Artificial Gravity Testbeds

[Coastal Ron]

humans are likely the only animal that will be in sufficient numbers in space for the foreseeable future

Absurd.

One small payload module (the Rodent Research Hardware System) has flown 300+ mice in nine years, as compared to ~630 humans total since the dawn of the Space Age.

And how much did we learn from those experiments that applies to mitigating health effects on humans being in space by using some amount of artificial gravity?

The goals of the those researchers were likely NOT human specific, but just science in general. Not applicable for the problem of humans in space.

Yeah, in general you need a large sample size for any human research studies, but the problem we need to address is pretty immediate. Well, immediate if we are actively trying to expand humanity out into space.

As usual you're arguing backwards from your personal goals ("let's put lots of humans in space for its own sake").

1. Thank you for making me sound as important as Elon Musk, but it is Elon Musk that has really been the driver regarding moving lots of humans into space. Remember he wants to colonize Mars with millions of people, and not once have I heard him mention sending mice to Mars. 

2. Besides the science that NASA does, which is important, if the goal of NASA is NOT to expand humanity out into space, then why are we spending so much money sending humans out into space? This is one of those "Duh!" questions, in that the answer should be pretty apparent.

So YES!!!! My goal is to put lots of humans into space, and I think if you ask others on NSF they will overwhelmingly agree that is the goal too.

However the fundamental problem is still economic, not motivational. We couldn't expand humanity out into space en masse even if we wanted to, because it's just too bloody expensive! So what we really need to do is make space stations that are extremely economical, in the same way SpaceX has done so for launch rockets.

Amazingly I agree with all of that, and yep, my goal has been to focus on lowering the cost of staying in space. And since my professional background has had a focus on cost, I think my designs are the least costly for the value they provide. My opinion of course, and competition is good, so if you think you have something better, great, we'll let the market decide.

That's really our main challenge. However we are facing enemies at the gate who violently oppose any attempt to achieve good economics when designing space colonies:

I think you are getting too emotional here. Who "violently opposes" any artificial gravity design? And when I say "who", I mean people that have money that could actually fund such endeavors. Because if you are talking about people on NSF that are disagreeing with you, that doesn't count, since none of us have the money to decide who is right or who is wrong. And while I HOPE that my designs could be adopted some day, for now I do this for fun and intellectual exercise.

So we may have to deal with small sample sizes in the beginning and then keep medically testing everyone that goes into space that spends time in some form of artificial gravity.

We don't have to do that, no.

We can fly research animals and humans at the same time, just as we always have. In fact, one of the (very few!) economical jobs for humans in space is "biomedical researcher" 

Paul451 is right: the best AG testbed just has racks of animal cages at different G levels. A Vast-style "stick" station is the best way to achieve that.

You have your opinions, and I have mine. And the market will determine who, if any of us, is right. Until that point I plan to continue iterating my ideas.

[Greg Hullender]

Certainly part of the debate has been about the three-dimensional layout of a station while it spins, and how the intermediate axis theorem (aka tennis racket theorem) could affect that rotation.

That's certainly a way for a rotating station to fail catastrophically, but much smaller perturbations could still make it unlivable. If it wobbles like a gyroscope, the result will (likely) be unacceptable, and every time someone moves around inside the thing, it's going to cause at least some amount of wobble.

This is a hard thing to simulate, because there's a certain amount of chaos involved, although--for sure--these days there are ways to cope with that. But nothing would beat orbiting an actual model. Still, either a really sophisticated simulation or a model would be highly informative. The fact that no one seems to be doing either one or even planning either one makes me think that any sort of rotating habitat is decades away.

[Coastal Ron]

Certainly part of the debate has been about the three-dimensional layout of a station while it spins, and how the intermediate axis theorem (aka tennis racket theorem) could affect that rotation.

That's certainly a way for a rotating station to fail catastrophically, but much smaller perturbations could still make it unlivable. If it wobbles like a gyroscope, the result will (likely) be unacceptable, and every time someone moves around inside the thing, it's going to cause at least some amount of wobble.

LOTS of discussion and debate about the potential imperfections of the rotation of future artificial gravity space stations, and all the more reason why we need to test out some of those solutions on testbeds.

