Kilometer Structures in Space

Zac Manchester is working on a phase 2 NASA NIAC (advanced innovation) study to create kilometer sized structures in space that will fit in one rocket payload.

Zac Manchester worked on Breakthrough Starshot which is trying to get a laser pushed solar sail to another star system. (aka 4+ light years in a human lifetime). Here is a video where Zac talks about breakthrough Starshot.

The Kilometer Structures in Space Using Mechanical Metametarials

Concepts for rotating space habitats as a means for generating artificial gravity date back more than a century. However, humans suffer discomfort from exposure to rotation rates as low as 3 RPM. To produce artificial gravity near 1g at rotation rates of 1-2 RPM, a kilometer-scale structure is needed.

The core of Zac’s solution is a high-expansion-ratio deployable structure (HERDS) built from mechanical metamaterials.

Specifically, they exploit two kinematic discoveries made in the last 5 years: shearing auxetics and branched scissor mechanisms. They intend to produce tube structures with an unprecedented 150x expansion ratio.

The Phase I NIAC study has demonstrated the viability of this approach and pointed us to several technical problems that must be addressed in Phase II.

The key technical work in Phase II will be focused on four specific thrusts:
1) modeling and understanding the complex deployment dynamics of our expanding hierarchical structure in detail;
2) mitigating jamming during deployment in the presence of manufacturing errors and external disturbances using simulation and design optimization;
3) rapid prototyping and hardware-based design iteration to calibrate models and evaluate sub-system components; and
4) experimental validation of meter-scale prototypes with thousands of links to demonstrate deployment without jamming and high expansion ratios.

In the near term, such structures would make sustained human habitation in cislunar space.

In the medium-to-longer term, such structures will be critical to sustaining humans in deep space. Finally, large structures will also advance astronomy by supporting large-scale telescope arrays.

29 thoughts on “Kilometer Structures in Space”

  1. A square kilometer array of thin-film solar cells could produce 350 MW of power for space manufacturing or to power the propulsion system of a Mars cycler or beam down 280 MW of power assuming an 80% efficiency in beaming power as microwaves and converting in power for the grid with rectennas.

  2. So why do you need a rigid structure between the ends? Strap two Starships together with a metal wire and you are done. Fixed it for ya.

  3. The cheapest way to fix this must be to connect 2 or 3 starships with tethers and just spin them. Perhaps SpaceX can develop a dedicated vacuum-only hub-ship that handles the tethers. Heatshields and engines can be swapped for something more useful. Perhaps solar panels and ion-drives.

    If the constellation of two ships or 2+hub-ship are connected on their sides, they can travel towards Mars or wherever with the rear facing the sun for maximum radiation shielding. Then all the fuel tanks and rocket hardware will act as shielding.
    I don’t know how hard it will be to adapt the interior of the starships for horizontal orientation from the current vertical.

    Perhaps, it would be much easier to have dedicated hotel-ships that never land in deep gravity wells. Just move passengers over to the shuttle/lander starship with heatshields and many engines before landing. The hotel-ships desperately need shielding from radiation and micrometeoroids. They also need a lot of life support equipment. Some of that mass can be offset by dropping heatshields, engines and fuel tanks not needed for landing.
    Another little problem is how to land a 70 m tall starship on an unprepared surface on Mars or the Moon. I haven’t seen a reassuring explanation how that is to be done.
    A horizontal design would solve many problems.

    More efficiency, safety, comfort and mass to target by using specialized versions of Starship…

    • Regarding landing in an unprepared surface, obviously the current Starships, which lack landing legs, cannot do it. However, once a set of legs is added, it’s not difficult to find a sufficiently-flat region on the Moon or Mars where it can land, even standing up. Also, even though landing a ship on its side is much easier and more stable, taking off again from that position is much, much harder, so obviously Starship cannot do that. So clearly SpaceX will continue to have it land on its tail.

      • There are a few things that speaks for having a different landing architecture for Earth, moon and Mars. Lack of atmosphere is one and that makes the grid fins useless or much less useful. The moon vacuum makes it possible to drop the heat shielding.

        Less gravity is another and this makes everything easier. One can drop many engines for example.

        The problem with unprepared surfaces is that the ground will/may not carry the weight and the rocket will topple over at some point. Legs will probably have to be much wider than what we have seen so far.

        There is also the problem with debris that will fly around at high velocities. It will be a danger to the ship and all ground equipment anywhere near the landing site both before landing and after take-off. The only way to fix that is to have the exhaust higher up in the stack and angled outwards. Not even that may be enough.

        Starship is 70 m high. Logistics with moving people and cargo in and out of the ship will be challenging, slow and perhaps risky.

