Elon Musk: glass geodesic domes

[Lampyridae]

50 kPa may be doable with higher oxygen content (somthing like 8psi, 32% Oxygen...or 8/32),

That would drastically increase your fire risk.

As for atmosphere, I'll check my source but my understanding was 8/32 had the right balance of increase percentage and decreased pressure to maintain a similar flammability.

Looking at what I have at hand, something which is on the just self-extinguishing at 14.7psi & 21% Oxygen will self-extinguish at 12/23 and 7/28. Therefore 8/32 is well over the line.

That's at 1g. At Mars gravity, flammability is even higher.

https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20130010991.pdf

In general, higher oxygen concentrations are bad news at low g. I didn't know this before. But 12-13psi / 0.8bar with the same oxygen concentration? That's fine. I'm breathing that right now.

MOC is the minimum oxygen concentration to continue burning (basically like 1% chance of sustained fire). ULOI is a concentration where there 50% chance that the material continues to burn.

[Lampyridae]

50 kPa may be doable with higher oxygen content (somthing like 8psi, 32% Oxygen...or 8/32),

That would drastically increase your fire risk.

[Nice FEA though. I'm curious how much difference it would make if the glass wasn't flat but was domed. For eg, doming it inwards (towards the pressure) would put some of the force in compression. AIUI, glass has a compressive strength better than two orders of magnitude higher than its tensile strength.]

Doming it in would reduce the radius of curvature and decrease the loading. An example from an ILC Dover expandable habitat:

http://www.simulia.com/download/scc-papers/Aerospace/structural-design-analysis-testing-lunar-habitat-2009-F.pdf

I had a better link which modelled all the elements, including the habitat window. The habitat window meanwhile was flat (ostensibly for viewing) and the principal load was along the long edges.

Pretty much all heavy-duty inflatables have this kind of lobing to reduce stress.

http://adsabs.harvard.edu/full/2005ESASP.581E.101L

A geodesic dome would ideally have hexagonal/pentagonal panels which would be better domed. The principle stress is usually along the longest, straightest edge of a panel.

[guckyfan]

On another thread, I think in the Mars section, a while back there was a video showing a room at sea level pressure with the air inside at somewhat lower oxygen partial pressure. Comfortable for humans to breathe but at that oxygen level open fire seemed impossible. Matches or lighters would not ignite. A candle or torch brought in from the outside extinguished quickly. In an enclosed habitat where fire is extremely dangerous this would be a good condition.

Assuming that most pressurized volume is made by tunneling with only few exceptions on the surface, like a geodesic dome for psychologic reasons, having full pressure would not be a major problem. Maybe for safety reasons this would be the way to go. I am coming from the position that half SL pressure with increased partial oxygen is desireable but I have changed my position. Fire is something that should be avoided at almost any cost.

[lamontagne]

On another thread, I think in the Mars section, a while back there was a video showing a room at sea level pressure with the air inside at somewhat lower oxygen partial pressure. Comfortable for humans to breathe but at that oxygen level open fire seemed impossible. Matches or lighters would not ignite. A candle or torch brought in from the outside extinguished quickly. In an enclosed habitat where fire is extremely dangerous this would be a good condition.

Assuming that most pressurized volume is made by tunneling with only few exceptions on the surface, like a geodesic dome for psychologic reasons, having full pressure would not be a major problem. Maybe for safety reasons this would be the way to go. I am coming from the position that half SL pressure with increased partial oxygen is desireable but I have changed my position. Fire is something that should be avoided at almost any cost.

I just happen to be working on underground subway stations these days.  Fire risk and exit is one of the driving forces in the design.  Fire codes requires that the station can be evacuated within 6 minutes.  During that time the ventilation system must clear out the smoke to prevent the passengers from being blinded, asphyxiated and cooked.  A minor trash can fire can reach a few MW in power, and a large vehicle fire 15 MW or more.  A tanker truck in 100 to 200 MW of power.
The fire can also damage the concrete of the tunnel and create spalling, weakening it to the point of structural failure.
The classic approach to fire protection is the joined triangle: Oxygen, combustible, heat.  Increasing the oxygen content is a bad idea, from this point of view.  Reducing the use of organic materials is a good idea, to eliminate combustible.  Paint is a big problem in this regard.  Ceramic coverings might be favored for this reason. 
To reduce heat, the classic approach is to add sprinklers to construction.  A fairly recent advance it to use high pressure misters rather than sprinklers when the equipment is water sensitive.  The mist cools the air and displaces some oxygen, attacking the fire on two fronts.
This has been replacing CO2 protection, that is pretty much phased out everywhere now, as it does the same job, displace oxygen, without the risk of asphyxiation.  Plus the cooling.
I expect any permanent colony would be entirely protected by sprinklers, including all the corridors, the domes, if any, and all the technical spaces as well.
BTW some authorities, such as Australia, have decided to install sprinklers in all road tunnels.  It's a growing trend.
I would also expect a high level of partitioning, for smoke control and to create safe spaces for egress.
I wonder if breathing masks with small 10min breathing bottle might be available, and if these could serve in case of pressure loss as well.  On the other hand it would be difficult to have a few hundred of these available quickly in a large dome, a cafeteria full of people, for example, that suddenly lot a section of windows.

