Hacker News Reader

Who cares about the efficiency when solar panels weigh over 50 kg per kW? Even at SpaceX prices, the cost of launching a solar panel into space is hundreds of times the cost of the panel. If your alternatives are "buy 500 panels and put them on the ground" and "buy one panel and put it in space," why the hell is anyone even remotely considering the second option?

Wasn't the original concept for SPSS (at least in O'Neils The High Frontier) to build them out of mined asteroid materials? At that point the cost of the proposal is disconnected from the cost of launching the panels. I've heard many analyses of that which do suggest the economics of it is inevitably atrocious for ground-launched SPSS. Cost numbers for asteroid mining vary by a couple orders of magnitude though, which mitigates the above...

So now instead of launching panels, you're launching a panel factory. Actually you're launching their entire supply factories too (mining, high-purity silicon boule manufacturing, wafer cutting, smelting metals for busbars and backing, etc). None of this has been done in microgravity before. Everything will need mass optimization, and you need provisions for maintenance and repair on all these systems.

Doable? Sure. But you have to make a lot of panels before it pays off.

It could be worse:

http://discuss.joelonsoftware.com/?joel.3.219431.12

Space-based solar may work in powering space probes that are too far from the sun. An array of solar concentrators could beam down just enough solar power downrange to make the exploration of the Kuiper belt and inner Oort cloud worthwhile (there is no loss in conversion since you're redirecting photons all the way to the spacecraft). The alternative thing may be orbiting microwave relays that receive power from Earth and then beam it at the distant spacecraft (either as laser or microwave power).

>solar panel is on-orbit, you get twice the solar energy - assuming that it is out of Earth's shadow

Not quite. When compared to stationary panels (which beat trackers economically, so they should be the baseline), space-based power harvests π (yes, pi) times as much sunlight due to lack of cosine loss, plus there's the bonus of reduced reflective loss. Atmospheric and weather loss is avoided entirely. I would estimate 4-5x the energy.

The big advantage of space-based solar imo is the ability to create giant collection areas defined by minutely pressurized helium structures. A giant parabolic mirror on Earth needs to fight gravity, winds, and the Earth's rotation (tracking). Similar concept: https://tec.grc.nasa.gov/past-projects/solar-concentrators/a...

Launch a 30,000 m^2 concentrating collector in a single launch? It's only 12 MW, but it's constant (unlike ground-based solar at ~18% capacity factor), and it's deliverable anywhere in the world, or even to other satellites. What if microsatellites could summon extra power by POSTing an ephemeris to an endpoint? Is there a market there?

It's an interesting niche tech, but spaceflight is just so expensive that I doubt it'll ever be viable.

Concentrators have problems.

If you use them with solar panels, then your panel will heat up and its efficiency will drop. Maybe lifetime as well. If you use a thermoelectric generator, you then have wearing mechanical low efficiency systems on board.

In both cases you need to reject heat, which is hard in space.

They also require precise pointing.

There's also the problem of beaming the power to earth. If you're in low orbit, there would have to be lots of receiving stations and the power would be intermittent unless you had a big constellation. And all the problems with inclination and the spinning earth... If you're in geostationary orbit, your sending and receiving antennas would need to grow huge.

What about beaming between satellites? You'd need a laser for that. Maybe they're becoming quite small and efficient with modern semiconductor technology developed for military uses. Still the receiving end might have low efficiency.

Yep, after writing all that I realized that heat rejection is going to be a problem. Dissipating 8.4 MW is no easy feat in orbit.

At that point the balance of system probably shifts to thin film PV. There would need to be a lot of advancement in that area before such a project becomes practical.

>What about beaming between satellites? You'd need a laser for that.

I was picturing a microwave rectenna, actually.

If you use a laser, the receiving satellite wouldn't need extra hardware, it could use its ordinary solar panels, if they were pointed in the right direction.

>>>It's an interesting niche tech, but spaceflight is just so expensive that I doubt it'll ever be viable.

Harvest the energy in space. Microwave it to the wing-surface receivers of ultra-light high-altitude gliders which use the energy to stay aloft.

Use the gliders as a platform from which to deliver rockets.

The thing is that in the time when solar orbital was first discussed, solar panels sucked and oil was super-expensive. I think the O'Neill designs were based on concentrating mirrors and turbines. He claimed it could pay for itself within a standard utility company's financial planning horizon of 30 years.

And then oil prices went back to normal.