        A horizontal architecture solves some problems and probably makes logistics much easier. One sci-fi system that could work for the moon is the old Eagle from the TV show Space 1999. Google “Space 1999 Eagle” and look at images.
        On the moon, one has the luxury of not having to have to deal with aerodynamics so a lander has more design freedom.

        Mars is very difficult and the atmosphere is not very helpful. You need heat shields and aerodynamic design but wont get much help slowing down once subsonic.

        Specialized landers with higher efficiency are weighted against mass production logistics back on Earth.

        Maybe the solution is to have a specialized lander that has the only role of paving and preparing landing pads for incoming traffic.

      • My own proposal was to equip the Starship, in orbit, with a special landing system that would mount to the rocket’s hold downs. It would include wide landing legs and auxiliary fuel tanks, but rely on the Starship for propulsion.

        Since the Starship has a fairly high thrust to weight ratio even fully loaded, landing on the Moon this landing system could outweigh the Starship itself several times over, carrying enough fuel that the Starship would land on the Moon with its own tanks still full, enabling the return to Earth with a full payload.

        The landing system would be left behind, to be converted to living space.

  4. The problem with these high expansion ratio structures is that they’re inherently flimsy, because there’s not a lot of actual structure there. It’s just spread out over a large volume.

    So you expand your package into a space station sized wisp that can’t hold air, or support people walking around in it.

    Now, if the goal is to use it as a template to deposit more material on, perhaps in some large scale application of sputtering, fine. It might work for that, and might ease construction, though not really reduce material requirements. Or, hey, maybe a microwave mirror, that doesn’t mind being mostly holes so long as they’re small, and isn’t under much load.

    But if your goal is to directly produce a large volume rotating to provide gravity, that will hold air, (Or even bowling balls!) this does not strike me as a plausible approach.

    The simple fact is that minimum structural requirements, size, and strength of materials, set lower limits on mass that there are simply no fancy ways to evade.

    If you want a quick variable G space station to test biological effects, forget this: Two Starships, connected by a tether. You can even have them at different accelerations by varying the residual fuel load.

    • I guess my assumption wasn’t that the 1km structure was to be enclosed ala O’Neil scale habitats but rather a more rigid alternative to a cable tether with solid habitats at the remote ends. This could scale to a classic torus with these expanded structures used for the spokes, which would speed construction, especially if their rigidity permitted use of partially completed torus structures.

      • Yah… toruses and donuts are attractive, but it is more cultural. Freed from the thousands of preexisting pictures of tori, I do wonder what elegant structures might be conjured forth.

        • Well let’s see:
          We can look at the basic physics. IF we want a constant 1g, then we’ve got all the living spaces at a constant radius from the axis of rotation. That is either a cylinder, or a subset of one.

          Assumes one rigid structure rotating at constant rate. If we have multiple rotational speeds and/or multiple axes of rotation then we can do all sorts of things, but that makes it really difficult to move from one part of the structure to another. At that point we might as well analyse it as multiple different structures, each of which is still a subset of a cylinder.

          If we have a subset of a cylinder, or a cylinder with a very high ratio of radius to axis length ie. a torus, then I don’t see what imagination you can apply except in choosing what subset you carve out from the maximum theoretical starting space of a full cylindrical surface. You can go wild in the carve out, constrained only in that you want your final structure to be mass balanced symmetrically around the axis of rotation. And you can play with that by using non-living space counter balances. A 1000 t solid rock mass at 500m from the axis will balance a 250 t living space at 2km radius on the opposite side. (Stick the solar flare radiation shelters in the 1000t block.)

          I suppose we don’t REALLY need 1g constant either. So we could have a spiral or something, starting at any value down to free fall at the axis and spiralling out to a 1.5 g storage/radiation shielding and gym deck on the outside. Then that can be carved up like a paper cut-out to give more or less surface area at each level.

          • One thing – if the counter-weight isn’t additional cargo or living space, you’d probably want to minimize it. So the center of mass/rotation of the system might be much closer to the living space, as opposed to closer to the counter-weight. At some point the mass of tether starts to become a factor too, especially since you may want more than one tether for safety and to minimize tilting due to minor internal mass shifts in the living space (i.e. someone walking around).

  5. How would we avoid these massive structures falling back to earth and making a bit of a mess? Seems to happen a lot with satellites

  6. Hi Brian
    I like the concept, but the justification is suspect. A small diameter rotating vehicle habitat needs to be tested for a range of spin-rates so the *actual* human response can be demonstrated one way or the other. There’s so much conflicting opinion on this question.

    • More or less. There is controversy whether 1.6 or 3.7 m/s² will yield long-term health in humans. 9.8 m/s² is *known* to work, so setting up an habitat for 1G may be more cost-effective than the thorough testing needed to fully characterise the biologic response to low gravity.