[Lampyridae]

I wonder if breathing masks with small 10min breathing bottle might be available, and if these could serve in case of pressure loss as well.  On the other hand it would be difficult to have a few hundred of these available quickly in a large dome, a cafeteria full of people, for example, that suddenly lot a section of windows.

However, a colony can voluntarily vent atmosphere to contain a fire. But I wouldn't go with increased O2 concentration at all, given the fire hazards in Martian g.

Multi-lobed fuel tanks are an actual thing, and have lighter skins than would otherwise be the case. However, the connections (nodes) weigh more because the load has to go somewhere.



And althought the noded composite tanks for the X-33 were a debacle (and an excuse to kill the program), the Al-Li tanks were just fine for the purpose.



http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.450.4264&rep=rep1&type=pdf

[Rocket Surgeon]

50 kPa may be doable with higher oxygen content (somthing like 8psi, 32% Oxygen...or 8/32),

That would drastically increase your fire risk.

As for atmosphere, I'll check my source but my understanding was 8/32 had the right balance of increase percentage and decreased pressure to maintain a similar flammability.

Looking at what I have at hand, something which is on the just self-extinguishing at 14.7psi & 21% Oxygen will self-extinguish at 12/23 and 7/28. Therefore 8/32 is well over the line.

That's at 1g. At Mars gravity, flammability is even higher.

Thanks! Do you mind if I ask what your source is for the above numbers? I'd like to read more into this and have a better knowledge of what prior work has been done.

[Paul451]

Do you mind if I ask what your source is for the above numbers? I'd like to read more into this and have a better knowledge of what prior work has been done.

I generally quote numbers from:- Hirsch, Williams, Beeson, "Pressure Effects on Oxygen Concentration Flammability Thresholds of Polymeric Materials for Aerospace Applications," Journal of Testing and Evaluation, Vol. 36, No. 1, 2008, pp. 69-72; because I've snipped out some of the tables/graphs.

The reduced gravity flammability comes from:- Olson & Ferkul, Evaluating Material Flammability in Microgravity and Martian Gravity Compared to the NASA Standard Normal Gravity Test. 42nd International Conference on Environmental Systems 2012, ICES 2012. Again, I've snipped some figures, so it's handy.

[Paul451]

However, a colony can voluntarily vent atmosphere to contain a fire.

That might be required just to deal with the thermal expansion anyway. Fire, hot. Using misters will also increase internal pressure, rule of thumb is 1600 times volume expansion. So your fire system will already have venting built in.

If the vented air (and misted water) smothers the fire, the remaining air will cool rapidly and the pressure will start to drop. Pump in stored nitrogen (cool from expansion) to keep pressure constant. Constant pressure lets the rescuers more easily enter the area (with portable oxygen), without much air-mixing at the hatches, reducing re-ignition/backdraft/flashover risks. Once the fire is definitely positively absolutely out and cold, you can use the same vents to flush the smoke while reintroducing breathable air. (Or just seal and vent to external pressure if you are writing it off or need to do a full rebuild.)

Even with a large fire, venting to Mars pressure might be unnecessary. Most fires might be dealt with by venting to a higher-than-Mars-pressure. Not sure how low you'd need to go to guarantee extinguishing, but it might be that the pressure can be kept high enough that any trapped humans can still breath (even if they are really, really unhappy) and self-evacuate.

But I wouldn't go with increased O2 concentration at all, given the fire hazards in Martian g.

Because of the issues of fire in a closed system, I've mentioned before that I'd rather spend the engineering costs and design to a higher pressure, with lower oxygen percentage. Reduce the flammability of materials below Earth normal.

But if venting (partial or fully) is effective enough, it might mean that fire isn't really such a great risk. (Lowering building pressure in response to a fire is not an option on Earth, so isn't part of our mind-set.)

[lamontagne]

However, a colony can voluntarily vent atmosphere to contain a fire.

