Artificial Gravity Testbeds

[Paul451]

Also, if they stick with the crane idea to lift it onto Super Heavy, then the structure will also need to support the loaded-but-unfuelled mass hanging from the nose.

Hence all the suggestions to 'hang' it from the nose for spin-gravity. The lifting lugs are after all right there...

You missed that it would therefore be designed to support itself under multi-g compressive forces and also under at least 1g tension. It's seems unreasonable to suppose that the structure suddenly becomes super weak if you reverse the direction of the forces.

The thrust structure must be able to support and transfer the compressive load from the engines through to the tank walls. (Along with tolerating the chaos and vibration of engine firing.) It also must be able to support the tensile loads of the inactive engines hanging from it (when stacked on SH, during ground handling, after landing, etc, etc.) The tank walls must be able to support both the compressive load of launch, while fuelled, and the tensile load of having themselves and the engines hanging from the nose during crane-stacking. The forward section must be able to support the entire rest of the ship hanging from it during crane-loading.

What additional force does hanging from the tail (for tail/tail AG) add that Starship wouldn't be able to support?

The only thing I can see is that the attachment points in the tail-section/thrust-frame for staging won't be specifically designed for tension. The thrust-frame must be able to. But the actual hard-points intended for SH stacking might not. But that's a fairly small modification. (If necessary. They might be fine. They have to handle the vibration and pitching of a fully fuelled Starship on top of the thrusting SH. That's a long arm of momentum across a small area. They must be stupidly robust.)

And horizontally, belly-side, at however many g's it pulls during re-entry, plus the bending force from aerodynamic pressure in that horizontal position.

Re-entry loads are spread across the entire heatshield surface. If you want to lift (or 'hang', for spin-gravity) in the same orientation, either you need to bring up a Starship-sized cradle or strengthen the structure for point-loading.

[cough]fins[/cough]

[edzieba]

What additional force does hanging from the tail (for tail/tail AG) add that Starship wouldn't be able to support?

I'm sorry, but trying to explain from first principles why structures designed to support a load in one direction do not automatically work with the inverse of that direction is getting into the realms of the absurd.

[Paul451]

What additional force does hanging from the tail (for tail/tail AG) add that Starship wouldn't be able to support?

I'm sorry, but trying to explain from first principles why structures designed to support a load in one direction do not automatically work with the inverse of that direction is getting into the realms of the absurd.

I didn't say "automatically", nor universally, I'm referring to the specific case of Starship, based on its specific design requirements.

[mikelepage]

What additional force does hanging from the tail (for tail/tail AG) add that Starship wouldn't be able to support?

I'm sorry, but trying to explain from first principles why structures designed to support a load in one direction do not automatically work with the inverse of that direction is getting into the realms of the absurd.

I didn't say "automatically", nor universally, I'm referring to the specific case of Starship, based on its specific design requirements.

Well, hey there chaps.  I didn't mean to be away from the forum for two months, but glad to see nothing has changed 

It occurs to me that this whole debate may become a moot point, because the design for the propellent transfer apparatus will depend on the in-orbit propellent transfer experiments that SpaceX is planning for the first set of orbital flights.  How confident are we that Starship will continue to use tail-to-tail propellent transfer at all? 

Playing devil's advocate, it is interesting that they chose to depict tail-to-tail transfer in the video, since pumping propellent *through* Superheavy up to Starship before Earth launch adds an extra element of complication when they could simply side load propellent to Starship from the strong-back/tower/crane, the same way they do with F9 second stage.  Likewise in orbit, prop transfer using milli-g acceleration ought to work no matter what orientation the Starships are in**.  Perhaps it might actually be better to have two starships dock "back-to-back" (offset to avoid tail fin conflict) and able to transfer propellent *and* crew, without having to perform multiple docking maneuvers.

