Yeah, but flipside, you're probably gonna need things on the Moon that you wouldn't for a terrestrial base. I think it's fair to assume that the $7bn a year figure is just a very rough number, but still a useful one for understanding just how outrageously expensive it would be to maintain a Moon base.
You kind of have a model of a lunar space station in the form of ISS. That's definitely not 7000 tonnes of supplies for a base. Maybe something like 15 tonnes per year or so.
ISS currently has 7 people on it, McMurdo Station varies between 250 and 1258. I don’t know the mass/year of the ISS resupply missions, neither including or excluding orbit reboosting fuel which a moon base can clearly ignore.
I'll take a look at the exact figures, but assuming the average number of people at McMurdo is 750, it's 9 tonnes per person per year, whereas in the case of ISS, it should be something like one third of this number per crew member at worst. Clearly they've already minimized the necessary supplies somewhat already.
A big difference is the ISS uses 24/7 solar where McMurdo can’t. Unfortunately powering a moon base is quite difficult do to the month long day night cycle, so it’s going to need nuclear or a lot of batteries.
That said, the ISS doesn’t do long distance missions dropping off equipment etc the way McMurdo and presumably any Moon Base would. So, the ISS isn’t that great of a model.
ISS doesn't use "24/7 solar". Insolation of the ISS depends on the beta angle and therefore ISS crucially relies on batteries for its continuous operation.
Here's the thing about a lunar base on the south pole of the Moon: It's assumed that it would manufacture its own propellant from local resources (water). That means that you're already storing electricity electrochemically because that's exactly how you make propellant from water. But that means that it's just a small leap to make the storage tanks slightly larger and siphon off a small portion of the hydrogen and oxygen to generate "baseline" electricity for the station. You already have all the equipment anyway, why not just add a fuel cell? We can do small fuel cells more easily than we can do small reactors anyway. And even just one spare cubic meter of your hydrogen tank provides you with 1.3 MWh of electricity. One extra meter of length on a 4.5 diameter tank provides you with over 10 MWh of electricity.
> You already have all the equipment anyway, why not just add a fuel cell?
According to the blog post, the dry mass of the fuel cells would mass similar to LiIon of the same energy storage capacity, not counting the mining equipment to get the lunar water.
That makes no sense whatsoever; fuel cell mass scales with power output, not with system capacity. Extra capacity incurs just slightly bigger tanks. No serious study I've seen assumes that long-duration storage (~100h or more) has fuel cells heavier than batteries. The last paper I've seen had batteries twice as heavy as the whole fuel cell system. I may try to find it again.
You need a part of the system anyway so you can't count those parts (water tanks, hydrogen tanks, oxygen tanks, electrolyzer) into any of the power options. They will be there regardless of whether you want them or not.
Also I've noticed that despite of what ben_w claims the article says, the article still assumes 100 kWh/t for a battery system and 1 MWh/t for a fuel cell system. So it doesn't really say that "the dry mass of the fuel cells would mass similar to LiIon of the same energy storage capacity", otherwise the numbers would be in the opposite order.
Right next to where it suggests 1 MWh/t “For the sake of argument”, it says:
"""Even using Lunar water, the mass of the cells, condensers, electrolysers, power electronics, storage tanks, and heat exchangers is comparable a Lithium-ion battery, while the round trip efficiency is much lower. And that’s not even including the mass overhead for mining lunar water!"""
Yes, but that makes no sense. Partly because as I said, lots of those components will be mandatory anyway so they should count into the fuel cell system at best only partially, and partly because pretty much the only part that scales with capacity is the tanks anyway. The rest scales overwhelmingly with average power instead.
Due to shipping and redundancy concerns redundant tanks are used to expand storage. So, the dry mass of tanks should scale nearly linearly with the total amount of water, hydrogen, and air stored.
Nobody said they don't. But 1) that doesn't make the fuel cells bigger (so the "similar dry mass of the fuel cells to batteries of a given capacity" is a category error), and 2) the additional tank mass is very mild. We can for example store 20.8 tonnes of hydrogen/oxygen mixture in a ~2 tonne tank. We know this because the Centaur stage does it. And those two tonnes of tankage correspond to ~50 MWh of electricity generated by a fuel cell. That's almost 300 tonnes worth of batteries in the form of 23 tonnes of mass, 21 of which might even not need transportation from Earth. Potentially up to a 100:1 win for Team Fuel Cells! ("Only" ~10:1 if you do carry water from Earth, but that's still substantial.)
Sure, dry mass of fuel cells vs dry mass for fuel cells vs total mass need for fuel cells.
Also, that’s just the start. Team fuel cells are also dramatically lower efficiency so you need a extra solar panels and all the associated equipment with that. Which again scales linearly with energy demands.