But my assumption has been that there will always need to be some sort of active system that tries to keep those imperfections within safe boundaries. And we have to assume that for most of the time that the rotating space stations WILL NOT have perfect rotations, and that they will rotate slightly off-center and have some degree of wobble. Which is why docking systems that can tolerate those imperfections are so important.

I do have an idea for a passive stability system, but I'll need a physics model to validate it.

This is a hard thing to simulate, because there's a certain amount of chaos involved, although--for sure--these days there are ways to cope with that. But nothing would beat orbiting an actual model. Still, either a really sophisticated simulation or a model would be highly informative. The fact that no one seems to be doing either one or even planning either one makes me think that any sort of rotating habitat is decades away.

We don't know how much Vast has looked into stability issues with their 100-meter-long spinning stick space station. As is I think that station will have intermediate axis issues, but the simple drawing on their website may not show important features that will solve that problem, so we need to wait and see.

My proposed "X" station on the first page of this thread is planned for testing out a partially passive stabilization system, but will have active stabilization otherwise. And that design should be inherently stable, but being a testbed it is better to load it up with potential features just in case... 

[Greg Hullender]

But my assumption has been that there will always need to be some sort of active system that tries to keep those imperfections within safe boundaries. And we have to assume that for most of the time that the rotating space stations WILL NOT have perfect rotations, and that they will rotate slightly off-center and have some degree of wobble.

Sure, and exactly how to make that work, how to determine what's tolerable, and how to handle failures gracefully are all tasks that need to be done and evaluated experimentally. But, as far as I can tell nobody is looking at this seriously. There's probably ten or twenty years of work before humans could actually occupy a rotating habitat--including one for a Mars vehicle. And no one has even started the work, so however long it is, it slips another day every day.

[Coastal Ron]

But my assumption has been that there will always need to be some sort of active system that tries to keep those imperfections within safe boundaries. And we have to assume that for most of the time that the rotating space stations WILL NOT have perfect rotations, and that they will rotate slightly off-center and have some degree of wobble.

Sure, and exactly how to make that work, how to determine what's tolerable, and how to handle failures gracefully are all tasks that need to be done and evaluated experimentally. But, as far as I can tell nobody is looking at this seriously.

Though certain NSF members may consider me a nobody (not you), certainly I am concerned about stability and have been looking into active and passive systems to manage it.

There's probably ten or twenty years of work before humans could actually occupy a rotating habitat--including one for a Mars vehicle. And no one has even started the work, so however long it is, it slips another day every day.

We definitely don't have a customer yet for an artificial gravity testbed, much less a full rotating space station.

But the physics involved in a rotating body are not hard to simulate, which means that solutions to instability issues should not be hard to simulate either. Remember that instability is caused by Newton's 3rd law, which states:

To every action, there is always opposed an equal reaction; or, the mutual actions of two bodies upon each other are always equal, and directed to contrary parts.

So the task is imagining what those potential inputs are, and then designing ways to counteract them. Inertia plays a big part of this too, as well as leverage, since inputs at the center of rotation would have little overall affect on the rotation of the body, though it move the body through space. But inputs away from the center of rotation have increasing amounts of affect the further they are from the center of rotation.

Having visiting vehicles docking frequently should not constitute much of an issue, assuming they dock in the planned controlled manner. So in normal operations the main input to the center of rotation and potential wobble would be from the movement of people and mass around the station. That is a slow enough and small enough change that I don't think it will require an immediate response from a stability system.

But what happens if the unexpected happens? Well my design has lifeboats that the crew can use to abandon the station, and I always assume that there is a non-rotating space station nearby rotating space stations, so the crew doesn't have to go far for safety. After that active methods can be used to stabilize the station.

There is an idea I have for a passive system that could slowly stabilize the station, but I would need to run some software simulations to validate it could work.

The bottom line though is that many of our questions and concerns could be addressed if we had a valid software simulation system for rotating space stations. In case anyone reading this would be interested in building such a system... 

[Greg Hullender]

Sure, and exactly how to make that work, how to determine what's tolerable, and how to handle failures gracefully are all tasks that need to be done and evaluated experimentally. But, as far as I can tell nobody is looking at this seriously.