If that came to pass, it would rule out the tail-to-tail spin-gravity option, and the nose-to-nose cabling option has always had its problems too.  I wonder if the most feasible goal for any testbed is still to aim for a dedicated spin gravity habitat that can work in LEO, and that every nth Starship carries with the fleet to Mars, so that astronauts can have those comforts/benefits periodically on the way to Mars, and (more importantly) to help acclimatise astronauts to higher gravity on the way back to Earth.

**EDIT: Just realised I've always assumed the milli-G is to keep the pumps primed that will do all the work of moving the fluids, but I suppose they could mean the acceleration is doing all the work (which would probably require tail-to-tail).

[Twark_Main]

Each of the modules on the X could have solar panels attach that would extend to provide power for the station.  Kind of like Dragon I service module or Soyuz.  Don't know why ISS didn't do something like this.

Nitpick: they did, on both the Zvezda and Zarya modules.

[snip]

**EDIT: Just realised I've always assumed the milli-G is to keep the pumps primed that will do all the work of moving the fluids, but I suppose they could mean the acceleration is doing all the work (which would probably require tail-to-tail).

Active pumping seems most likely. That would minimize both refilling time and ullage propellant use.

I always thought tail-to-tail was mostly about eliminating the second connection, and re-using the on-pad filling lines. "The best part is no part."

For crew transfer, I like the idea of putting a NDS port on the cargo door (as seen in one of the SpaceX renders). That way you can easily swap out the door, so you only carry the mass penalty on those missions that require it. Two cargo doors provides redundancy.

[edzieba]

What additional force does hanging from the tail (for tail/tail AG) add that Starship wouldn't be able to support?

I'm sorry, but trying to explain from first principles why structures designed to support a load in one direction do not automatically work with the inverse of that direction is getting into the realms of the absurd.

I didn't say "automatically", nor universally, I'm referring to the specific case of Starship, based on its specific design requirements.

And there are no design requirements that involve Starship being hung inverted from the thrust structure under and appreciable acceleration (gravitational, rotational, or linear).

[Twark_Main]

What additional force does hanging from the tail (for tail/tail AG) add that Starship wouldn't be able to support?

I'm sorry, but trying to explain from first principles why structures designed to support a load in one direction do not automatically work with the inverse of that direction is getting into the realms of the absurd.

I didn't say "automatically", nor universally, I'm referring to the specific case of Starship, based on its specific design requirements.

And there are no design requirements that involve Starship being hung inverted from the thrust structure under and appreciable acceleration (gravitational, rotational, or linear).

Obviously. But their point is (and they're right) that the primary structures SpaceX is using for the design loads would easily support those out-of-design loads.

[edzieba]

What additional force does hanging from the tail (for tail/tail AG) add that Starship wouldn't be able to support?

I'm sorry, but trying to explain from first principles why structures designed to support a load in one direction do not automatically work with the inverse of that direction is getting into the realms of the absurd.

I didn't say "automatically", nor universally, I'm referring to the specific case of Starship, based on its specific design requirements.

And there are no design requirements that involve Starship being hung inverted from the thrust structure under and appreciable acceleration (gravitational, rotational, or linear).

Obviously. But their point is (and they're right) that the primary structures SpaceX is using for the design loads would easily support those out-of-design loads.

I do not agree. SpaceX are already optimising tank wall thickness for a non-inverted orientation.

[Paul451]

Obviously. But their point is (and they're right) that the primary structures SpaceX is using for the design loads would easily support those out-of-design loads.

I do not agree. SpaceX are already optimising tank wall thickness for a non-inverted orientation.

They are optimising the tank walls for compression under full fuelled launch loads, and internal tension from the huge forces produced by over 1000 tonnes of prop under compression during launch. The "hanging" tensile strength is simply a bonus from the demands of those other two things.

I actually suspect that the nose-crane vehicle stacking idea wasn't something that was added to the requirements, ie, strength that will need to be added to the vehicle's structure, but instead simply fell out of the strength required for everything else. (With the exception of the actual hard-point in the nose, presumably.)