That might be required just to deal with the thermal expansion anyway. Fire, hot. Using misters will also increase internal pressure, rule of thumb is 1600 times volume expansion. So your fire system will already have venting built in.

If the vented air (and misted water) smothers the fire, the remaining air will cool rapidly and the pressure will start to drop. Pump in stored nitrogen (cool from expansion) to keep pressure constant. Constant pressure lets the rescuers more easily enter the area (with portable oxygen), without much air-mixing at the hatches, reducing re-ignition/backdraft/flashover risks. Once the fire is definitely positively absolutely out and cold, you can use the same vents to flush the smoke while reintroducing breathable air. (Or just seal and vent to external pressure if you are writing it off or need to do a full rebuild.)

Even with a large fire, venting to Mars pressure might be unnecessary. Most fires might be dealt with by venting to a higher-than-Mars-pressure. Not sure how low you'd need to go to guarantee extinguishing, but it might be that the pressure can be kept high enough that any trapped humans can still breath (even if they are really, really unhappy) and self-evacuate.

But I wouldn't go with increased O2 concentration at all, given the fire hazards in Martian g.

Because of the issues of fire in a closed system, I've mentioned before that I'd rather spend the engineering costs and design to a higher pressure, with lower oxygen percentage. Reduce the flammability of materials below Earth normal.

But if venting (partial or fully) is effective enough, it might mean that fire isn't really such a great risk. (Lowering building pressure in response to a fire is not an option on Earth, so isn't part of our mind-set.)

You're not injecting a lot of mist though, so that's not so bad.  And water condenses out fast, so it's less of a problem than other gases.  The expanse in volume of the hot air from a fire is quite high, so you would want to stop it fast to prevent overpressure, that's a good point.  Doubling or tripling the pressure in a glass dome is not a good idea.  I agree that standard air pressure is probably, best.  There will be so many other things to deal with, let's try to keep at least one parameter constant in the experiment that a Mars colony would be.

Once you get everybody out, the simplest may indeed be to just vent the dome.  Save the people, then save the building.  Basic fire fighting.  The sprinklers are really there for the first few critical minutes, to ensure you don't get to the latter catastrophic situation.
Wonder what type of pressure/fire door would be best?  Hopefully, not a gigantic door, with big locking crenelations, that slams down and that your hero just barely manages to slip under.  And not the other door, where the sidekick gets caught on the wrong side and the hero watches helplessly as his best friend dies right in front of the tiny little window.

[Paul451]

Wonder what type of pressure/fire door would be best?  Hopefully, not a gigantic door, with big locking crenelations, that slams down and that your hero just barely manages to slip under.  And not the other door, where the sidekick gets caught on the wrong side and the hero watches helplessly as his best friend dies right in front of the tiny little window.

As long as the life support system counts oxygen levels in seconds-remaining, I'm happy.

[Paul451]

I'm curious how much difference it would make if the glass wasn't flat but was domed. For eg, doming it inwards (towards the pressure) would put some of the force in compression. AIUI, glass has a compressive strength better than two orders of magnitude higher than its tensile strength.

Just to be clear, by "dome inward", I didn't mean the entire structure is inverted, a la



I still mean a mundane outward curving shape, but the individual glass panes would curve inwards, to convert tensile forces on the glass into compressive force. (The frame is still tensile.) I'm curious if it would make glass more viable.

For eg,

[Lampyridae]

I'm curious how much difference it would make if the glass wasn't flat but was domed. For eg, doming it inwards (towards the pressure) would put some of the force in compression. AIUI, glass has a compressive strength better than two orders of magnitude higher than its tensile strength.

Just to be clear, by "dome inward", I didn't mean the entire structure is inverted, a la



I still mean a mundane outward curving shape, but the individual glass panes would curve inwards, to convert tensile forces on the glass into compressive force. (The frame is still tensile.) I'm curious if it would make glass more viable.

For eg,

Well, these would function as a set of (the tops of) vacuum bell jars. For 18" / 45cm you have up to 8mm glass thickness. Bell jars are designed for cylindical stresses, not spherical ones - the dome is therefore overdesigned.

Light bulbs are about .6mm in thickness, and they are crushed at depth as a low-energy seismic source.

http://saund.stanford.edu/saund1/Light_bulbs.pdf

Following the link we have a handy dandy selection of light bulbs, thicknesses, volume and their crush pressures.

Generally, the average (E27) 60mm diameter light bulb 0.6mm thick can take up to 1000kPa (yes 10 bar) of extra pressure before it implodes (neglecting internal pressure, some 80kPa). Since thickness is proportional to pressure and radius, we can easily scale these values up. Some implode at 200KPa so we'll keep this diameter-to-thickness ratio, and call it a safety factor of 2 as glass is a treacherous beast.