For non tracking panels at 100% efficiency and 200W/kg assuming an optimistic 1/2 of total weight for solar collection is panels runs around 50 kg per MWh per month on the moon.
Batteries at ~90% bump that to 55kg/MWH. Fuel cells at ~30% bump that to ~170 kg or another 110kg /MWh, not that significant but still important.
How does panel efficiency go into this? Those 200 W/kg are post-conversion, so whatever efficiency they have is already included.
> assuming an optimistic 1/2 of total weight for solar collection is panels
What does that even mean? 200 W/kg is system efficiency, so not just cells but all the structures involved.
> runs around 50 kg per MWh per month on the moon.
I don't get it. "MWh per month" is 114 watts. A 50 kg array at 200 W/kg has 10 kW peak. At the south pole, a non-tracking unifacial horizontally mounted and elevated array will have at least a ~30% capacity factor, a bifacial array will have a ~60% capacity factor. That's 3 kW or 6 kW on average, respectively, far from your 114 watts.
EDIT: Should have really been 1.4 kW instead of 114 watts - I calculated with 1 MWh per year by a mistake. Still a substantial difference from 3/6 kWh.
> Fuel cells at ~30%
A round-trip from a fuel-cell system will be at ~40%. Maybe ~50% if you use top equipment. But even bog-standard fuel cells are at ~50-55% (LHV) efficiency and bog-standard PEM electrolyzers are at ~80% efficiency, which should give you no less than a 40% round-trip.
It doesn’t but a measure of output from solar panels ignores the storage efficiency which is what we care about here.
> 200w/kg is system efficiency.
You listed that as the panels in your post.
First actual array output falls over time. Whatever it would be new is irrelevant you scale for how long it’s in use. Panels on the moon need to be mounted at an appropriate angle for maximum efficiency, cooling, they need wires to move electric from the panel to your power conversion, electronics to handle that power etc. You also get losses from lunar dust, etc. Simply saying all that doubles weight is a reasonable ballpark.
Don’t forget the moon is dealing with days of full sun panels need to maintain low temperatures for efficiency and you can’t just dump all that heat into the moon.
> I don’t get it
A 50kg array consisting of 25kg of solar panels, at 30% efficiency produces: 25 different 0.2kw panels for 30 percent of the time over 28 days times 24 hours = 25 * 0.200kw * 0.30 * 28 * 24h = 1008 kWh or 1.008 MWh.
Edit: 1.4kW * 28 days * 24h/day = 940 kWh or 0.94 MWh.
> A round-trip from a fuel-cell system will be at ~40%.
I have yet to read about an actual working system over 30% water to hydrogen to water. Do you have any citations or is this assuming some unknown breakthrough?
> It doesn’t but a measure of output from solar panels ignores the storage efficiency which is what we care about here.
So why even mention panel efficiency?
> First actual array output falls over time
You always compensate for it, it's not hard.
> Panels on the moon need to be mounted at an appropriate angle for maximum efficiency, cooling, they need wires to move electric from the panel to your power conversion
All included in that figure.
> electronics to handle that power
That is admittedly not included in that figure but PPUs are still fairly lightweight these days.
> Don’t forget the moon is dealing with days of full sun panels need to maintain low temperatures for efficiency
We've already dealt with these things for geostationary satellites, and those are almost always insolated thanks to their orbit.
> and you can’t just dump all that heat into the moon
On the south pole you most likely can, since even though there's lots of sunlight, the extreme incidence angle means that the surface is disproportionately cool.
> A 50kg array consisting of 25kg of solar panels, at 30% efficiency produces: 25 different 0.2kw panels for 30 percent of the time over 28 days times 24 hours = 25 * 0.200kw * 0.30 * 28 * 24h = 1008 kWh or 1.008 MWh.
Heh? A 50 kg array at 200 W/kg produces 10 kW peak and even at 30% capacity factor generates ~2 MWh in a lunar month: 100.324*29.5 = 2.124 MWh.
> consisting of 25kg of solar panels
I already said that the ~200 W/kg for UltraFlex/MegaFlex arrays includes structures, so there's no "25kg of solar panels". There are in fact no panels at all - on UltraFlex/MegaFlex, individual "naked" cells are attached to the flexible substrate of the array directly.
> I have yet to read about an actual working system over 30% water to hydrogen to water. Do you have any citations or is this assuming some unknown breakthrough?
No, it assumes perfectly standard system components. But here you have NASA's RFC demonstrating 50% efficiency in 2006 (already fifteen years ago!): https://ntrs.nasa.gov/citations/20060008706
Those fit on satellites in zero g. The moon has gravity so you support them at an angle you need bracing on something else, or for extra long panels connected to a structure you need bracing inside these ultra flex panels.
