Moonwards

MolbOrg

00:26

just of gags, 300+ f pages of analysis
>In particular, Project Destiny is taking the challenge of Elon Musk—t o put one million colonists on Mars within the next hundred years at the cost of $200,000 per colonist—and calculating the numbers.

engineering.purdue.edu/AAECourses/aae450/2017/spring/docs/…

This is their conclusion: "If there are no major design changes to the architecture presented by Elon Musk, our team concludes that it is not feasible to establish a 1 million colony on Mars within 100 years using
any of the models we have presented.

Specifically, it is not possible to manufacture the necessary vehicles fast enough for any of the models we investigated. We base this conclusion on our design requirements, assumptions and constraints for how we established our colony on the surface.

yep, moon is superior - with launch window each day

"If the mission were to be completely privatized, the cost per ticket would come out to around $5.7 million per person." - about the price in your timeline

at the end

kim holder

00:45

ok, i'm back

:D i love it

but they didn't put in any new technology, right?

they didn't build nuclear rockets, or tethers, or anything like that?

i think remotely operating robots is also a big deal

kim holder

01:18

@MolbOrg you know, i wanted to talk about the eutectic system aspect of the heat store.
I haven't found information about other possible eutectic systems made out of the elements found on the Moon, but considering the number of minerals there are, and the great complexity of the crystal lattice, and the option of adding one or two percent of something just to tweak the formula, it seems like one can hope there is a formula out there that would allow the pipes to run at 1300 K, or maybe lower.

because i did note how much better the system performs if only you can lower the operating temperature of the mirror pipes.

this is one of those moments where i wish i had an engineering reference book.

MolbOrg

@kimholder Haven't read it yet - conclusion is from a good samaritan. consider to read, because I would like to challenge them and so they could wind up their analysis a notch next year ))

@kimholder and that is the problem with your timeline (but I'm half trough so should not say so but so far)

^will not discuss

kim holder

but if you are half way through, you should already be through the part where tele-operation of robots is an important thing

MolbOrg

yes, that is the reason I'm saying that at all

kim holder

alright, that is mysterious for me then

MolbOrg

would be not if we talked about bridge already

MolbOrg

01:30

eutectic

eutectic usually does not mean few percentages, 5-10-15-20-40-50

kim holder

you really can't stand not talking about your bridge concept, can you?

MolbOrg

it depends on what you mix and how much componetns

kim holder

That Alex Ellery paper you posted a few days ago? The one at Space.com, where he is building self-replicating robots? i feel a need to say they are many years away from having a working model. I have that paper, in full. It was sent to me by one of his former grad students.

i got into a conversation about that same paper with someone else, who emailed me about it. He may well feel the way i suspect you do - that is the giant game changer. But only when it happens. They aren't exactly close to their goal. They can barely make their pieces do anything at all, never mind assembling them into an actual robot capable of doing anything.

So, hopefully you will be satisfied when you get to the part about building the giant city. That's where you will find your robots.

MolbOrg

however, considering there is some industrial base, and ability to refine things - calcium probably could lower the melting point. the need to refine and prepare steals some pros - but realistically speaking some preparing process is expected, so not a big deal.
as I said earlier good candidates could be metals, especially if there is not need to reduce them, if it is possible just collect pure metals, even if it is not high percentage yield in the dust.

but at the end of the day, it needs to look at Ca+SiO2 for the tank

kim holder

hm.

you know, if it wasn't necessary to have 3 million tons of aluminum to fill the tank, i'd think about it

but it is just too much to think about having that capacity until long after this point in the colony

especially when you'd need hundreds of the things down the road.

MolbOrg

01:46

metals have their own drawbacks - they are reducing agents so they may react with oxidizing agents aka ceramics thus they may require special ceramics. For aluminum, there was some, not sure for iron.

kim holder

still doesn't seem as big an issue as getting thousands of gigatons of the stuff.

oh sorry - only thousands of megatons

a few gigatons

MolbOrg

the problem with the 3 million number is that I didn't found that nasa link about a content of pure metals inclusions in the dust, it had some percentages.
but why you talk about 3 million tons of aluminum with 10MW of electrical

kim holder

oh look at that. zinc boils at 1180 K...

