6
$\begingroup$

Rockets are basically horizontally or vertically assembled and integrated. SpaceX selected horizontal integration for its current Falcon rockets, essentially for reasons of lower cost. This orientation was significantly different from the choices made by the vast majority of the other rocket operators.

All evidences show that the new vehicle, Starship is vertically assembled with all sorts of cranes, and SpaceX renderings show vertical integration of Starship and the booster. So for what reasons did SpaceX choose vertical integration instead of -- allegedly simpler and cheaper -- horizontal integration for Starship?

$\endgroup$

3 Answers 3

6
$\begingroup$

There are two separate operations mentioned in the OP: Manufacture of the stages that make up the stack, and Integration of the stack prior to launch.

Integration

The integration of a 2 stage rocket consists of 2 separate operations: integration of the payload to the second stage, and integration of the second stage to the first stage.

Integration of the first stage to the second stage is traditionally done inside a building. For smaller rockets such as the Falcon 9 and Soyuz, it makes sense to do this inside a standard building. These tend to be too low to stand a rocket up in. You then have the issue of standing the rocket up afterwards. For very large rockets such as the Saturn V and the Space Shuttle integration of the first and second stage is done vertically. This requires an extremely large specialized building and access equipment, but avoids the complications and risks of erecting the complete stack once it is integrated. The larger the rocket is, the more effort will have been made to save weight on the structural design, and the more challenging horizontal integration and subsequent erection will be.

In the case of Starship, the booster is designed to take off vertically, land vertically, and be ready to reuse quickly. The booster needs no heat shield as it only operates at suborbital velocity and most of the parts requiring servicing (in particular the engines) are at the bottom. The booster should never need to be laid down horizontally during normal operations. Hence vertical integration of the upper stage to the booster is the only option. And we know that it will be done outside, not in a building.

To my knowledge we have no information on how the payload will be integrated into the upper stage. It may well be done horizontally for some payloads. The upper stage will require a greater range of maintenance tasks than the booster, and I think it likely that it will need to be laid down horizontally at some point for work on the upper fins and heatshield. After that it would need to be stood up vertically. Here's a video of the space shuttle being stood up vertically and mounted to the external tank. Payloads could be integrated with the orbiter either horizontally or vertically, see section 7.3.3 of the shuttle user guide.

We know from the structural failure of Starship SN3 that SpaceX have cut the structural strength to an absolute minimum, so that for some operations the tanks need to be pressurized in order to withstand the structural load even in the vertical orientation (to say nothing of the horizontal orientation or the transition between the two). So the structural loads for any maintenance or integration operations done in the horizontal will need to be considered carefully. SpaceX are not going to add weight to the Starship for ground maintenance, so any additional bracing will have be be added as external jigs.

Some payloads require vertical integration, and SpaceX wishes to carry these, so vertical integration is definitely going to be available. Payloads requiring vertical integration are often those which have not been designed to take the force of gravity in a lateral direction (in order to save on structural weight) and therefore must be kept vertical at all times.

Manufacture

The most comparable rocket vehicle sections in terms of size and shape to Starship are the Saturn V and the Space Shuttle external tank / SLS tank which were / are manufactured horizonally. These parts are designed to handle both gravity and launch stresses without internal pressure. Starship is not so we can conclude that its tanks will be less rigid. The space shuttle external tank has both longitudinal and circumferential stiffeners. Current Starship prototypes appear to have no circumferential stiffeners, and longitudinal stiffeners only on the bottom skirt.

The individual rings from which starship is made are strong vertically, but if stood horizontally they would deform into ovals, to an extent sufficient to complicate welding of the segments together. Hence SpaceX keep the segments vertical when welding them together to avoid ovalling under gravity. The alternative would be to build a huge jig as long as and as wide as the starship to rotate it and keep the parts true round while welding. You can see from photos inside SpaceX's factory that they have facilities to rotate Falcon rockets horizontally in order to work on them.

Balloon tank design

SpaceX have opted for a lightweight but relatively uncommon structural design concept: the balloon tank. To my knowledge this has only been used before in the examples referenced in the link. The idea is to maintain an internal pressure in the tank to give it rigidity, enabling a reduction in structural weight.