And the same applies to tail-hanging. Everything else means Starship is over-engineered for the comparatively low tensile loads produced by hanging a nearly unfuelled Starship from its thrust-frame.

Eg, in the F9 Payload User Guide, the vibration loads can be up to ±2g laterally at the same time as +3.5 to -1.5g axially. Obviously the upper stage needs to handle similar loads, going from compression to tension, including the bending/twisting force on the structure and connections due to the lateral loads. Oh, notice that negative 1.5g axial load. In other words, hanging from the interstage at more than 1g. (I assume the negative g-loads are mainly the jerk force during MECO and staging.)

Starship could be a gentler launcher, but designing a vehicle to "hang" 1000+ tonnes from the interstage at negative 1g at SH MECO makes the issue of "hanging" a couple of hundred tonnes (ship, payload, landing prop) at a 1 pseudo-g during rotation seem trivial. (Just as dealing with 2g lateral loads means that a fraction-of-1g bending loads during spin-up/down is trivial.)

I'm sorry, but trying to explain from first principles why structures designed to support a load in one direction will support lower loads in the same direction is getting into the realms of the absurd.

FTFY

[Twark_Main]

What additional force does hanging from the tail (for tail/tail AG) add that Starship wouldn't be able to support?

I'm sorry, but trying to explain from first principles why structures designed to support a load in one direction do not automatically work with the inverse of that direction is getting into the realms of the absurd.

I didn't say "automatically", nor universally, I'm referring to the specific case of Starship, based on its specific design requirements.

And there are no design requirements that involve Starship being hung inverted from the thrust structure under and appreciable acceleration (gravitational, rotational, or linear).

Obviously. But their point is (and they're right) that the primary structures SpaceX is using for the design loads would easily support those out-of-design loads.

I do not agree [that the primary structures would easily support those loads]. SpaceX are already optimising tank wall thickness for a non-inverted orientation.

Do you have some numbers to support that? Because my math disagrees.

SpaceX varies the tank thickness vertically because of hydrostatic pressure, but remember that for the proposed usage the tanks will be nearly/entirely empty, and won't be subject to 3+ g flight loads.

The tank wall thickness optimization you mentioned means the top of the upper tank is the thinnest, but it still must be thick enough to support 6 bar in-flight (plus margins on top of that). That's a hoop stress of 2700 kN/m. If we assume the propellant tank are not vented but remain pressurized at 6 bar (this is intentionally a worst-case assumption that sandbags my argument), that pressure will cause a longitudinal stress of 1350 kN/m. But the metal is thick enough to support the hoop stress of 2700 kN/m, so it has 1350 kN/m of tensile "overhead." Multiplied by the ~28 m circumference, that's a total of 38,000 kN, or about 3,900 tonnes-force.

So as long as your fairing/payload section doesn't weigh more than that, there's plenty of strength. Heck, even assuming 1 bar (eg thinner metal in the payload section), that's still over 600 tonnes-force of weight.

TL;DR pressure vessels deal with much greater forces than gravity structures. Incidentally this is why internal pressure, not weight, dominates the structural design of space/surface habs.

[edzieba]

Unless you intend them to be permanently and irrevocably converted, any Starship would need to contain sufficient propellant to deorbit and return. That may be low for LEO, but for other uses (e.g. HEO or a Lagrange station) is it appreciable.


But the entire discussion seems meaninglessly moot. It is akin to arguing whether or not you could tow trailer from the bike rack on the back (microgravity fluid couplers), purely because you refuse to walk to the other side where the tow hitch is (actual lifting hardpoints).

[Paul451]

microgravity fluid couplers

No. Don't strawman. No-one has suggested using the fuel line itself to attach the two Starships. The fuel coupler is not what is being used to attach the Starship to the Super Heavy, so it isn't what people are proposing for this.