MolbOrg

we talking about Al

kim holder

yeah, i know, i just wanted to put that down before i forgot

the unit as it is produces about 700 kW with a tank that holds 11 million .... oh - no, its kg...

ok, yes. big difference. :P

still. 3 kilotons of aluminum

the only metal i have ever heard about being present in the regolith in pure form is iron

there has been discussion that in the central mountains of some craters - the ones where the impact was at a low enough speed - part of the original asteroid would still be there

and there you might find rich deposits of pure metals

still. asteroids are not known for their aluminum content.

and at any rate that is another gap in my knowledge base. i need a bunch of good breakdowns of the metal content of various kinds of meteors.

let's see what Lunar Sourcebook has on pure metal content. i bet it is going to be parts per million.

(dinner time)

(note to self - zinc content of asteroids is much better...)

MolbOrg

02:06

hm, you won. my IronWill isn't that strong, I have to speak.
I do not care about the guy (with the self-replicating robot). I do not need to read papers to see that it is a crap. For me, it is enough just one photo they have. The reason why the link was posted - because they made the contract with DARPA(this month), and because they got financing of $750'000 and $500'000 in April this year from nasa and some satellite provider.
The idea itself is good, but context of it just pure crap so as the current state of development.

Ok, 3000 tons of alluminium - let's say there is 0.5% in the dust in pure form - we should refine 200 times more of the dust - less than 200'000 cubic meters of dust

correction
Ok, 3000 tons of alluminium - let's say there is 0.5% in the dust in pure form - we should refine 200 times more of the dust - less than 240'000 cubic meters of dust

it is questionable but not unrealistic

the question is - Is I correct about pure metals in regolith dust, and how much and how to extract.
but if I'm correct then opportunity here

kim holder

@MolbOrg no.

there is iron.

MolbOrg

I know that you disagree on that. But I certain what I have seen and what I have remembered.
but if you agree on iron, then also a good pick

everything is good if it is easy to extract - any metal can be extracted magnetically - any conductor.

for iron, it is required 2000 ton/overall efficency of the system - so 6-8 thousand tons

and iron would be a good pick, because it does not melt at 1300K, because it is relatively ok heat conductor, because you can use other ways to heat it etc

considering that overall efficiency of the system is about 26%

kim holder

02:47

i have spent a lot of time looking at the Lunar Sourcebook. i'll go look for a source in a while, can't right now. i'm pretty confident there are no samples of lunar material that have free metals in anything more that parts per million quantities.

as in, 50 ppm, that kind of level

however my interest is in something with a low melting point. i am (as you know) very attached to the system working just as well at night as it does during the day, and that requires getting the heat of fusion out of it. iron won't do.

MolbOrg

I'm not suggesting - I just saying about some pros to use iron as an example.

mirrors and radiators could be one block with generator and engine.
A system of let's 100kW output(10x20m). On a single frame, assembled in the base, transported to installation pace, plugged electric cable.
A tank is used as heat storage with iron, heated by electricity(as an example), used only overnight to support the base only.
Expenses are 3 times more electric generators(iron, aluminum), 3 times more heat engines(iron) (100W/kg), electrical lines instead of pipes(aluminum)

@kimholder you should crunch numbers, I suggest new sheet tab - lets compare possible medias for tank

*I definitely should find the thing I saw, I tried one-day but without success

MolbOrg

03:39

Ok a bit ugly in therms of input data, but more or less represents the situation, the sheet heattank

aluminum beats the crap out of regolith - the only drawback collection of the alluminium, the rest is sweet for it

Iron it needs about 4 times of the regolith by mass, but it is about 1.3 more by volume

it needs to separate metals, but it also gives metal heat conductivity - less pipes, iron gives no phase change which is good in some regard(can be used)

MolbOrg

04:01

lowering heat media from 1450K to 1100K (in mirror system) increases overall efficiency 2 times from 16% to 33%

according to spread sheet

kim holder

well, that's a lot to think about. i'll have to get to it tomorrow though. very interesting.

except - what is q/kg?