The real conceptual competitor at the time -- in fact, it's still the real competitor -- is plain old fission.

But nuclear is evil and not cleantech.

Solar energy is actually nuclear energy.

All energy is actually nuclear energy.

(Of course Earth-based nuclear power doesn't have the entire mantle of the sun as radiation shielding. So there's that…)

No, some is gravitational: tidal, arguably hydroelectric, though that's a combination of solar (feeding the water cycle) and gravitational (capturing gravitational potential of lifted water).

Geothermal is in part latent residual heat of formation, which is arguably either kinetic or gravitational energy again. Some is _also_ the result of radioactive decay, though I don't have the precise breakdown.

I suppose there might be some ways of tapping planetary or stellar magnetic fields, which aren't directly caused by nuclear power.

Similarly energy systems based on collapsed stars (white dwarves, neutron stars, black holes) would not be directly driven by nuclear processes.

But yes, your point generally stands: most available energy on Earth is largely ultimately derived from nuclear reactions.

True, tidal isn't nuclear. Basically everything else is.

>Some is _also_ the result of radioactive decay, though I don't have the precise breakdown.

Neither do I, but here's my logic: if there were no nuclear power in the center of Earth, Lord Kelvin tells us it would solidify in only ~40 million years[1]. The Earth being 100 times older, the amount of residual heat left today must be insignificant.

Stellar magnetic fields come from the convective mantle, which is nuclear powered. Not sure about planetary magnetic fields, but the above analysis suggests nuclear as well. Collapsed stars will be heated by gravitational potential, but we don't see a lot of those 'round these parts.

[1] https://en.wikipedia.org/wiki/William_Thomson,_1st_Baron_Kel...

My point wasn't that the vast, vast majority of energy available to humans on Earth isn't ultimately nuclear in origin. Just that there are energy potentials in the Universe which aren't nuclear in origin. If you trace back the formation of the Solar System itself, as a generation 3-4 system, even the gravitational potential represented in it originates from nuclear reactions.

The Big Bang (background microwave radiation) and Dark Energy would also come to mind as non-nuclear energy systems.

Not of much practical use to us on Earth, however.

Oil prices went back to "normal" for a bit but have since gone back up. This chart http://inflationdata.com/Inflation/Inflation_Rate/Historical... is out of date, but we are roughly back at the 1980 peak in real terms again (Brent was $113 last time I looked, US prices are probably slightl lower).

Absolutely, but in the meantime solar panels have improved enormously also. The third part of the equation is launch costs, which haven't really fallen at all.

Launch prices basically are energy costs, plus a markup, so they are never going to make this work unless you could make the whole thing extremely light...

O'Neill's plan was to bootstrap manufacturing in situ, starting with robotic mining of moon dust or rock. The ore would be sent to L5 by linear launchers and then smelted into simple metals to make reflective surfaces.

Seems to me that a significant advantage of 'space solar' is that you can ditch a significant amount of ground based distribution infrastructure. Transformers are pretty efficient but lose about 2% per step up/down[1], power lines are another 1-4%/100 miles[2], then you have losses related to storage since you can't turn solar on/off at will like you can with hydro or fired plants.

If a single solar sattelite [multiple for redundancy, in reality] can replace a large proportion continent's energy infrastructure, it starts to sound a lot more attractive.

1. http://en.wikipedia.org/wiki/Transformer#Energy_losses

2. http://en.wikipedia.org/wiki/Electric_power_transmission#Los...

You'd still have to distribute the power from ground-based receiving stations via the grid. No free lunch.

Yes, but I believe his point was that you could have multiple smaller stations close to where the power is needed. If it isn't safe to have the receivers near cities, then I don't think there's much chance that any receiver (anywhere) will get past the regulatory hurdles.

>a single solar sattelite [multiple for redundancy, in reality] can replace a large proportion continent's energy infrastructure

Woah there. Exactly how much power would each of these hypothetical satellites provide?

He's not exactly wrong but "it's ridiculous" is never a good gauge of what's feasible. Most of what we use on a daily basis is pretty ridiculous if looked at from the right perspective.

And while I don't have time right now for a complete rebuttal to Musk's criticisms I will point out that he's being somewhat of a hypocrite here. He's complaining about photon to electron to photon to electron conversion, but consider what might be his ideal: ground based solar power and electric cars. The only reasonable way to make that work is through photon to electron to battery to electron to battery to electron systems, which are every bit as ridiculous and ineffecient as what he is complaining about.