We're saying:
1: "the frame of the trailer is capable of being pushed or pulled from either direction at walking speed".
2: "these other attachment points on the back of the trailer, here here here, are designed to mount the trailer vertically on top of the roof, while it's being subject to hurricane force winds, therefore it'll probably be able to hang from those same attachment points inside my shed."

[edzieba]

microgravity fluid couplers

No. Don't strawman. No-one has suggested using the fuel line itself to attach the two Starships. The fuel coupler is not what is being used to attach the Starship to the Super Heavy, so it isn't what people are proposing for this.

Neither did I. Whatever mechanical couplers are used for holding Starships together tail-to-tail for microgravity fluid transfer will be designed for... microgravity fluid transfer. Not Earth-g or Mars-g.

We're saying:
1: "the frame of the trailer is capable of being pushed or pulled from either direction at walking speed".
2: "these other attachment points on the back of the trailer, here here here, are designed to mount the trailer vertically on top of the roof, while it's being subject to hurricane force winds, therefore it'll probably be able to hang from those same attachment points inside my shed."

There are no existing couplers sitting at the bottom of Starship intended to hang the entire thing under 1g loads. While Starship (or any upper stage) is stacked on the booster, the predominant force is going to be a combination of gravitational acceleration and the acceleration of the booster. Many rocket get away with a simple slip-fit of upper staged onto lower stages.

[Twark_Main]

microgravity fluid couplers

No. Don't strawman. No-one has suggested using the fuel line itself to attach the two Starships. The fuel coupler is not what is being used to attach the Starship to the Super Heavy, so it isn't what people are proposing for this.

Neither did I. Whatever mechanical couplers are used for holding Starships together tail-to-tail for microgravity fluid transfer will be designed for... microgravity fluid transfer. Not Earth-g or Mars-g.

We're saying:
1: "the frame of the trailer is capable of being pushed or pulled from either direction at walking speed".
2: "these other attachment points on the back of the trailer, here here here, are designed to mount the trailer vertically on top of the roof, while it's being subject to hurricane force winds, therefore it'll probably be able to hang from those same attachment points inside my shed."

There are no existing couplers sitting at the bottom of Starship intended to hang the entire thing under 1g loads. While Starship (or any upper stage) is stacked on the booster, the predominant force is going to be a combination of gravitational acceleration and the acceleration of the booster. Many rocket get away with a simple slip-fit of upper staged onto lower stages.

Paul451 already addressed this upthread:

The only thing I can see is that the attachment points in the tail-section/thrust-frame for staging won't be specifically designed for tension. The thrust-frame must be able to. But the actual hard-points intended for SH stacking might not. But that's a fairly small modification. (If necessary. They might be fine. They have to handle the vibration and pitching of a fully fuelled Starship on top of the thrusting SH. That's a long arm of momentum across a small area. They must be stupidly robust.)

[Paul451]

There are no existing couplers sitting at the bottom of Starship intended to hang the entire thing under 1g loads.

Actually, there are. As I mentioned regarding the F9 user guide. Payloads (and hence presumably the upperstage, hence the interstage connectors) not only have to handle positive axial loads of up to 6g, they also have to handle negative 2g.

Negative. Meaning it's hanging forward. Under tension. At 2g. Fully fuelled.

The equivalent for Starship is when it masses over 1000 tonnes. If Starship can tolerate hanging 1000 tonnes, under tension, off the thrust-frame at 1g, let alone 2g, then it can tolerate hanging a couple of hundred tonnes under 1g rotational in the same direction.

In my opinion, at this point in the argument, the burden is on you to show that the Starship's thrust-frame would be significantly weaker.

[edit: I like that I'm contradicting Twark Main quoting me. Coz I can no longer "see" that the staging couplers aren't "designed for tension". They clearly must be.]

