aren't
assumed for Island One.

The natural illumination option needs to reject more
heat,

Why do you assume this? The primary mirror can be coated with an
optical filter to absorb non wanted frequencies. What is then
cioncentrated and beamed in to the cylinder is close to 100% useful. In
comparison, no artifical lighting comes close to being this efficient.

Just trying to let the other side of this debate have a point.


--


Regards,
Mike Combs
----------------------------------------------------------------------
By all that you hold dear on this good Earth
I bid you stand, Men of the West!
Aragorn

Alex Terrell wrote:

(Any thoughts on this? - how effective is 5cm
of steel, followed by 50m of vacuum followed by 4 metres of moon rock
foam, in absorbing a nuclear blast?)

That depends on whether the "blast" delivers any concussive force or
not. Nukes in space can only generate a concussive force with the mass
attached to the nuke (not much) and - if it detonates really close to
the target - what it can ablate off the target.

Otherwise, you only have radiation from the nuke to do damage, and the
radiation from the nuke won't be more penetrating than cosmic rays,
which the shielding can stop.

Mike Miller

Above I stated that I prefer reflected light, because its possible to
remove the unwanted spectrum outside the habitat. What is beamed in can
then be used at close to 100% efficiency.

I also said the cost of the power is negligible compared to the cost of
the station.

However, one consideration that favours electrical illumination, is the
fact that colonies may well be grouped in pairs. If one were to have
day, whilst the other has night, the same power station could
illuminate both colonies. (Limiting it to 50% light cycle on average).

This could in theory be done with sunlight, but the mirror mechanism
would be complex. Furthermore, it would take time to move the mirrors
to switch the power, so each colony would get a little less than 50%
average of the sunlight.

Complexity should be avoided, not for reasons of cost (as said, its
peanuts compared to the colony construction), but for reliability. If
natural light is chosen, then each of the two colonies would probably
have its own mirror, which would be blacked out at night.

There may be some facility to transfer some of the light from one
colony to the other to provide back up illumination, though this might
be better provided electrically.

So now I'm undecided.

Mike Combs wrote:
I've lost count of how many times on these newsgroups I've encountered
someone saying that providing space habitats with natural sunlight via
mirrors and windows would just be too complicated, and we'll instead use
artificial illumination. Sometimes the advantages of using red LEDs to
raise crops are touted (less wattage, less need for heat rejection).

I've always argued against this, mostly from instinct, but also from knowing
the actual solution the original space settlement studies did settle on. I
was concerned about conversion efficiencies (why convert light to
electricity and then back to light again when light is what you wanted in
the first place?), and it always struck me that aluminized Mylar and glass
panels by the square mile would be pricey, but PV arrays and artificial
lights by the same magnitude would be more so.

Here's a paper that seems to agree:

"Effect of Environmental Parameters on Habitat Structural Weight and Cost"
http://www.nas.nasa.gov/About/Educat...eres/II-1.html

In the midst of looking into a variety of parameters for habitats for 1,000
(early construction shack), 100,000 (intermediate range earthlike habitat),
and 10,000,000 (long range habitat), and for toroidal, spherical,
cylindrical, and Crystal Palace geometries, the paper compares the costs for
artificial vs. natural illumination.

Table 6 indicates that natural illumination with mirrors should have only
about 20% the cost of the artificial illumination option.

The paper does agree that concentrating solar energy so as to minimize
window area is to be recommended. They say you could get up to a solar
concentration level of 70x before getting into problems with the glass
softening.

--


Regards,
Mike Combs
----------------------------------------------------------------------
By all that you hold dear on this good Earth
I bid you stand, Men of the West!
Aragorn

Joe Strout wrote:
I was recently
reading about this actually being put to use in real buildings -- light
collectors on the roof, fed to fiber optics, which then light up glow
tubes right next to the fluorescent lights. Light sensors automatically
shut off the fluorescents when the piped light is bright enough.

However, the losses in the fiber optics are pretty severe -- IIRC, both
because of the wide bandwidth of the light, and because of its
intensity. I regret that I don't have the figures handy, but the upshot
was that this was considered impractical for anything but the top floor
of the building.

I'm surprised by this. End losses (getting light into and out of
fibers) can be 10% or so at each end, but the best fibers have only a
few dB/km loss internally and are good from visible light through a
couple of microns wavelength in the infrared. Are those fibers too
expensive for buildings? And electromagnetic radiation is linear. Why
should intensity should make any difference until it's high enough to
start heating the fibers.

Joe Strout wrote:

This is the incorrect part. Light losses in the
fiber optic cables will be severe. (And of
course, all that lost light gets turned into heat,
that your colony then has to reject.)

This looks like about state of the art for broad
spectrum optical fiber for lighting use, and with
ten meters surely enough to get the light through
the hull, 100 dB loss per kilometer over much
of the spectrum doesn't look like too much to
suffer.

http://www.polymicro.com/products/op...fibers_fbp.htm

If that's too much infra-red transmissivity, an IR
mirror could be put in front of the collection
surface, I suppose.

You may be working on old information.

Alternately, I may be entirely incompetent to read
that graph. I sure can't read the math in the formal
papers on Rayleigh scattering with much facility.

FWIW

xanthian.

wrote:
Alex Terrell wrote:

(Any thoughts on this? - how effective is 5cm
of steel, followed by 50m of vacuum followed by 4 metres of moon rock
foam, in absorbing a nuclear blast?)

That depends on whether the "blast" delivers any concussive force or
not. Nukes in space can only generate a concussive force with the mass
attached to the nuke (not much) and - if it detonates really close to
the target - what it can ablate off the target.

My assumption was that you have massive radiation exposure of the outer
shell

Otherwise, you only have radiation from the nuke to do damage, and the
radiation from the nuke won't be more penetrating than cosmic rays,
which the shielding can stop.

But in the volume radiation from a nuclear bomb, I think a large
section of the outer shield will be instantly vapourised. However, the
"instantly" will not be as "instant" as the nuclear flash. The plasma
of iron nuclei will still continue to absorb the nuclear radiation.

This plasma will then generate a concussive force, which will transfer
significant heat and momentum to the next level of the shield. How much
I don't know.

What is clear is that the multilayer design is orders of magnitude
better than a single layer shield.