MolbOrg

04:18

J/kg. Heat joule

kim holder

the figures for regolith don't look right

MolbOrg

yes, took average

kim holder

i have to go though. it's already past our bedtime

MolbOrg

thats why ugly data

if we take 1500 instead of 1250 heat capacity , nothing changes until melting point(100% melt of regolith) is begins slightly beat aluminium after that

hm, however, yes probably there is error

no, no error

if we take 1500 J/K,kg

however there is no easy way to account that melting point of regolith isn't fixed, it is possible to create more accurate table but it does not change end result at 1450+ temp (until iron melts), and it does not change results for lower temperatures

however results aren't precise by any mean, they are a just illustration of the situation, nothing more than that

MolbOrg

10:41

arxiv.org/pdf/1410.6865.pdf

A more readily available source might be native Fe in the regolith, although the
concentrations are quite low (~0.5 wt%; Morris, 1980). Native Fe in the regolith has at
least three sources (e.g. Morris, 1980): meteoritic iron, iron released from disaggregated
bedrock sources, and iron produced by the reduction of iron-oxides in the regolith by
solar wind hydrogen. The latter component occurs as nanometre sized blebs (often
referred to as ‘nano-phase Fe’) within impact glass particles (‘agglutinates’) and is

If the practical difficulties could be overcome, a native Fe concentration of
~0.3 wt% is not an entirely negligible amount, corresponding to about 5 kg for 1 m3
of
regolith. Moreover, meteoritic iron will be associated with siderophile elements, some
of which (e.g. nickel, the PGMs and gold) are valuable owing to their catalytic and/or
electrical properties; at concentrations typical of meteoritic Fe, 1m3
of regolith could
potentially yield 300g of Ni and 0.5g of PGMs.

I crunched numbers a bit (rough estimation to get some sense of things)

using pure metals sucks if their extraction cost is more than 100kJ per kg. so just extraction by reduction sucks

however, those nanoparticles will be bound with He and other Solar wind implanted volatiles - which are required

kim holder

14:20

ah, that is indeed a fine paper. haven't read it in a while

kim holder

14:34

here is an alternate source

lpi.usra.edu/publications/books/lunar_sourcebook

that is from rocks, i suppose bulk regolith would be the same

but basically, your choice is iron, or nickel, and there is a great deal of variation in the nickel depending on the site

volatiles can be collected by just pumping out the thin vapors that come off heated regolith. but that won't get those nanoparticles

the cost of pure metals in terms of energy needs to be considered in terms of how much of that energy can be gotten from direct sunlight.

the fluidized bed reactors are designed with their own parabolic mirrors for that reason.

and the dust roaster is an interesting item as well. i thought about calculating how long you'd need to run a bunch of them to get enough aluminum to fill a tank, but left it.

btw, the numbers you have for heat capacity don't consider phase change, which is the part the entire power plant is based on. and it doesn't take the data from the reference doc we've been using, and has been listed here before.

the range that counts has an average J*kg/K of 2900

it's even on our spreadsheet, along with the values over a range of other temperatures

@MolbOrg i don't understand the approach of putting the absolute amount of energy in a kg of material. Isn't what counts the amount of energy drawn by cooling it? In which case, the relevant thing is the energy per kg per K?

ptable.com/#Property/Heat/Specific

Also, major note. i didn't notice that there were more tabs on the spreadsheet until you mentioned adding one

i was going to ask you to share your heat field spreadsheet, but decided it should probably wait. but it is right there

MolbOrg

17:59

"btw, the numbers you have for heat capacity don't consider phase change, which is the part the entire power plant is based on."
they do, pay attention to green cells and previous values for the column

@kimholder no, my choice is anything that is easy to extract, if at all.

@kimholder a looong

yes, it is the sum of hidden capacity and regular capacity.
with regolith cells

\=if(B20>$F$4, $G$4, 0)+(B20-$C$6)*$B$4+if(B20>$F$4, (B20-$F$4)*$C$4,0)

kim holder

@MolbOrg that was an english phrase there, what i meant was one's choice is iron or nickel

alright, that was sloppy on my part

MolbOrg

it is sum of 3 - first is hidden capacity if temperature above melting point, second heat capacity before melting, third heat capacity after melting

I could make some error, but it look legit, and it is good seeking for them, they happen

no, I'm not targeting for the nickel for sure, the iron percentage is on the border, nickel is a percentage of the percentage, Al(if there is) or Iron(those 0.3-0.5 wt%, of parts of meteorites)

kim holder

i have never seen any reference to the presence of aluminum in stony or metallic asteroids

MolbOrg

anything which requires less than 100kJ/kg to extract, if there is nothing for that energy than using metals for the tank is a bad idea because it slows down the process significantly.