In space panels can radiate from their back sides into space, but that doesn’t work on the moon as it hits 260 degrees Fahrenheit in the day. Look up panel efficiency curves with temperature. Cooling during the day is a major concern for a long term lunar base, but also needed for panels.
Further you need cables from wherever they are to where up your using power.
Lunar dust again is an issue for solar panels on the moon vs satellites.
Sure, and that worked wonderfully a lab. But, production systems need to worry about a great deal of stuff that doesn’t apply in a lab setting. I don’t mean this dismissively yes I completely agree in theory it could work, but it’s just not working technology yet. I haven’t looked recently so hopefully there is a system demonstration out there.
>Those fit on satellites in zero g. The moon has gravity so you support them at an angle you need bracing on something else
UltraFlex/MegaFlex is rated for 3g acceleration on those satellites even when fanned out, why do you think it would have problems with 0.16g? That's one twentieth of what they're supposed to withstand.
>but that doesn’t work on the moon as it hits 260 degrees Fahrenheit in the day
> Further you need cables from wherever they are to where up your using power.
High-voltage cables are not that heavy.
> Sure, and that worked wonderfully a lab. But, production systems need to worry about a great deal of stuff that doesn’t apply in a lab setting. I don’t mean this dismissively yes I completely agree in theory it could work, but it’s just not working technology yet.
Of course you don't; you're just disingenuously moving the goalposts.
As dredmorbius points out, the ISS gets solar every hour of every day, so it really is 24/7. What it’s not is 60/24/7, because up to ~45 minutes of each hour may be in earths shadow.
I may not be a native English speaker, but Wikipedia tells me that "In commerce and industry, 24/7 or 24-7 service (usually pronounced "twenty-four seven") is service that is available at any time and usually, every day ... Synonyms include round-the-clock service (with/without hyphens), especially in British English, and nonstop service".
I struggle to fit "at some times during every hour" into this definition. If there's a 24/7 shop, I definitely don't expect not being able to enter it at some times, or even not being able to enter it 75% of the time.
In practice it’s common to use similar terminology for intermittent services.
For example a subway train might only be running every 30 minutes but it still lists hours of operation. The hotel help desk might not have someone there every second, but theirs a difference between someone putting up a “be back in 15 minutes sign” when their on a bathroom break rather than a closed sign when they are leaving for the day. Informally, the existence of a modest wait doesn’t preclude 24/7 service.
I definitely don't picture a 24/7 service as waiting up to 45 minutes every hour. By that time you're operating six hours out of every twenty four. You might not even be able to advertise yourself as 24/7 at that point in some places.
> For example a subway train might only be running every 30 minutes but it still lists hours of operation.
But that's a train. It never runs every single moment, so that doesn't make the times with once-per-30-minutes trains different from the times with once-per-10-minutes trains. On the other hand, if it's for example a grocery store, you definitely do expect being able to enter at any moment and purchase something if it advertises itself as "24/7" or "nonstop".
As a native English speaker, I agree with your assessment of “24/7”; linguistics aside, the core point of the argument is (or should be) that ISS power cycle is 90 minutes whereas a moon base would be 28 days, unless you can get power from the sun side to the dark side.
I’m unclear why the blog post is saying this is difficult, as it should be fairly straightforward to make aluminium cables in-situ and just leave them exposed on the surface, the atmospheric pressure on the moon is far lower than the minimum of the graphs on the linked Paschen's law Wikipedia page. (I’m not a physicist, there is a high chance I’m overlooking something any vacuum engineer would consider obvious).
If you're mentioning the cycle lengths, the ISS actually has it worse relatively to the length of its cycle than a south pole base on the Moon would have it since ISS can spend up to around 40% of its time in Earth's shadow whereas lunar night can be as little as several days long on some places of the south pole. And if you elevate the solar array by as little as several meters, I believe I saw some papers predicting a period without power as short as 2.5-3 days or so. So rather than a ~500:1 ratio of storage difficulty, it's at least improved to ~100:1 or so. Still fairly bad but not as bad as the simple cycle length ratio would suggest.
The notion of connecting several places with cables is a nice one, and also a neat optimization problem. We should be able to calculate the immediate power curves in different places. Finding a minimum cost combination could be the topic for an interesting study. Long-term, this is definitely what you'd want to do.
> you definitely do expect being able to enter at any moment and purchase something if it advertises itself as "24/7"
But what’s an acceptable delay before purchase. I have definitely spent more than 45 minutes waiting in the line for events etc while their currently open.
It’s all kind of moot as it’s clear what I meant is exactly the underlying reality, but I can see why you object.
