With computer flight controls where they are today, what would stop us from building a relatively large space station on earth, but then using 10 or 20 or even 100 falcon heavy or Starship or SLS rockets to launch it to orbit?

For an example, let’s say we built a von Braun station with a diameter of 300M that had rockets placed under it evenly every ~47 meters to launch it in one go into orbit? This assumes an extremely strong structure, but in theory, if the launches and flight paths are coordinated well they could avoid any twisting or other non-consistent pressures? Even an empty structure with nothing but the pressure vessel and basic airlock equipment would solve a lot of initial problems. I am assuming lower payload per rocket because the aerodynamics would obviously be affected at lower altitudes.

Is this remotely possible as an alternative to current space construction issues and limited fairing sizes, or are there issues I’m not considering or underestimating?

  • 2
    $\begingroup$ Yes, a massively parallel rocket question needs to be asked! As long as there is a gap between rockets for air to flow, drag per rocket may be not much larger than for a single rocket. Different but related: What would the challenges be in developing a Falcon Heavy with three or four strap-on boosters? and Ballpark comparison of a hypothetical Falcon 'Quad' Heavy with cross feeds and How big could a rocket get? $\endgroup$
    – uhoh
    Commented Jan 15, 2021 at 8:11
  • 3
    $\begingroup$ Purely from a risk standpoint, not a good idea. It's one thing to lose 5% of a space station if 19 out of 20 launches (containing parts) are successful; quite another to lose the entire investment in one failure. And as you suggest, why make the entire structure launch-force-compatible when that's gross overkill for the final deployment? $\endgroup$ Commented Jan 15, 2021 at 12:56
  • $\begingroup$ As noted above, the risks are too high. Also, the aerodynamic effects will be sufficient to significantly increase the fuel costs for launching such a design. This, in turn, would necessitate an increase in the amount of fuel. Further, this leads to an increase in total mass. If it were expedient, then developments in this direction would already be carried out. The James Webb Telescope will be deployed in space. That is, it will be launched in a folded state. There are objective reasons for this. $\endgroup$
    – TommyJo
    Commented Jan 15, 2021 at 14:54
  • $\begingroup$ This resembles OTRAG, a launch system comprised of lots of identical rockets bundled together (and the bundles stacked, for multiple stages). It was a very interesting project that went nowhere for reasons only tangentially related to the design itself. But it seems like an OTRAG-like system could get this sort of work done. $\endgroup$
    – SF.
    Commented Jan 15, 2021 at 17:56
  • 1
    $\begingroup$ Don't forget that there's another reason for assembling structures in orbit from smaller parts: A large space station that was just strong enough to withstand the worst-case stresses that it would encounter in flight might be incapable of supporting its own weight on the surface of the Earth. A sufficiently large spacecraft that is strong enough to not collapse under its own weight here on Earth might be massively over-built for orbit. $\endgroup$ Commented Jan 15, 2021 at 18:26

3 Answers 3


This is not necessarily as far fetched as you may think, but you seem to focus on lifting a structure and building a rocket system around it.

For instance: in 1968, Boeing studied a post-Saturn vehicle using four Saturn-V cores, with first and second stages in parallel arranged below a massive payload fairing.

Saturn V-4X(U)
Saturn V-4X(U) (Hazegrayart)

This would allow for million pound launches to LEO, and thus is an example of a superheavy vehicle built from multiple other launchers all lifting in parallel.

In terms of your original question, a ring station or similar is extremely unaerodynamic due to the physics of a cylinder being rather disgusting.

It would make more sense to forego the fully constructed station and instead use a couple launches of a more optimised vehicle to lift and construct parts of the station.

