dyson sphere stability

From: Eugene.Leitl@lrz.uni-muenchen.de
Date: Fri Dec 01 2000 - 10:21:21 MST


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Jeff Davis writes:

> The predicted instability of the solid version of the dyson sphere has been
> a source of disappointment for moi ever since I heard about it. The idea

While the idea of man-made celestial crystal sphere is rather neat, I
do not see any essential features not covered by a cloud of orbiting
objects, which can be made even optically dense. It would be
interesting to see how we can use this to intercept the majority of
the solar wind, too. Implanting them into periodically recycled
sacrificial layers; using magnetic field traps?

So, whence the disappointment?

> of the fully enclosed star was so delightful to me, and so elegant, that I
> pondered long over the matter seeking some remedy, some bit of phenomenon
> to rescue me from my despair. I considered the predicted instability, and
> concluded that it was factually supported, but based on +considerations of
> gravity and gravity alone. Gravity, to the exclusion of all other forces
> in the universe, as if no other forces existed. Fortuitously, this is not
> the case. Also fortuitously, the gravitational 'instability' is neutral.
> That is, gravity does not actively drive the system to collapse, but rather
> stands by, having no effect at all. This is in contrast to orbital

Hmm. Even if you could construct it, what about self gravitation
(collapse is autocatalytic), and tidal forces? Also, the thing as a
whole is in orbit (since static sphere can only be too gossamer to be
of much use), so the equator will be doing allright, but you're in
deep shit as soon as you start deviating from the equatorial
orbits. Orbital cloud solves this by changing height and inclination
of the orbit, so stuff never collides (this won't work if the orbiting
stuff is dumb and has no control over own orbit). With the cloud, you
have only the fragmentation catastrophe, and you can lick that with
active orbit control.

> Radiation pressure, baby. It ain't big, but it's real.
 
If your orbit has no radial component (wheee!), you can't held
anything relevant afloat, even if the floating stuff is heated to
white incandescence. Do the math. Magnetic plasma sails do much
better, but you can't operate another one within few 10 km radius.
 
> Solar radiation on the interior of the sphere will tend to drive the sphere
> so that the center of the sphere describes an 'orbit' around the star. To
> see why this is so, consider what happens if the star is off-center. The
> portion of the sphere nearest the star has the greater density of incident
> photons to both absorb and reflect. It will reach a higher temperature and
> thus a higher albedo, so it will reflect (rather than absorb) a higher
> percentage of the incident photons than the more distant (and thus cooler)
> interior surface. These two factors, photon density and correlated albedo
> values, should be the determining factors in the net thrust derived from
> radiation pressure. Higher albedo means higher reflection, and a reflected
> photon delivers twice the thrust of an absorbed photon. This thrust should
> act as a restoring force driving the sphere back towards the concentric
> position. There being no particular reason for it to stop once it reaches
> the concentric position, it would tend to overshoot, decelerate and then
> reverse direction, oscillating in a back and forth manner. This
> oscillation, in two dimensions, and out of phase, would stabilize to a
> condition where the sphere center describes an 'orbit' around the star.
 
Have you considered that you're heating the star up, by reflecting the
bulk of radiation back in in form of shortwave photons? (Instead of
making the orbiting objects as black as possible). Whatever the star
is going to do, this is not going to be sustainable. You're going to
heat the star, which will bloat and emit bluer and bluer. You'll
likely can boost mass loss that way, by making the surface hotter.

> The exterior of the sphere will likewise have an asymmetric radiation
> emission pattern resulting in a net thrust. A worst case scenario is that

Why assymetric? A fraction of the photons get absorbed, and is heating
the envelope. The asymmetric thing is more interesting if you can make
the cloud radiate anisotropically, predominantly in one
direction. However, this impacts recoil upon the cloud, shifting it's
orbit. Will you cancel out the impulse impacted upon the residual
system (minus the radiated photons), if you use photonic pressure to
readjust the disturbed cloud component orbits? I have no idea. Someone
please tell me, as it is rather significant.

> the (waste) heat is transmitted through the shell locally and re-emitted
> locally. Clearly, it would be re-emitted--dumped--at a lower wavelength
> resulting in a lower net thrust (I think I got that right) than that caused
> by the radiation forces on the interior. The interior and exterior
> radiation reaction forces appear to work in opposite directions, but since
> the radiation incident on the interior surface is a restoring force and the
> greater of the two, the net effect is a restoring force.
>
> One piece of this puzzle that goes beyond what I can handle is what is the
> effect of the radiation in the interior of the sphere after it first
> strikes the shell? If it's absorbed I have no problem, but if it's
> reflected and then absorbed, or bounces around for several generations
> before finally being absorbed, what is the net effect? I think first order
> effects will dominate as described, but I'm prepared to hear other points
> of view.
>
> Someone may say that radiation forces are puny in the inertial arena. I
> agree, but if they're the only forces in play, then they must dominate.
> Response times are just very long.
>
> Candidate number two is something of a mystery to me. I know it's there,
> but I frankly don't know how it acts on the system or how it compares in
> magnitude of effect to radiation pressure.
>
> Magnetic field effects.
>
> The star may have a magnetic field. The field lines may extend out into
> space. How far out do they extend? They may interact with the dyson
> sphere. They may be conducted through the dyson sphere. What is the
> sphere made of? If it were made metallic and isotropic, would the star and
> sphere be magnetically drawn together, or held stably concentric? Could a
> non-isotropic or non-metallic sphere design take advantage of or circumvent
> magnetic effects to passively achieve spherical concentricity?
>
> Help me. I've fallen and I can't get up.
>
> One last note. Though the 800 lb gorilla appears neutral in this matter,
> that's only first order effects. Second order gravitational effects may
> come into play. If the non-uniform heating of one side of the sphere
> causes it to distort, or tidal forces cause it to distort, then
> gravitational forces may again come into play. The devil's in the details.

The devil is that you can't build a sphere of any relevant size
without infinitely strong material (Unobtainium still isn't in the PSE
last time I looked). Denser objects are sufficiently small, but
they're typically fierce and have literally astronomic gravitation
gradients. You might construct you classical Dyson sphere around a
very small singularity (a primeordial relict or a newly minted),
though. If it has something like a few g/s mass loss as Hawking
radiation, and has the right wavelength (I'm rather not warm my toes
directly on a hard xRay source), you could build your sphere around
it. Even fill the sphere envelope with some gas/dust, so that you can
homeostate the hole in the right regime by feeding just as much as it
evaporates, and attenuate the hard radiation from the eye of the
dragon (hole and accretion disk). The whole thing doesn't have to have
more than a few Gt mass, probably, and would be rather cold on the
surface. Invisible unless there are lots of them, or you look at one
from real close. And rather efficient, of course.

- -----BEGIN PGP PUBLIC KEY BLOCK-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.5 and Gnu Privacy Guard <http://www.gnupg.org/>
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=JX2n
- -----END PGP PUBLIC KEY BLOCK-----
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.5 and Gnu Privacy Guard <http://www.gnupg.org/>

iEYEARECAAYFAjon3f0ACgkQhudim1f5z9MTGwCg6tzzoY070d8LqsyS8TV2FRZP
BH4AnRbRCGekdvMoJUTPyJ7sVOkNLlPA
=FWlB
-----END PGP SIGNATURE-----



This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:32:09 MST