I looked into some numbers on this, comparing theoretical costs of nuclear power with a solar PV solution on Mars.
Kilopower (NASA's research project for a Martian nuclear fission reactor, from the article): 7,200 kg for 40 kW. [1]
ISS solar arrays: 14,515 kg for ~100 kW in Earth orbit. [2] If we assume 40-50% Earth solar insolation on Martian surface, the ISS PV array probably isn't far off 40 kW output on Mars
Conclusions:
Nuclear reactor would have approx. 50% launch weight. SpaceX estimate $45/kg payload with a Falcon Heavy, so about $324K for Kilopower or ~$600K for the ISS arrays.
The build cost of the ISS arrays was around $300 million (space PV is way more expensive than terrestrial). The development and test costs of Kilopower is around $15 million; build cost of final units is unknown.
You would have to automate and/or remotely control all of the nuclear power plant operations. Dust storms would be a challenge to a PV solution, though not insurmountable.
They actually look very comparable. Nuclear has an edge due to it weighing half as much as the equivalent PV generation system. I think there is definitely value to a simpler system that's more decentralized... but that's harder to quantify.
Edit: people noted I forgot to factor in batteries. 40 kW = 480 kWh per 12 hours. 1 Tesla PowerPack = 220 kWh @ 50 kW. Let's assume a worst case that the base needs the same electrical power during the night as it does in the day, so we need around 5 PowerPacks to get us through each night. 1 PowerPack weighs 1,622 kg, 5 = 8,110 kg. Wow, we need another Falcon Heavy trip just to bring us enough batteries for 1 night! Let's not mention those dust storms that can last for a month or so...
Now nuclear looks much better... and that has its own complexities. Colonising Mars is going to be very hard. :)
We have commercial satellites that also use much cheaper arrays with much higher efficiency.
Using 30 year old solar tech with a very mass-inefficient design is putting your thumb on the scale of such a comparison.
I like kilopower and am a fan (it really helps for super deep space missions that currently have to rely on the tiny amount of Plutonium-238 we have and can make), but we're going to use a LOT of solar power on Mars just as we do today.
A robotic rover can afford to shut down during the night but humans aren't able to do that. So you need to average out the power generated over the day night cycle and add mass for enough batteries to store daytime power for use at night. Last I did the calculations that roughly doubled the weight. So more like a factor of 8 than 2. But still, power generation won't be the majority of the mission weight so it's not a slam dunk.
On the moon it would be. Half-month long nights mean you'd need a lot more batteries if you're making a permanent settlement and aren't at one of those spots on the south pole where you can always see the sun traveling along the horizon.
EDIT:
I think solar panels have gotten a lot lighter since the ISS went up and when I last did the math nuclear and solar were a lot closer.
You are correct. I updated my post with some battery calculations. Only enough battery energy for one night adds another Falcon Heavy of payload. The dust storms will be a huge intermittency problem too.
>The build cost of the ISS arrays was around $300 million (space PV is way more expensive than terrestrial).
That was 20 years ago, on a government contract, for panels that must survive the rigors of actual space.
Panel manufacturing costs have dropped almost 10x since then, and the environment (assumed under the protection of martian atmosphere) which they will operate will be far less extreme than Earth orbit, necessitating less expensive materials. I'm not so sure your numbers account for this.
Space applications use single crystal GaAs (Gallium arsenide) process instead of wide spread c-Si (Crystalline silicon, currently usually multi-Si), due to their superior performance and weight.
Process of producing c-Si panels has become a lot cheaper but this doesn't apply to GaAs process.
Space applications are now mostly using triple-junction cells with GaAs as one of the layers. Spectrolab, Azur Space, and SolAero's product lines are dominated by triple-junction cells. Azur Space also still sells space-qualified silicon cells (lower efficiency and much lower radiation tolerance, but cheaper):
The Martian radiation environment is mild enough that crystalline silicon might still be competitive if you needed large stationary arrays. But on the Martian surface, where you can't rely on near-constant sunlight, nuclear is going to be tough to beat. (Orbiters are a much better match with solar.)
Good points. They do need to be tough enough to withstand Martian dust storms, but apparently they aren't as bad as they look due to the low atmospheric pressure. I read the biggest problem is cleaning the dust off them.
Thanks for this concise comparison. As others have noted, you also need to account for batteries, and battery churn. When the batteries need to be replaced, you have quite a bit of additional payload that needs to be delivered. I am sure that nuclear fuel rods (or pellets, or whatever fuel type they are using) will be dramatically lighter and easier to transport.
Dust will need to be cleaned off of the panels daily, which adds further expense associated with dedicated external missions to the array of panels.
Most of the ancillary considerations weigh in favor of the nuclear option.
Better windmills :D This green energy hoax barely sustains communities on earth, why would sane someone risk on Mars? Mission needs reliability - we have nuclear power sources reliably working for decades, powering bigger things than calculator.
Can you explain what you mean by "green energy hoax [that] barely sustains communities on earth"?
Renewables are part of having a diverse power grid and because of the investment needed they are situated in areas with consistent/reliable wind/solar radiation.
I drive past the massive wind farms of Minnesota and he Dakota's regularly - they are almost always generating electricity. My understanding is that the US has significant potential renewable energy reserves.
Kilopower (NASA's research project for a Martian nuclear fission reactor, from the article): 7,200 kg for 40 kW. [1]
ISS solar arrays: 14,515 kg for ~100 kW in Earth orbit. [2] If we assume 40-50% Earth solar insolation on Martian surface, the ISS PV array probably isn't far off 40 kW output on Mars
Conclusions:
Nuclear reactor would have approx. 50% launch weight. SpaceX estimate $45/kg payload with a Falcon Heavy, so about $324K for Kilopower or ~$600K for the ISS arrays.
The build cost of the ISS arrays was around $300 million (space PV is way more expensive than terrestrial). The development and test costs of Kilopower is around $15 million; build cost of final units is unknown.
You would have to automate and/or remotely control all of the nuclear power plant operations. Dust storms would be a challenge to a PV solution, though not insurmountable.
They actually look very comparable. Nuclear has an edge due to it weighing half as much as the equivalent PV generation system. I think there is definitely value to a simpler system that's more decentralized... but that's harder to quantify.
Edit: people noted I forgot to factor in batteries. 40 kW = 480 kWh per 12 hours. 1 Tesla PowerPack = 220 kWh @ 50 kW. Let's assume a worst case that the base needs the same electrical power during the night as it does in the day, so we need around 5 PowerPacks to get us through each night. 1 PowerPack weighs 1,622 kg, 5 = 8,110 kg. Wow, we need another Falcon Heavy trip just to bring us enough batteries for 1 night! Let's not mention those dust storms that can last for a month or so...
Now nuclear looks much better... and that has its own complexities. Colonising Mars is going to be very hard. :)
[1] https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/201600...
[2] https://www.nasa.gov/mission_pages/station/structure/element...