Though certain NSF members may consider me a nobody (not you), certainly I am concerned about stability and have been looking into active and passive systems to manage it.
.
.
.
The bottom line though is that many of our questions and concerns could be addressed if we had a valid software simulation system for rotating space stations. In case anyone reading this would be interested in building such a system... 

Grin. I think having access to such a system is a prerequisite for the seriously part. :-)

To start with, for a given build, I'd like to know how much instability is caused simply by people walking around. Is it too small to notice? Or comparable to a cruise ship at sea? Is it possible to damp it out just by pumping water around? If so, exactly what do you measure in order to compute when and where to pump the water? What other things (besides docking) can cause instability?

That might be a great application for AI, by the way; assuming your simulation is good enough, you can train your networks the same way AlphaZero was trained to play chess. You don't need human-labeled data to identify good and bad choices; the outcome of the simulation itself tells you that. (E.g. the longer the system stays stable past a given decision, the better the score you give that decision.) That would be nice if it saved having to work out the physics, although the inputs to the neural net would still need to be sufficient to (in theory) compute the instantaneous axes and rates of rotation at (probably fairly frequent) intervals of time.

[Coastal Ron]

To start with, for a given build, I'd like to know how much instability is caused simply by people walking around. Is it too small to notice? Or comparable to a cruise ship at sea? Is it possible to damp it out just by pumping water around? If so, exactly what do you measure in order to compute when and where to pump the water? What other things (besides docking) can cause instability?

Using water, or other types of weight, has been proposed. I'm not enamored with the idea because pumping water requires a large amount of infrastructure. It would be great if the water being pumped was part of the radiation mitigation efforts, but I'm not sure that is possible, so all that water would just be dead weight being moved around.

I'm looking at a weight system on the outside of the station, but not ready to discuss it yet.

That might be a great application for AI, by the way; assuming your simulation is good enough, you can train your networks the same way AlphaZero was trained to play chess.

In order to create an A.I. system of any sort, you need a massive amount of training data, which we won't have. In any case, a physics simulator should do just fine to run Monte Carlo what-if scenarios.

[Greg Hullender]

In order to create an A.I. system of any sort, you need a massive amount of training data, which we won't have. In any case, a physics simulator should do just fine to run Monte Carlo what-if scenarios.

Ah, but you do have a massive amount of training data. :-) Every output from your Monte Carlo simulator is training data. The input is the state of the station at time t. The action is whatever operations are available to stabilize the station. (E.g. instead of pumping water, imagine that you have dozens of weights on cables hanging from the rim of the station. Your possible actions are just to raise or lower one or more weights.) The label is whether the simulation says the station would stay stable for (say) 10 or 20 more rotations.

Once you have an initial net, you refine it by making it a part of your simulation. That is, you don't assign the label by freezing everything and running the simulation to see if it's stable; you assume the old net is still running. It's as though you're second guessing the old net at a single point. "What if we did this instead, but then went back to normal?"

There are a variety of ways to set it up, but all of them use the simulation to generate (effectively) an infinite amount of training data. This is pretty much how things like AlphaZero learn chess and go.

[Coastal Ron]

In order to create an A.I. system of any sort, you need a massive amount of training data, which we won't have. In any case, a physics simulator should do just fine to run Monte Carlo what-if scenarios.

Ah, but you do have a massive amount of training data. :-) Every output from your Monte Carlo simulator is training data.

It is data, but training data for an A.I. needs to identify what data results in a correct outcome, and what doesn't. In other words, I'm not sure an A.I. system is going to be any better at controlling balance and wobble than a regular software program that is rule based. And I've been around long enough to remember when neural nets and fuzzy logic were all the rage, so I'm trying not to get overly excited about the current A.I. stuff - it does some amazing stuff, but it doesn't do everything amazing... 

But computing systems will evolve quite a bit between now and when a rotating testbed is finally built in space, so I don't want to rule anything out, but I don't need to debate it either. I just think some sort of computer system will be able to recommend how the station can be kept in proper rotation.

The action is whatever operations are available to stabilize the station. (E.g. instead of pumping water, imagine that you have dozens of weights on cables hanging from the rim of the station. Your possible actions are just to raise or lower one or more weights.)...

All I will say to the last part of your paragraph is 

[Twark_Main]

humans are likely the only animal that will be in sufficient numbers in space for the foreseeable future

Absurd.