  • 1
    $\begingroup$ True, I was focused on the details on the original question, in particular launching an intact rotating wheel space station, which, by its distributed structure, is very demanding aerodynamically. structurally and in terms of flight control. I don't see huge issues with clustered rockets for super-heavy lift. $\endgroup$
    – Galerita
    Commented Apr 29 at 0:10
  • $\begingroup$ Perhaps just semantics and it doesn't change your point, but it wasn't actually designed it was just part of a study. Along with a lot of other ideas such as strapping four 13-ft diameter solid boosters onto a Saturn IC. The artwork shown is a screenshot from someone who does a lot of CGI videos. $\endgroup$ Commented Apr 29 at 13:09
  • 1
    $\begingroup$ @StevePemberton Yes I'm aware: I used the wrong terminology, and I'll edit as such. Best depiction I could find as Astronautix nor NTRS has a good depiction that I could find. $\endgroup$ Commented Apr 29 at 23:20
  • $\begingroup$ AnarchoEngineer - I saw the Astronautix article also, interesting that it wouldn't have been just four standard Saturn IC cores but they each would have been stretched. Would be nice to find the actual Boeing study but I don't know if it's available anywhere. Closest that I found was a Marshall study that lists lots of ideas, mainly strap on solids. $\endgroup$ Commented Apr 30 at 13:19
  • $\begingroup$ @StevePemberton Yeah I can't seem to find it anywhere unfortunately. However I do like that it wouldn't require new tooling and machinery for most of the manufacturing process right up until full stacking and connection of the payload service module to the cores. I am rather fond of it I must admit $\endgroup$ Commented Apr 30 at 20:59

It may be possible, but it would be both inefficient and unwise (risky). I agree with many of the comments made below the question. Fundamentally there two competing design criteria: 1) a von Braun Station (vBS) orbiting at 400 km using centripetal force to simulate gravity; and 2) a payload that must withstand one - and only one - launch, and the aerodynamic and thrust forces normal to these centripetal forces.

Sizing the Problem

The 300 m diameter vBS will be too massive to be launched using the 20 Starship/SuperHeavy proposed.

Consider SpaceTech's vBS. It's 190-meter in diameter and composed of 24 cylindrical modules, each 20 meters long 12 meters in diameter. It has a corridor (a torus) and shafts descending between modules. There are 4 main spokes giving access to a central docking point.


We can scale an ISS module to estimate the mass of each vBS module. The Destiny Module is 8.4 m long, 4.2 m in diameter and has a mass of 14,500 kg. The surface area of a cylinder is $2\pi r(r + l)$. Destiny has a surface area of ~138 m^2; the vBS module ~980 m^2, or 7.1 times the area. For a given pressure, pressure vessels normally have wall thicknesses proportional to the radius of the vessel. If we neglect this and assume a linear scaling 7.1x14,500 ~ 103,000 kg.

They assume 40 launches of the SpaceX Starship to "send all the pieces up there". If we assume the additional features - the corridor torus, spokes, and shafts - don't contribute too much more to the mass (unlikely), 24 simultaneous Starship launches could launch this vBS to LEO. (Note the structure will necessarily be much heavier if strengthened to withstand launch as an intact structure.)

Atmospheric Drag

Each Starship has a cross-sectional area of $\pi r^2 = 4.5^2 \pi = 64 m^2 $. I'll assume Starship has similar aerodynamic properties to the Falcon 9, which someone has estimated has a drag coefficient (Cd) of about 0.75. Maximum dynamic pressure ($P_m$) is around 35 kPa for most rockets. The drag force is $D = P_m C_d A = 35*0.75*64 = 1,680 kN$.

The SuperHeavy has a thrust of 74,400 kN. Assuming an all-up weight of 5,000 tonnes, 25,350 kN is used to accelerate the vehicle and the remainder is used to hold the rocket in the air (assuming it is going vertically, which is roughly true at Qmax). Only ~ 6.6 % of the "excess thrust" is used to push through Qmax.

Approximating the vBS as 24 cylinders, each cylinder has a cross-sectional area of 20*12 = 240 m^2. If we continue the calculation on a per module/rocket basis, Starship would likely be anchored in the middle of the connecting shaft serving two modules & hence not well obscured by the slipstream of the vBS. In addition part of the torus corridor and spokes should be added to the per-booster cross sectional area. A minimum cross sectional area would be 300 m^2 per module, and likely more.

Someone has kindly estimated the Cd for Starship in its belly-flop entry mode, with Cd ~ 2 in this configuration near Mach 1, which is also close to Qmax.

Cd Starship belly-flop

Redoing the calculation (on a per module basis), $D = P_m C_d A = 35*2*300 = 21,000 kN$. This likely underestimates the drag, but it is very similar to the excess thrust of 25,350 kN for each SuperHeavy.

The vast configuration would likely have to delay passage through Qmax until a greater altitude and lower air density. It would also be unlikely to reach it's desired orbital height, if it could reach orbit at all.

In contrast, when launched individually, each module - in an appropriate fairing - would easily be lofted to a 400 km LEO.