It's possible the existing interstage-couplers might not be suited to tail-tail AG. (You'd want something designed more as a grabble/docking/mating system, after all, compared to the existing connectors.) However, everything I can find says that the hardpoints on the thrust-frame that those couplers are attached to are more than adequate for such loads. Probably an order-of-magnitude over-engineered for it. You may need to design a new coupler designed for the specific job, you don't have to "permanently, irrevocably" redesign the rest of the Starship.


The more I look into it, the more convinced I am that the tail-tail orientation is probably better engineered for this than the nose-nose variant. (Previously I preferred nose-nose or even side-to-side, for other reasons.)

So the real issue isn't structural. It's just the annoyance of having to operate some systems upside down. A lot of that can be solved by designing modules that can be flipped, and the advantage of having plumbing that works under gravity IMO outweighs the disadvantage of designing them to rotate.

Which brings us to MikeLepage and Lamontagne to design those systems.


Sorry I didn't link to this before.

https://www.spacex.com/sites/spacex/files/falcon_users_guide_10_2019.pdf

(Page 15 for a neatly graphed acceleration map.)


[Edits: Formatting, phrasing, spelling, etc. Bunch of stuff.]

[edzieba]

The payload guide includes all flight phases, including those long after booster separation, where you have a near-empty upper stage kicking around with RCS.

Load encountered by the interstage links is going to be be from 3g towards the booster, to a small force away from the booster (static load under microgravity from the pressurised pushers in the event a latch fails to unlatch and must shear as a failsafe to ensure separation).

Think of it logically: what force and from where is going to be applied to the upper stage while the booster is attached and under acceleration that could both counter the ~3g acceleration from the booster and provide negative 2g (i.e. accelerate a fuelled upper stage at 49ms^2, or in other words apply a minimum of 4.5 MN of force)?

[Paul451]

Think of it logically: what force and from where is going to be applied to the upper stage while the booster is attached and under acceleration that could both counter the ~3g acceleration from the booster and provide negative 2g (i.e. accelerate a fuelled upper stage at 49ms^2, or in other words apply a minimum of 4.5 MN of force)?

A single, simple example: The jerk-force (delta_acc) during MECO would produce a rebound pseudo-force. The effect on an acceleration-compressed liquid, methane/LOx, only needs to differ from the effect on a solid vehicle by 1m/s² in order to produce a momentary 2g forward acceleration on the vehicle. I'm sure there are more; including the twisting lateral forces produced by supersonic flight under rocket thrust, the length of the upperstage+payload (ditto Starship) gives a huge moment arm, that's a lot of torque force on the interstage couplers.

(The same jerk-force during SECO is probably what accounts for the 6g rearward.)

The payload guide includes all flight phases, including those long after booster separation, where you have a near-empty upper stage kicking around with RCS.

So you think the 400N Draco thrusters can produce 6g's?

(And where do you think these thrusters are attached to when they are producing 6g of thrust? Is it the thrust-frame of the upperstage? With none on the nose? How does that produce negative 2g's? If we're "thinking logically".)


You've provided no arguments except your own personal incredulity. You've produced nothing factual to contradict anything that I or TM have provided or calculated except your own disbelief. Do you understand that, logically, an argument that boils down to you saying "I don't believe it" requires no counter-evidence except the statement "And yet it is"?

It also seems that you've given up arguing that the tanks can't handle the tensile force required. You seem to have given up arguing that the thrust frame can't handle the tensile force required. And I hope you're not arguing that the nose section can't handle the tensile force required.

So you're down to arguing that only the connectors (shaft and collets) that hold the upperstage to the interstage can't handle the tensile force required. If so, can we just stop and say the case is proved, and we can all agree that we just need to add a new attachment to the thrustframe? And then move on to talking about the actual problems with tail-tail AG?

[edzieba]

So far, the only argument offered for "Starship's existing latching hardware can support an inverted Starship" has been "the Falcon payload guide has a -2g payload survivability requirement", which is obtuse at best.