One small payload module (the Rodent Research Hardware System) has flown 300+ mice in nine years, as compared to ~630 humans total since the dawn of the Space Age.

And how much did we learn from those experiments that applies to mitigating health effects on humans being in space by using some amount of artificial gravity?

RRHS isn't an artificial gravity research program. You're confusing it with MARS (Multiple Artificial-gravity Research System. https://www.nature.com/articles/s41598-017-10998-4

But you're also missing my point.  RRHS is small. On any testbed station, there's going to be plenty of room for rodents to outnumber people.


The goals of the those researchers were likely NOT human specific, but just science in general. Not applicable for the problem of humans in space.

Yeah, in general you need a large sample size for any human research studies, but the problem we need to address is pretty immediate. Well, immediate if we are actively trying to expand humanity out into space.

As usual you're arguing backwards from your personal goals ("let's put lots of humans in space for its own sake").

1. Thank you for making me sound as important as Elon Musk, but it is Elon Musk that has really been the driver regarding moving lots of humans into space. Remember he wants to colonize Mars with millions of people[/quote]

You want to put people on a space station for its own sake. Elon Musk does not.

Elon Musk wants to put people on Mars.

Those are not the same.

, and not once have I heard him mention sending mice to Mars. 

... that's because Elon Musk wants to colonize Mars. In this thread, the goal is (by definition) medical hypo-gravity research stations.

Different goals, different solutions.

2. Besides the science that NASA does, which is important, if the goal of NASA is NOT to expand humanity out into space, then why are we spending so much money sending humans out into space? This is one of those "Duh!" questions, in that the answer should be pretty apparent.

Yes it should be.

First Google result for NASA mission statement: https://www.nasa.gov/careers/our-mission-and-values/

So YES!!!! My goal is to put lots of humans into space, and I think if you ask others on NSF they will overwhelmingly agree that is the goal too.

That's fine. Me too.

Where you run into trouble is when you assume that your investors and customers share that passion for your goal, and you start spending their money wastefully and justifying it by saying "but... but... the goal is space!!"

yep, my goal has been to focus on lowering the cost of staying in space.

Which Coastal Ron am I talking to again?? 

I think you are getting too emotional here. Who "violently opposes" any artificial gravity design?

Apparently you never any of the replies to my posts.

people on NSF... don't count

Lol. Thanks for proving my point.

[Twark_Main]

As for stabilization, the solutions are well-known.

For course-grain balancing, you use water tanks and pump water around. The pumps and tubing are far lower mass than having large cables/winches moving large weights.

For vibration, the challenge is similar to that encountered in tall buildings. Presumably the analysis tools (FEA software) and solutions will be the same: tuned mass damper water tanks/pendulums.


Naturally these systems could be combined into a single Balancing / Dampening Water System (BDWS).

[Greg Hullender]

It is data, but training data for an A.I. needs to identify what data results in a correct outcome, and what doesn't.

Yes, of course, but what you may not know is that those labels (e.g. good/bad) don't have to be very accurate. As long as they're roughly correct, systems can often use them to find very good solutions.

When I was at Microsoft, we once trained a handwriting recognizer where the labels simply told it which letters were in the word, but not what order they were in. Given enough data, it figured it out anyway. If you look at how AlphaZero was trained or even ChatGPT, you'd be amazed at just how poor the labeling is, considering how good the results are.

So to give a half-assed label to your simulation data, say, you get no labels for the first 10 rotations (or so--you pick the window) of the simulation. But for every time slice after that, the sample from 10 rotations ago gets a label of "GOOD." If it ever goes out of control (i.e. wobbles more than allowed), then you stop, and everything for the past 10 rotations gets a label of "BAD."

Initially, you're really just exploring the space randomly, but later, you're using nets trained with your earlier data to guide the system. Or, whenever it wobbles, go back 10 or 20 rotations and make random tweaks that the net didn't propose to try to get fresh data. There are a lot of ways to play with it.

[Greg Hullender]

As for stabilization, the solutions are well-known.

Are they? Do you have any papers you could cite? I've not found anything, but perhaps I used poor search queries.

[Coastal Ron]

humans are likely the only animal that will be in sufficient numbers in space for the foreseeable future

Absurd.

One small payload module (the Rodent Research Hardware System) has flown 300+ mice in nine years, as compared to ~630 humans total since the dawn of the Space Age.