@kimholder we leave that for now, there is Fe, it's good enough for the discussion, if there is Al well that's good too, if not not a problem.

kim holder

yeah, i am mostly thinking about the possibility of filling tanks with metal over the long term, after there is much more infrastructure

MolbOrg

18:11

@kimholder wrong answer

kim holder

i'm afraid you don't get to decide that

i was going to add, what i really am more hopeful about, is the existence of a eutectic system with a lower melting point

MolbOrg

this is some rough assumptions for materials needed per 500kW setup. Wery rough, and more on lower end of the things

pink fields is amount of days needed for those 500kW to produce enough energy to extract melt etc those materials required to building the thing and probably the building energy itself

the first pink number - is without tank materials, the second with tank materials

I assumed 10MJ/kg should be enough to melt, form the material, bend and anything else to make the construction

it is a low number for Al if there is need to extract it from Al2O3 it is low in about 10 times then

It is ok number for Fe if we mix available raw iron (that 0.5 %) with extraction from oxides.

kim holder

i could get into a pretty complicated discussion about this

MolbOrg

the efficiency of engines generators - overall efficiency assumed to be 25% for the setup

so, it should be seen that using pure metals for the tank is way more time expensive idea, than do not use them

using them for a tank requires about 4 years compared to 3.5 days in this scenario (which is imperfect but still)

kim holder

long term. what does that mean to you?

MolbOrg

18:25

so the optimal strategy is not using metals for the tank, in the assumptions.

kim holder

there are a lot of assumptions there

at any rate, we got into this mostly because we were talking about how helpful it would be, if aluminum could be used. so helpful it was worth checking if there is a way

there isn't in the time frame the current colony deals with.

in the longer term, there is potential for possibly doing it, but i don't even want to get into it. if i have time, i'll start sketching an initial rough draft of part of a potential solution.

i'm mostly interested in eutectic systems at this point.

MolbOrg

18:51

Ok, the last word from me about the previous topic. Energy payback time is an important property of any system and should be considered, especially if you interested in faster growth and you have to choose between technologies and such. Same as - energy returned on energy invested (EROI).
If the link is not misleading(too good from what I have expected, and especially considering +30% solar intensity on moon)
https://en.wikipedia.org/wiki/Crystalline_silicon#Energy_payback_time
Mono-Si 3.3y, MultiCrystal-Si 2.2y, amorphous-Si 0.9y (seems to be too good to be true, and needs verification, b…

"i'm mostly interested in eutectic systems at this point." - what is the question here

kim holder

energy payback time, however, should not count energy you get mostly for free, or energy spent to do something else, that yielded a byproduct

MolbOrg

"i'm mostly interested in eutectic systems at this point." - what is the question here

kim holder

there isn't really a question about eutectic systems, unless you have insight on how to go looking for one. i was actually planning to just note in the presentation of this that, though it would work, it would benefit hugely from having a heat storage mass that melts at a lower temperature, and can be composed of common materials on the Moon, and explain eutectic systems

hang on - i needed a minute

i'm writing a blog post now about the new model. i'll explain the issues there, and note the greatly expanded spreadsheet.

i have a presentation at a conference on Aug 9 - at the Starship Congress.

i must have mentioned that... at any rate, there is going to be a panel discussion afterwards, and that panel now includes Haym Benaroya, David Livingston, John Jossy, and me.

i am particularly concerned about Haym. he is one of the world's leading experts on construction in space.

he is on the panel because David talked him into it. i would really like to make a good impression on him, it is a very unique opportunity.

now that the model is done, i need to set it up for presentation, which is fairly time consuming

and i need to finish a bunch of modelling on the habitats of the colony.

i'd love to get into a long complex conversation with you about how to properly think about energy budgets in an extremely complex long-range colony - but i am resisting because i am anxious my time is already too short to prepare properly

i did go look at your suggestion regarding calcium and silicon dioxide, but i didn't get far. there must be something about the proportions.

MolbOrg

19:32

yep, Ca slicate was bad guess of mine. It is part of cement chemistry, and basically, one of the things which-which makes cement a cement.