As for the argument that propellant compressibility would cause significant acceleration on rebound, that I also don't buy. From here, LOX compressibility at 66K is a mere ~7.5 x 10^-6 in^2 /lb.f (~ 1.1×10−9 m^2/N), and RP-1 at 266K is 3.75 x 10^-6 in^2 /lb.f (~ 5.4×10−10 m^2/N), so both effectively incompressible.

The 6g rearward force is a much more mundane "you've strapped your payload to an accelerating rocket" induced force rather than a more exotic transient. Dracos are present on Dragon, F9 stage 9 just has N2 cold-gas thrusters (which can not only produce more acceleration than you would expect on a depleted stage, the stage itself can produce apparent negative gs relative to the payload adapter when it reorients, or even deliberately spin as in Starlink deployment).

It also seems that you've given up arguing that the tanks can't handle the tensile force required. You seem to have given up arguing that the thrust frame can't handle the tensile force required. And I hope you're not arguing that the nose section can't handle the tensile force required.

No, I've just seen zero evidence to indicate either can.
While Starship is not quite a full balloon tank (in that it can self-support unloaded), it is sufficiently balloon that skin thickness is already orientation optimised just to withstand normal operating conditions. Design for structures under compression is not the same as design for structures under tension, and there remains no evidence that Starship has been designed to sustain negative g loads when hung from the thrust structure. Offer some, or at least some back-of-the-napkin math, and it may be worth considering (you know, the usual "What can be asserted without evidence can also be dismissed without evidence.").

[Paul451]

Design for structures under compression is not the same as design for structures under tension, and there remains no evidence that Starship has been designed to sustain negative g loads when hung from the thrust structure. Offer some, or at least some back-of-the-napkin math, and it may be worth considering (you know, the usual "What can be asserted without evidence can also be dismissed without evidence.").

TM showed (with maths!) that propellant tanks are designed to handle more tensile loads than is required for vertical hanging. I've showed that the F9 requires payloads to handle both tensile and compressive loads in excess of 1g in any direction.

You've offered literally nothing except your own personal incredulity, and now insults.

[mikelepage]

TM showed (with maths!) that propellant tanks are designed to handle more tensile loads than is required for vertical hanging. I've showed that the F9 requires payloads to handle both tensile and compressive loads in excess of 1g in any direction.

SpaceX varies the tank thickness vertically because of hydrostatic pressure, but remember that for the proposed usage the tanks will be nearly/entirely empty, and won't be subject to 3+ g flight loads.

The tank wall thickness optimization you mentioned means the top of the upper tank is the thinnest, but it still must be thick enough to support 6 bar in-flight (plus margins on top of that). That's a hoop stress of 2700 kN/m. If we assume the propellant tank are not vented but remain pressurized at 6 bar (this is intentionally a worst-case assumption that sandbags my argument), that pressure will cause a longitudinal stress of 1350 kN/m. But the metal is thick enough to support the hoop stress of 2700 kN/m, so it has 1350 kN/m of tensile "overhead." Multiplied by the ~28 m circumference, that's a total of 38,000 kN, or about 3,900 tonnes-force.

So as long as your fairing/payload section doesn't weigh more than that, there's plenty of strength. Heck, even assuming 1 bar (eg thinner metal in the payload section), that's still over 600 tonnes-force of weight.

Repeating Twark_Main's calculations for reference (thanks for doing the calcs!).

If so, can we just stop and say the case is proved, and we can all agree that we just need to add a new attachment to the thrustframe? And then move on to talking about the actual problems with tail-tail AG?

Agreed.  I'm more keen to discuss the issues with keeping landing propellent cold, and deploying/retracting solar panels in any spin-gravity scenario.  At least for the latter, I suspect deployment/retraction of solar panels has to happen before/after spin phases, but I haven't yet conceptualised what that looks like.

As I've said in previous posts, ideally from a design perspective, spin gravity should not be an *additional* cost to running the ship during cruise phase, rather, it should actually make many things *easier*.