Nominally, Starship is supposed to have sufficient structural strength to handle its own weight when standing on the launch pad, but require internal tank pressure to maintain structural rigidity during launch and flight. Therefore the collapse of Starship SN3 due to loss of pressure in its oxygen (lower) tank during a cryotest in its methane (upper) tank came as a surprise.

I have seen it said that the structure would not have failed if the weight in the methane tank had been that of methane, not nitrogen. While this may be true, it's worth considering what would have happened if the methane tank had been full and Starship had been carrying a payload.

The density ratio of nitrogen/methane is 808kg/m3 / 657kg/m3 = 1.23 . The propellant capacity of Starship is 1200 tons, of which a fifth (assuming ideal fuel / oxygen ratio) is methane. 240 tons of methane x 0.23 = 55.2 tons so the excess mass caused by loading the tank with nitrogen instead of methane was around 55.2 tons.

Therefore we can conclude that if SN3 had been carrying a full load of methane and a payload of 55 tons when the pressure in the oxygen tank was lost, it would have collapsed. It shows just how tight Spacex are cutting the structural design on Starship, and it really cannot take any unanticipated loads. I believe they will either need to redesign the oxygen tank or arrange for the launch structure to provide some support to the upper part of Starship, to guard against structural failure in case of loss of pressure in the tanks on the launch pad.

$\endgroup$
3
  • $\begingroup$ most shuttle payloads were integrated vertically $\endgroup$
    – user20636
    May 3, 2020 at 7:53
  • $\begingroup$ Since Starship remains rigid when empty, it's not considered a balloon tank. "By contrast, non-balloon tanks in other liquid-propelled rockets remain rigid while empty due to an internal framework, although they do also depend on internal pressurization to support thrust and launch loads." (From the balloon tank link as of the date of this comment.) $\endgroup$ Nov 25 at 16:49
  • $\begingroup$ @CharlesStaats Empty weight with no load above isn't much. Gravity load on a tank from the payload and any filled tanks above it is the situation that we should consider; that is significantly lower than a launch load. Starship isn't rigid under these conditions. Collapse of Starship 3 under gravity load from part filling the upper (CH4) tank with N2 shows that the empty O2 tank below couldn't withstand gravity load from above while standing on the pad without pressurization, qualifying it as a balloon tank. Also, it has no internal stringers. The booster is probably proportionally even weaker $\endgroup$ Nov 26 at 0:36
6
$\begingroup$

Vertical integration should allow for quicker reuse, if you can do it right. Bring the landed rocket horizontal, loading it, then vertical takes about 2 hours, which if SpaceX does what they ultimately intend to, would be a significant part of the down time between rocket launches. By going full vertical, they can reduce that time.

$\endgroup$
1
$\begingroup$

Likely the change was because of scale.

While a Falcon 9 is not small (230 feet tall and 9 feet wide) in comparison to the Starship/Super Heavy (350 feet tall (final height to be determined) and 30 feet wide) combo it really is small.

The Falcon 9 was specifically designed to be road mobile, that is wide enough that it can fit under bridges and inside a normal car lane so it can easily be moved.

SS/SH is most definitely NOT designed to be easily moved. In fact moving it any time will be complex.

As a consequence, the plan is to actually land in fairly close proximity to the launch pad, making it a task for a crane to move it over, and not require much other transport.

(How they get to the launch site is an excellent other question I look forward to seeing answered in the real world).

Thus while it is true that they will be vertically integrated, in the sense of one loaded on top of the other, it is not really moved the way a Falcon 9 would be, so perhaps the comparison is not entirely apt.

$\endgroup$
1
  • 1
    $\begingroup$ In Boca Chica they're just using Roll Lift heavy transport vehicles to drive them down the road to the launch site vertically. Once the full stack is operational the quick way to move Starship will be to launch from site A and land at site B (although seaborn will still be an option). Supeheavy will need to either be built on site or shipped via barge. KSC and VAFB both have infrastructure to support seaborne delivery; and there are plans (moving very slowly or possibly on hold until closer to when needed) to build a connection from Boca Chica to the Port of Brownsville. $\endgroup$ May 3, 2020 at 4:11

Your Answer

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

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