      Of course, if we can get it we probably should do this characterisation *as well,* but it may be cheaper to go straight for a full standard gravity from the word “go.”

      • The structural requirements do scale pretty heavily with the level of acceleration you supply, so the lower you can get away with, the better. Although the mass savings don’t kick in at small scale, due to the necessity for radiation shielding if the habitat is outside the Van Allen belts. For smaller habitats, for this reason, yeah, might as well just go to 1 G.

        O’Neill’s largest colonies were scaled so that the structural mass alone would provide sufficient radiation shielding. Even so, at lower gravity levels they could have just been larger. So at large scale it makes a difference.

        The real question for us here is colonization of existing bodies, such as the Moon or Mars. If 1/6G is inadequate, to build colonies on the Moon you’d have to build in centrifuges. Aside from the need for them to be conical rather than cylindrical, perfectly feasible.

        On Mars, if 1/3G isn’t enough, it gets dicey, because there’s enough atmosphere to make really large rotating structures hard to pull off.

        Ironically, if humans require 1G for health, the case for colonizing the asteroids instead of Mars gets stronger, just because it’s easier to build rotating habitats in zero G without an atmosphere present.

        This is actually a question that needs to be definitively settled fairly early in Musk’s Mars colonization plan. And, as I say above, it’s fairly easy once he gets the Starship working. He can just take two of them outfitted as though for the trip to Mars, only with more medical gear, and connect them with a tether bolo style, and conduct the research in low Earth orbit, while doing long duration tests on all the systems.

        • I don’t think I would want to volunteer for that blobs-on-a-bolo test bed. There is just SO much shît that would prove rather hard to compensate for, should the bolo break. The blobs would have to be rescued. And let us say what, 500 m, or half a kilometer diameter?

          a = ω² • radius

          ω is radians per second. radius in meters. acceleration in m/s²
          Léts see … we want ‘a’ to be say 5 m/s² or so.

          5 = ω² • (½ 500)
          ω = 0.1414 rad/sec = 1.35 RPM

          OK, it is not so bad. A little more math shows 225 meters per second tangential velocity for the bolo-tied blobs.

          I worry too much.

          • That’s the point of using two fully equipped Starships: If your tether breaks, you’ve got… Two fully equipped Starships at zero G in very slightly different orbits. You’ve got very little to worry about except the possibility of accidentally ending up on a collision course with something. I assume the autopilot is capable of keeping track of that, and doing a quick burn to negate it in the event.

            They can rescue themselves. You just have to make sure that nothing disasterous is going to happen if the ‘gravity’ suddenly goes away.

            And, how is the cable going to break? It’s not going to be one strand, after all, or running near the breaking point. There’s a multi-strand tether design that actually gets stronger if a single strand breaks, the “Hoytether” (Due to the remaining strands realigning to better carry the load.) and that would certainly be used. Basically you’d need a foreign object almost as large as the Starships themselves to nail that tether, assuming the anchor rings were full diameter.


            I actually expect manned missions to Mars to be conducted in exactly this way, because it would allow you to spend the entire trip acclimating to Martian gravity. And if returning, reacclimating to Earth gravity.

            • Maybe wear at the tether connection point due to movements within the life support hab shifting it around could weaken the tether? Not likely to be a fast process, and you could pretty easily design multiple attachment points – so if it did break, it’d only let loose a bit before catching. So at worst your hab jerks and tilts a bit as one tether effectively lengthens a bit, – making it obvious that you need to go make repairs.

          • That’s the point of using two fully equipped Starships: If your tether breaks, you’ve got… Two fully equipped Starships at zero G in very slightly different orbits.

            I actually expect manned missions to Mars to be conducted in exactly this way, because it would allow you to spend the entire trip acclimating to Martian gravity. And if returning, reacclimating to Earth gravity.

        • I agree with everything that you said above, but the point of this research is the very early bootstrapping tech. By the time we can do anything even approaching O’Neill scale, we can probably do 3 standard gravities if we liked (not that there would be any point to it). If we’re going to wait until then we might as well do the partial-gravity research.

        • Another set of unanswered questions: Do we need a constant 1g to maintain health? Maybe 8 hours while sleeping is OK? 8 hours but we have to be up and walking around? What if it turns out we just need three half-hour exercise sessions at 1g? Would we be willing to tolerate some of the effects of low gravity on long space trips and on the Moon and Mars, if the heart and other organs could be kept healthy by reasonable amounts of exercise at 1g?

          NASA’s human space program would have been looking for the answers to these sorts of questions for the past couple decades, if they were serious about getting to Mars or establishing a permanent presence on the moon. Of course, those answers are not necessary if they’re just using those ideas to maintain NASA funding.

Comments are closed.