And how much did we learn from those experiments that applies to mitigating health effects on humans being in space by using some amount of artificial gravity?

RRHS isn't an artificial gravity research program.

Well then it doesn't apply to the goals of an artificial gravity testbed, does it?

Look, you want to study non-human animals, and that is fine. My goal is to as quickly as possible determine what the minimum gravity will be to allow humans to not only survive in space, but also thrive. And since Mars will be a popular destination, it makes sense to test for Mars level of gravity as quickly as possible to contribute to the total science that will be taking place when (not IF!  ) Elon Musk starts settling Mars.

And if Mars gravity is good enough, then the question becomes how much less gravity is good enough? What about the gravity on Earth's Moon? And so on. Replicating the gravity of planetary bodies should be cheaper on an artificial gravity testbed than going to those planetary bodies. That is the theory I'm operating under, and nothing you can say can change that because you are not a potential customer. 

1. Thank you for making me sound as important as Elon Musk, but it is Elon Musk that has really been the driver regarding moving lots of humans into space. Remember he wants to colonize Mars with millions of people

You want to put people on a space station for its own sake. Elon Musk does not.

Elon Musk wants to put people on Mars.

And my goal is to find out if humans can not just survive, but thrive in Mars gravity, which can be simulated on an artificial gravity testbed. And as I've noted on the Realistic, near-term, rotating Space Station thread, one of my stations is specifically designed for Mars level gravity, which I believe will become one of the gravity standards for rotating space stations.

So I think validating Mars level gravity in a rotating space station is important. And if you disagree, I don't care. You do your thing, and I'll do mine. Stop trying to manage my free time... 

So YES!!!! My goal is to put lots of humans into space, and I think if you ask others on NSF they will overwhelmingly agree that is the goal too.

That's fine. Me too.

Where you run into trouble is when you assume that your investors and customers share that passion for your goal, and you start spending their money wastefully and justifying it by saying "but... but... the goal is space!!"

See, this is where you are getting too emotional, because THERE ARE NO INVESTORS YET FOR ANY SPACE STATIONS.

Geez, haven't you realized that we are enthusiasts, and that the likelihood of anything you proposing turning into something real is pretty darn low. Same for me too, but you don't seem to realize that.

Plus, and this is important to remember, YOU ARE NOT MY TARGET CUSTOMER. Get over it.

[Coastal Ron]

It is data, but training data for an A.I. needs to identify what data results in a correct outcome, and what doesn't.

Yes, of course, but what you may not know is that those labels (e.g. good/bad) don't have to be very accurate. As long as they're roughly correct, systems can often use them to find very good solutions.

A.I. is a popular topic in my social and work circles, and I just attended a speech by an A.I. expert recently that was hosted by a friend of mine that is also an A.I. expert (going back to neural network days).

My personal thinking is that while A.I. could be used in this application, I'm just not sure it will be any better than a rules-based system.

Also, current LLM's are not good at math (horrible actually), and while regular machine learning algorithms could be taught based on a rules-based Monte Carlo simulation, I'm not sure if that would end up being better than just relying on a regular rules-based system.

However some sort of control system will be needed, it just that who knows what will happen in 10 years?

[Twark_Main]

humans are likely the only animal that will be in sufficient numbers in space for the foreseeable future

Absurd.

One small payload module (the Rodent Research Hardware System) has flown 300+ mice in nine years, as compared to ~630 humans total since the dawn of the Space Age.

And how much did we learn from those experiments that applies to mitigating health effects on humans being in space by using some amount of artificial gravity?

RRHS isn't an artificial gravity research program.

Well then it doesn't apply to the goals of an artificial gravity testbed, does it?

Again you spectacularly missed my point, even to the point of intentionally editing it out of my post when you quoted me.

On an AG testbed, you can easily have racks and racks of RRHS-like animal habitats.

Look, you want to study non-human animals... My goal is to as quickly as possible determine what the minimum gravity will be to allow humans to not only survive in space, but also thrive.

In other words, you want to jump to risky large-scale human tests without retiring any of the risk on non-human and unmanned tests first.

Does the Titan come to mind, anyone?

nothing you can say can change that because you are not a potential customer. 