Except then you consider the alternatives :-/

photon to electron in space to photon' to electron on ground ≃ photon to electron on ground

photon to electron (on ground) to battery (filling station) to electron to battery (car) to electron to kinematics VS raw oil (limited on this very planet) to petrol to kinematics

Also not considering resistance loss and petrol transportation. A little bit more complicated :)

Why would you put solar panels up there? Why couldn't you just focus the energy into a beam and point it towards earth like space based mirrors for one of these solar power stations[0] where sunlight is focused on a 'boiler' which generates steam for a conventional turbine. I've no idea if the physics works out but i'd be interested to hear from someone who does. [0] http://en.m.wikipedia.org/wiki/PS20_solar_power_tower

Quite unintuitively, you can't really focus light from the sun to a long distance, since the sun is not a point source.

Basically, you can't exceed the power per steradian of the sun with mirrors. So you would have to fill a large portion of the receiving end's sky to transmit any real power. Which means the mirrors would have to be really close (like they are in heliostats) or if they're far away, they would have to be really huge.

Does not work.

>Basically, you can't exceed the power per steradian of the sun with mirrors.

This is something I've wondered about. It seems like one should be able to make a counterexample by focusing an enormous area on a tiny one. Is it just because due to "fuzzy focus," the concentrated focus volume unavoidably grows as the mirror size grows?

well, first thing that comes to mind is, with that anyone that builds a device that focuses a large amount of solar energy into a tight beam aimed at the Earth, I'll have a hard time believing that's going to be used for some benevolent energy-harvesting purpose.

And even if it is intended to, what will be the accuracy of this beam? what's to say some orbital junk won't crash into it and "oops" scribble a nice kilometres-wide Lissajous figure of scorched Earth onto our planet, like a hit and run graffiti tag?

Even if, according to WP, NASA is "a distinctly civilian (rather than military) orientation encouraging peaceful applications in space science", the US government has proven itself rather incapable of keeping itself in check, so the odds of it going to turn into "hey we could also turn it into a gigantic solar death ray space weapon", somewhere, at some point given its life cycle, seem pretty high to me?

I'd be surprised if most didn't think of this at some point when reading the original article, let alone when talking about concentrating huge energies from space in general.

That said, perhaps people don't consider it feasible, as in, "not in MY world". Sounds dystopian sci-fi and so on.

Space structures can be truly enormous, hold their shape with no gravity issues, be aimed with trivial amounts of energy, have no weather issues (no cloudy/rainy days). Any number of reasons this is a win.

Mr. Musk is just ranting. He's used conversion efficiency as his whole argument. The only argument worth having is total cost of ownership, which relates to payback time.

I imagine micrometeoroid impacts would be a significant "weather" issue.

Yes; of an entirely different nature than hurricanes, torrential rains etc. Unfortunately the best locations on earth for truly enormous high-efficiency solar are equatorial.

What if instead of solar panels, you put up big mirrors, and instead of earth you have them orbit far closer to the sun?

Btw, great quote.

yes, and shuttling antimatter batteries back to earth.

And people get scared of "giant solar death rays" from beaming solutions for power transmission. Anti-matter batteries sound terrifying.

Heh, that sounds great.

Someone make a stupid mistake on the ground, and your ship crash, and then BOOOM, a 20km wide hole shows up in your space center.

This seems to assume that you have the ground area on Earth for the panels. That's true where Musk is working in the US, but not for many smaller countries.

Of course I guess it that for those countries it would still be easier to put the panels somewhere else and build out the connection infrastructure than to put it in space.

There are plenty of countries that are net importers of energy now.

Surface area is at an even higher premium in space. One house > one satellite.

I know Musk is worshipped by the Hacker News crowd (and rightly so) but as with all things related to technology it's really hard to make predictions considering the fact that there could be a "black swan" type breakthrough or accidental discovery that completely changes everything.

He didn't predict anything- he analyzed the existing technology and infrastructure to conclude that the cost-to-benefit ration was too high. His argument stands until a 'black swan' type breakthrough occurs, and by its very nature there is no guarantee that will happen.

What about countries that get very little sunshine through winter season, scandinavian countries for example?

> and that double conversion is going to get you back to where you started, basically

That feels overly simplistic. What are the actual losses?

I think beaming the energy down is the wrong approach, instead that solar energy needs to be converted to antimatter.