Calcium silicate

Calcium silicate (often referred to by its shortened trade name Cal-Sil or Calsil) is the chemical compound Ca2SiO4, also known as calcium orthosilicate and sometimes formulated 2CaO.SiO2. It is one of a group of compounds obtained by reacting calcium oxide and silica in various ratios e.g. 3CaO•SiO2, Ca3SiO5; 2CaO•SiO2, Ca2SiO4; 3CaO•2SiO2, Ca3Si2O7 and CaO•SiO2, CaSiO3. Calcium silicate is a white free-flowing powder derived from limestone and diatomaceous earth. It has a low bulk density and high physical water absorption. It is used in roads, insulation, bricks, roof tiles, table salt and occurs...

melting point around 1540C

and the thing added for different reasons than lowering melting point - making the glass water resistant, making proper forming temperature etc (my guesses, understanding for the moment)

so basically geology which you looked earlier is good source of such compositions

Olivine

The mineral olivine ( /ˈɒlᵻˌviːn/) is a magnesium iron silicate with the formula (Mg2+, Fe2+)2SiO4. Thus it is a type of nesosilicate or orthosilicate. It is a common mineral in the Earth's subsurface but weathers quickly on the surface. The ratio of magnesium and iron varies between the two endmembers of the solid solution series: forsterite (Mg-endmember: Mg2SiO4) and fayalite (Fe-endmember: Fe2SiO4). Compositions of olivine are commonly expressed as molar percentages of forsterite (Fo) and fayalite (Fa) (e.g., Fo70Fa30). Forsterite has an unusually high melting temperature at atmospheric pressure...

kim holder

yes, but the information on eutectic systems i've found is all about ones found in nature, not ones that have been deliberately designed.

MolbOrg

The ratio of magnesium and iron varies between the two endmembers of the solid solution series: forsterite (Mg-endmember: Mg2SiO4) and fayalite (Fe-endmember: Fe2SiO4). Compositions of olivine are commonly expressed as molar percentages of forsterite (Fo) and fayalite (Fa) (e.g., Fo70Fa30). Forsterite has an unusually high melting temperature at atmospheric pressure, almost 1,900 °C (3,450 °F), but the melting temperature of fayalite is much lower (about 1,200 °C [2,190 °F]).
The melting temperature varies smoothly between the two endmembers, as do other properties. Olivine incorporates onl…

so fayalite - 1473K

so seems that there no much combinations and none of them seems to be at low point

kim holder

but say that in consideration of fayalite being the common mineral with the lowest melting point, that mineral is purified - which we already have a process for (maybe not optimized, but we have one)

and then mixtures are sought where fayalite is combined with other chemicals, some of which may even be imported, if they are worth the trouble.

kim holder

20:45

21

Q: Why do felsic materials have lower melting points than mafic?

It is clear from Bowen's reaction series that more felsic minerals have lower melting points than mafic minerals. As far as I know, the same is true of quenched glasses. Felsics have a higher degree of SiO2 polymerization in the solid phase, which I would have thought was energetically favorable...

academic.brooklyn.cuny.edu/geology/powell/courses/geol7040/…

MolbOrg

21:24

I should say that negative result is also useful result

needs to go back a bit. What is the reason to lower temperature in tank?

kim holder

my big thing about my design is it needs to get most of its heat from the freezing process overnight. that way the heat engines can be run at the same temperature all night, and have the same output as they do during the day

material that freezes at a lower temperature allows more heat to be brought to the tank from the mirrors, and then be stored for use overnight

geo.arizona.edu/xtal/group/pdf/AM92_1573.pdf

^ The albite fusion curve re-examined

MolbOrg

21:47

So, you need a melting point to have the constant temperature on the hot end of heat engine - is that correct?
And on second first place is because it has great deal in storing energy - is that correct?

kim holder

yes

MolbOrg

good.

MolbOrg

22:00

let's take look at constant temperature
imagine 10 separate well insulated buckets of water at 100C. through them goes a pipe - one bucket by one bucket (sequential).
we pump a liquid/gas trough the pipe(out heat carrier) at let say 0C input at relatively slow speed
which buckets will cool faster?

kim holder

the pipes inside the heat tank have been designed to distribute the temperature overnight so that it doesn't freeze first on the walls and prevent heat from reaching them. that is the reason for some of the details of the heat tank.

if that is what you are getting at....

MolbOrg

No, I do not.

I'm getting at constant temperature for a heat engine

kim holder

well, the first bucket cools faster because the incoming fluid is coldest there and it has the biggest heat gradient.

we have to simplify it that way

MolbOrg

good, what about exit temperature?

kim holder

we went over this to some degree for the cold end, didn't we?

of course there isn't a constant temperature, but the piston is in a chamber filled with gas, the heat there will redistribute quickly.

MolbOrg

22:09

maybe, but as I said, we roll back a bit

we still about buckets of water from our experimetn

kim holder

and i'm still a very busy person

can you skip to the end and see if i understand without going through a bunch of steps?