[repetition snipped]

"I won't accept any constructive feedback from anyone but Elon Musk himself" is a weird position. Doesn't seem like a recipe for success, but okay.

[Coastal Ron]

Well then it doesn't apply to the goals of an artificial gravity testbed, does it?

Again you spectacularly missed my point, even to the point of intentionally editing it out of my post when you quoted me.

On an AG testbed, you can have racks and racks of RRHS-derived animal habitats.

I don't care about livestock in space. Or rabbits in space, or mice in space. They are not my target market.

Look, you want to study non-human animals

It's not that. I'm just pointing out that you're rejecting it for weird and unjustified reasons.

By weird you mean I'm focused on my target market, and not some unrelated market? Yep.

And here's the thing, I don't think I'll have any problem finding research subjects. And 100% of the science returned will be related to the ultimate goal of finding out whether humans can survive in artificial gravity that is significantly less than one Earth gravity.

But if you want to study animals instead of humans, go do it. Just STOP BUGGING ME ABOUT MY GOALS.

My goal is to as quickly as possible determine what the minimum gravity will be to allow humans to not only survive in space, but also thrive.

In other words, you want to jump to risky large-scale human tests without retiring any of the risk on non-human and unmanned tests first.

Does the Titan come to mind, anyone?

What did I tell you about bad analogies? You throw a word out that in your mind contains your entire argument, yet it turns out that your audience sees no relevance - because there is none.

And you have yet to outline what any of your SPECIFIC concerns are, since we do human research in space all the time and you don't seem to object to that.

But let me be honest. Long ago you stopped providing any useful feedback, so I would be quite happy if you just considered me to be hopelessly clueless and left me to stumble around on my own. How about that? Because it should be quite clear by now that we just don't see eye to eye on pretty much anything, and you have become boorish to me.

[Coastal Ron]

If artificial gravity testbeds are seen as a mandatory step to take before committing to full-sized operational rotating space stations, then there will be a need for human testing at some point.

Artificial gravity testbeds like the Vast 100-meter-long spinning stick space station are simple designs that could offer the least expensive and quickest method of doing human testing in space, but they don't offer a lot of floor space for test subjects to stay in specific artificial gravity ranges.

However science has been able to find human participants to endure many types of experiments, and NASA just announced that they have started a 4-person 378-day experiment oriented towards sending people to Mars. Pretty low fidelity, but four people will be sharing a 1,700-square-foot habitat, so not exactly a mansion.

Besides testing out specific gravity gradients (i.e. Moon gravity, Mars gravity, etc.), what other tests need to be done to pave the way for building full-sized operational rotating space stations?

[leovinus]

If artificial gravity testbeds are seen as a mandatory step to take before committing to full-sized operational rotating space stations, then there will be a need for human testing at some point.

Artificial gravity testbeds like the Vast 100-meter-long spinning stick space station are simple designs that could offer the least expensive and quickest method of doing human testing in space, but they don't offer a lot of floor space for test subjects to stay in specific artificial gravity ranges.

However science has been able to find human participants to endure many types of experiments, and NASA just announced that they have started a 4-person 378-day experiment oriented towards sending people to Mars. Pretty low fidelity, but four people will be sharing a 1,700-square-foot habitat, so not exactly a mansion.

Besides testing out specific gravity gradients (i.e. Moon gravity, Mars gravity, etc.), what other tests need to be done to pave the way for building full-sized operational rotating space stations?

One way to answer your question "what other tests?" is to look back and study in more detail the published literature since the early 60s. A vast amount at NTRS and elsewhere. If you are interested in the angle from an "historical artificial testbed" thread incl. human and machine responses then happy to contribute papers and material.

[Coastal Ron]

...
Besides testing out specific gravity gradients (i.e. Moon gravity, Mars gravity, etc.), what other tests need to be done to pave the way for building full-sized operational rotating space stations?

One way to answer your question "what other tests?" is to look back and study in more detail the published literature since the early 60s. A vast amount at NTRS and elsewhere. If you are interested in the angle from an "historical artificial testbed" thread incl. human and machine responses then happy to contribute papers and material.

I see that you have a big interest in the historical perspective of spaceflight, and sure, if you can find references to what early NASA thought would be the challenge with less-than-Earth gravity for long periods of time in space, that might be interesting.