Using the above information, assuming the large size poses no manufacturing issues, we can confidently make a 6 metre diameter inverted dome section that is 6cm thick with a lightbulb-glass safety factor of 2. Or, a 60cm window 6mm thick that has a safety factor of 2.

https://ecatalog.corning.com/life-sciences/b2c/US/en/Glassware/Jars/PYREX%C2%AE-Bell-Jars-with-Top-Knob/p/6885-222

Taking a different method, a Pyrex bell jar from Corning's catalogue is 225mm in diameter and 4.5mm thick. It's meant to be pretty tough and withstand all sorts of experimental nasties. Make it 2.25m in diameter and the glass will be 45mm thick.

http://www.hyvac.com/Products/Systems/Bell_jar_specs.htm

This website lists an 18" (45mm) bell jar that's 8mm thick. So again, scaling up would give us something 40mm thick.

The fact that bell jars have a cylindrical component is what requires them to be thicker, as well as handle little explosions inside, rapid pressurisation etc.

Tempering improves the strength of glass by "freezing in" compressive forces that cancel tensile forces before the glass yields. This is energy intensive and time consuming.

[Rocket Surgeon]

Ask and ye shall receive!

So just quickly...I believe fixing the edges IS an accurate model, as the edges of the glass would be set into a frame, probably locally made steel, with a very tight fit to provide a seal. This would "stop" the glass from moving in a specific direction, or rotating at the edge, effectively fixing it.

For a more accurate simulation, I'd have to make the glass a Solid and "weld*" steel Plate to the edge of it. Doable, just a little more time consuming.

Again, I assuming yield strength is 7 MPa.

Results for a 500mm Equilateral:
- 50 kPa, 10mm, the highest stress is 11.47 MPa
- 50 kPa, 30mm, the highest stress is 1.252 MPa
- 50 kPa, 50mm, the highest stress is 0.438 MPa

- 25 kPa, 10mm, the highest stress is 5.736 MPa
- 25 kPa, 30mm, the highest stress is 0.626 MPa
- 25 kPa, 50mm, the highest stress is 0.219 MPa

As a comparison, the 1000mm Equilateral:
- 50 kPa, 10mm, the highest stress is 49.25 MPa
- 50 kPa, 30mm, the highest stress is 5.448 MPa
- 50 kPa, 50mm, the highest stress is 1.943 MPa

- 25 kPa, 10mm, the highest stress is 24.62 MPa
- 25 kPa, 30mm, the highest stress is 2.724 MPa
- 25 kPa, 50mm, the highest stress is 0.971 MPa

So the shorter, the better, and the thicker, the better. In fact, even at 50 kPa, a 30mm thick panel is more than sufficient structurally, depending on your safety factor.  Such a panel would only be 9 kg, excluding the framework.


*This is just a FEMAP use term, not actually welding in real life.

[Paul451]

I believe fixing the edges IS an accurate model, as the edges of the glass would be set into a frame, probably locally made steel, with a very tight fit to provide a seal. This would "stop" the glass from moving in a specific direction, or rotating at the edge, effectively fixing it.

Since it's under outwards pressure, the alternative is to let it self-seal against the inner face of the frame. (A few light clips to hold it in place when the dome is unpressurised.) That way it can expand and shrink at a different rate than the frame (within reason) and the seal is tight regardless.

Eg... (flat panes because I forgot about domed panes until I'd finished)

[Lampyridae]

I believe fixing the edges IS an accurate model, as the edges of the glass would be set into a frame, probably locally made steel, with a very tight fit to provide a seal. This would "stop" the glass from moving in a specific direction, or rotating at the edge, effectively fixing it.

Since it's under outwards pressure, the alternative is to let it self-seal against the inner face of the frame. (A few light clips to hold it in place when the dome is unpressurised.) That way it can expand and shrink at a different rate than the frame (within reason) and the seal is tight regardless.

Eg... (flat panes because I forgot about domed panes until I'd finished)

An example from real life, the LEM windows were mounted on sprung metal inside a teflon seal and were thus "floating," taking only the force of cabin pressure.

Aircraft cabin windows are designed as cutouts, so the stress can go around them. The principle is the Kirsch hole-in-an-infinite-plate. Stress distribution is independent of the size of the hole, always triple right at the edge and then dropping off one diameter away. /notanengineer



SpaceShip Two has windows spaced somewhat over one diameter apart, and Burt Rutan gave a sharp welding hammer to the SpaceShip One pilots and told them to try and break the polycarbonate windows during a pressure test.



