[Clewza313]

The game changer here is SpaceX's Starship. If you can lift 100T to orbit in one shot, that suffices to deploy around 6 MW worth of the current solar panels we use for the ISS, which clock in at around 120KW/2T. Solar panel efficiency has improved greatly since they were designed and built, so even accounting for the overhead of assembly, refueling, the transmission component etc, 10MW+ per launch seems quite feasible, particularly if Musk gets costs down to the mooted $20/kg or so.

Charlie Stross does the napkin math here: https://www.antipope.org/charlie/blog-static/2021/09/fossil-...

[marcosdumay]

Interesting. That's 0.20$/W on launch costs, what is actually not a showstopper.

I still doubt it will happen any time soon, because by the time we have enough launch capacity to use on things like this, the 0.2$/W will be a sizeable fraction of the costs of solar on the ground. The advantage is that you don't need batteries, but like photovoltaics, the floor on the costs of batteries is very low.

[giantrobot]

Their math is wrong. The ISS solar panels plus supporting truss elements mass about 60t or 2KW/t. At the fantastical (unbelievable) $20/kg to LEO that's $10/W. That doesn't include the mass of thermal control elements, maneuvering systems, power conversion, or power transmission system.

Also that's only to LEO which is useless for an SPS since the system wouldn't dwell over any receiver on the ground long enough to transmit a useful amount of power. You'd need all that mass in a geosynchronous orbit which trebles or quadruples that cost.

[giantrobot]

You're missing a few details.

1. The current (public) Starship design has no facilities for cargo delivery. While it can lift a lot of mass it doesn't have doors to unload large payloads in space. SpaceX certainly can build cargo versions but that will require non-trivial changes to the craft. Cargo doors aren't load bearing structures.

2. Starship's "100t to orbit" is only to LEO. To get a payload to a better orbit you're going to need to also launch some sort of bus and lots of fuel. Even if it's just a buddy fueling Starship that's still more launches and rendezvous.

3. Musk's claims of $20/kg are unbelievable. He has a long colorful history or over promising and under delivering. A fully fleshed out Starship production pipeline will bring launch costs down but not that much. It'd be nice to be wrong but I'd take all of Musk's claims with a giant grain of salt.

4. You're vastly underestimating the amount of mass you'd need to put in orbit and then assemble. Solar panels aren't load bearing so they all need to be attached to some sort of support truss (like the ISS truss). Then you've got the power conversion system, thermal control system, maneuvering system, and power transmission system. Each of those systems will easily mass the same as the actual solar panels.

5. The ISS solar panels with the support truss elements mass about 60t [0], so they generate about 2KW/t. Assuming power conversion was 50% efficient (I think that's reasonable) a gigawatt SPS would need at least 1000t of panel assemblies with at minimum another 500t of supporting equipment/infrastructure. I'd argue the support structure would be at least 1000t.

6. To make that SPS even remotely useful you'd need it in a geosynchronous orbit, any lower and it would pass too quickly over the ground to transmit any power.

So an SPS, sending just a gigawatt of power to the ground, would require a minimum of 1500t (~4x ISS) launched into a geosynchronous orbit. That's in addition to the fuel and buses launched to facilitate construction. All of that is just for assembly. None of that counts the component construction on the ground of the space-qualified hardware.

Meanwhile a gigawatt of renewable power on Earth is orders of magnitude cheaper and doesn't require several square kilometers of rectenna on the ground. A bit of debris that breaks a ground based panel also won't cause a chain reaction that can destroy the entire installation.

[0] https://space.stackexchange.com/a/9758