MolbOrg

sorry, if you say stop, I'll stop.

kim holder

you are helping, a lot, i'm just trying to make the best use of my time i can. so it might be useful to take the chance i'll understand your point without the thought experiment, which might take a while.

MolbOrg

we just one question from conclusion - this bucket experiment is important, I'm not sure you know or understand it

if it is hard to you, then I can say that exit temperature will be about 100C(a bit less, and it depends on situations, but good enough)

kim holder

well, for future reference, try me first, and then go backwards. there is a lot i don't know, but the moon sunlight thing is fresh in my mind.

MolbOrg

22:13

the answer is that you can have constant temperature

you do not need melting point for that

roughly speaking, but close to the reality

kim holder

the heat you draw out lowers the temperature in the tank. unless the heat you draw is from a phase change. if we were just talking about a few hours, i could see how you could juggle it around, but over 2 weeks of night?

MolbOrg

time is not important there, it depends on the amount of energy available, it is like a syringe filled with heat. and moving heat carrier like pressing on the syringe-piston

the deal is that those buckets do not have to exchange heat at fast rate between each other

and you might have spare buckets, and when the exit temperature is below some temperature, add them in the system - if it is easier to imagine

kim holder

they are still getting colder as they lose their heat. unless you do some trick like compression / expansion to extract more heat by forcing a phase change in the carrier

MolbOrg

"they are still getting colder as they lose their heat. " - you have insulation fro that

kim holder

insulation doesn't seem relevant. the heat in the system decreases.

MolbOrg

22:23

yes, but it decreases at different speed at different parts of the system, and you have control over it

kim holder

there is also the fact that the phase change has the highest energy storage of any temperature

aside from trying to drain the heat store sequentially, the tank would have to become much larger to get the same energy.

and sequential draining would require the tank be split into all sorts of sections, and have a much larger surface area.

the nice thing about finding a eutectic system that has a good freezing temperature and doesn't require a lot of processing or imported chemicals is that it would be so elegant.

And - this point is important - it doesn't require me to change the entire freaking model when i only have 7 weeks before the conference

it seemed worth it to spend a couple of hours and quickly check about that, because it makes the whole idea more appealing.

MolbOrg

@kimholder you do not have to

kim holder

Majorly change the model? Can't do it. It's too late. I don't have time.

MolbOrg

essentially they are the same

kim holder

the heat will move unless it is kept in the same place

MolbOrg

22:31

@kimholder I said, no changes required. they are the same things

@kimholder it depends on heat conductivity, which is small for regolit rock or dust

size

kim holder

not for molten regolith or anything molten that might go in there. Not for 2 full weeks.

besides, the tank has been specifically designed to circulate the heat.

MolbOrg

the difference between molten 1450K and solid 700K is 2 times in size of the dimensions - 2 times higher, 2 times wider

kim holder

2 times more pipes

and still - how do you keep the heat in one place?

MolbOrg

no one keeps it in one place, it just flows too slow between parts of the system

kim holder

that's avoiding the real question here.

MolbOrg

22:38

which one?

kim holder

keeping heat inside a well-insulated tank is one thing. keeping heat in all sorts of places within a tank is another.

MolbOrg

it is done by flow of the heat media

do you have link to tank - current state drop it here

kim holder

among other things... it's the other things that are problematic.

MolbOrg

eh? what is problematic there

kim holder

i'm writing up the blog post. you'll see it all in detail.

MolbOrg

22:43

kk

tank link

kim holder

i'll upload the current model to the repo. the tank is on a layer with a few other things.

MolbOrg

kk

kim holder

ok, it's there. here is a shot that shows you what layers it is on

now, in getting that ready, it occurs to me that i set up the stone mass around the central tank to be a buffer zone that also acts as a heat reservoir, and i set up the pipes to run through it and distribute the inner tank heat to do the very thing you say

but, it didn't seem to me that it could be enough to keep the heat even on its own. it is there to keep the heat inside the whole tank complex better.

Hoh explained it as a counterflow.... something....

i made them fewer, actually, because i decided i'd gone overboard.

here is a shot of the pipes that go to and from the hot end of the stirling engines. by reversing the direction of flow, the heat can be drawn mostly from the wall of the lava tank during the day, and mostly from the outer edges of the stone buffer mass during the night.

there it is with the stone in dark grey, and the heat buffer made of powder regolith outside of all of that to prevent heat seeping out of the whole heat system, in light gray

Transcript for