New Shepard windows are huge, and seem to keep this one diameter spacing.



SpaceShip One, Two and New Shepard all have curvature in their windows for streamlining but may also help distribute stress.

I don't know at what point the assumption becomes that the windows can no longer be treated as simple structural cutouts.

Article on Apollo program windows: https://www.lpi.usra.edu/lunar/documents/apolloSpacecraftWindows.pdf

[Paul451]

Colour version of that Apollo LM window.



[edit: Just to be clear, found it on the interwebz, I didn't colour it.]

[Rocket Surgeon]

Hmmm... those configuration would still constrain the glass in T(x,y,z) but would allow rotation, possibly reducing stress... AND still allow the model to solve!

[Oersted]

In "envisioning amazing habitats" we already went through the dome discussion around 60 pages and three years back. Yup, glass domes are actually no good on Mars.

For those who are new to this forum and to discussing bases on Mars that thread is a good long read.

It will also explain to you why so many oldtimers in here seem to fancy tunnels (as does Musk).

That wasn’t actually a consensus opinion.

I agree and I didn't say it was.

[rarchimedes]

I'm just catching up on all this after a week in the hospital from foot surgery, so let me take a few shots. I agree with the initial comments that I have seen on domed windows, but the one thing that I do not understand is the references to glass thicknesses as directly proportionate to radius when every other thing in this world that I know is proportional to the square of the change in radius.

Moving on, truly hemispherical segments make the most sense because they transfer all lift forces perpendicular to the base plane and therefore to the frames. The hemisphere is the strongest possible form to present to both forces from within and dangers from without. If you were flying these, the weight of hemispheres as opposed to almost any other glass form might be prohibitive.

I renew my objection to any "window" exposed directly to a line of sight leading to the atmosphere on Mars. That atmosphere will hardly slow or filter or burn up incoming rocks or radiation of any kind, leaving a necessity for complex support and emergency repair systems as well as thicker and more complexly produced glass, itself. Mirrors can be used to bring light and images (properly erected) to properly protected windows, but I have seen no evidence that lack of external lines of sight is the kind of psychological problem that has been evinced in theses parts. Some of the sunniest states in our union have some of the highest suicide rates. I might not find it ideal, but I might prefer it to exposure to unmediated Mars. Explosive decompression is also not much fun and can be difficult to remediate. Referring to the ISS or to the LEM isn't of much help as both of those are horrendously expensive and not easily translatable to hectare size exposures. Plastics can easily be formed into hemispheres of any particular thickness with actually better optical clarity than glass, especially at the greater thicknesses. The hydrogen content of most plastics make them better radiation filters than most glass and at much lower densities.

As noted elsewhere, the process to produce rocket fuel can easily be tied to a process to produce plastics and without heavy duty mining, crushing, separating and possible extended chemical processes necessary to make glass of any useful quality. The only mining operation necessary early on, on Mars would be for the production of water which would then feed almost all other processes.

Elsewhere in these posts there have been discussions ad nauseum about production of various metals and silicon, none of which seem likely to be done with the equipment available to the early cycles to Mars. Pure heat separation may possibly produce small amounts of the lower temp metals, but the closest nearly pure metals are allegedly to be found in metallic asteroids. If that is so, bringing metals from asteroids might be far cheaper than from Earth or through production on Mars. Unlike on Earth, where an elevator to geosynchronous orbit is just marginally physically possible, putting in an elevator to areosynchronous orbit has no technical issues, it will be fairly easy to establish a Mars elevator to bring down such riches to the Mars surface while making the safety of access to and egress from the surface magnitudes of order safer and easier than at present. That is how we get our windows and almost everything else.

[lamontagne]

If small glass panes are advantageous as far as mass goes, since they are under less strain and can be thinner, then we might want to forget about float glass and perhaps go with another forming process?  Perhaps even molds, rather than a continuous process?  Not that is matters much if we have waste glass, as it seems typical to re-melt part of the production.  I think it would be nice to have rounded corners in the glass panes, as per the LEM.

Or perhaps go with a very narrow production line, perhaps 200mm wide?  Would certainly save on mass and complexity.

Might be interesting, if the parts are small enough, to do some of the treatment for toughness that Corning uses for it's glass?  Potassium should be available, or can come from Earth if need be.  Anyway it'll be required for other things so might as well produce it locally.

https://www.corning.com/au/en/innovation/the-glass-age/science-of-glass/the-secret-of-tough-glass-ion-exchange.html

We only need a tiny production line, after all, a few tonnes per day will already be quite a success.

A small industrial installation here on Earth seems to be 400 tonnes per day.