Further Considerations

I agree with many of the comments above. In particular I think it highly unlikely that current flight control software could keep 24 SuperHeavy's sufficiently coordinated to prevent breakup of the vBS. In any case, a much heavier support structure would be required to withstand the point loads of the booster and differential loads through the structure due to inevitable minor coordination errors. Vibration of the entire structure would also be an issue - potentially setting up destructive harmonics. Beneath a fairing a single module would be easily dampened against vibration. Finally, the modules and structure would likely need to be strengthened against Qmax.

Launching even a small vBS in this manner would be highly inefficient compared to assembling it in orbit. Such a launch would not be attempted unless there was absolutely no alternative.

p.s. I noticed this lonely question had been sitting by itself for some time, so I thought I'd give it a go. Have pity!

  • 1
    $\begingroup$ your flight control software point makes me wonder about a custom controller also attempting to minimize structural stress as measured on the payload. If asked to justify it I might be able to get away with something about space tugs $\endgroup$
    – Erin Anne
    Commented Apr 28 at 19:32

After reading through all of the comments, I'm inclined to agree with you that in principle this may be a feasible idea. The advancements in flight controls have been considerable. We can now reliably land a booster and the possibility of catching a booster or a ship seems like a reasonable next milestone. The possibility of launching multiple rockets in a tight formation seems reasonable, with our present level of technology.

For some launch systems, reliability is high enough to launch humans to space. Because we have good data on the probability of a single rocket launch failing, it is possible to calculate the probability of success for a mission involving multiple rockets that are launching simultaneously.

The question then becomes, is it possible for multiple rockets to launch in a tight formation and share responsibility for lifting a large payload? And, if that is possible, how does the addition of the payload-sharing requirement affect the overall probability of success for the mission?

Let's consider an architecture in which N rockets are placed on N separate launch platforms. The N platforms are pre-arranged in a tight formation in the center of the large ring-shaped space station that we want to launch into orbit. The rockets are all then individually connected to the station with lots of guy wires made out of a strong light-weigh material such as carbon fiber. (See figure below illustrating the concept for just one rocket connected using just a few guy wires).

Rocket connected to station by guy wires

Theoretically it should be possible to launch all the rockets simultaneously and have them together carry the space space station to orbit.

The space station and numerous guy wires will generate more aerodynamic drag than the individual rockets would normally experience, so the amount of payload each rocket can launch to orbit will be lower than normal, but this is not a showstopper. Only a small fraction of a rocket's delta-v is typically used to overcome aerodynamic drag.

With lots of rockets, the probability of scrubs occurring will be higher than normal too, but again this does not appear to be a showstopper.

It would obviously help if the space station were architected up front to be amenable this style of launch. The plan could involve over pressurizing the habitation modules to give them more structural rigidity to resist the aerodynamic pressure experienced during the launch. Or, inflatable habitat modules could be used to reduce the aerodynamic cross section during launch. We could also give the station some spin on the way up to help it resist hoop stresses and buckling forces as the acceleration increased.

So, "Is this remotely possible as an alternative to current space construction issues and limited fairing sizes..?"

Yes, absolutely. It would also be a spectacular achievement that would no doubt go down in the history books.

The main challenge to moving forward with such an idea is simply that an alternative approach, such as building the station as separate modules, launching them individually, and later assembling them in space, could have certain advantages.

  • $\begingroup$ I didn't think of that setup. One giant cluster of connected SuperHeavys would solve the control issues. The aerodynamics wouldn't change and the structure would still need to the massively reinforced. There are new structural issues in terms of the guy wires. The reinforcement of the ring station would have to be considerable to withstand being lifted at an oblique angle. It reminds me of a giant helicopter, except not gently lifting a large irregular structure, but literally rocketing it into space. I still think it's nuts unless absolutely necessary. $\endgroup$
    – Galerita
    Commented Apr 29 at 0:31
  • $\begingroup$ As well as being bundle, the SuperHeavys would need some form of robust capping fairing, to which the guy wires are attached. If the guy wires were attached to the top of each Starship, The forces would pull the bundle apart and may even "bend" the individual rockets. $\endgroup$
    – Galerita
    Commented Apr 29 at 3:00
  • 1
    $\begingroup$ There are no "bundles", Starships, or SuperHeavys in the architecture proposed above. Just proven rockets such as Falcon 9 or SLS. The guy wires would attach to the second stage's engine mount and to points at the top of the second stage where the rocket's payload would normally be placed. The forces from the guy wires will be similar to the forces from the rockets' regular the payload. $\endgroup$
    – phil1008
    Commented Apr 29 at 3:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.