Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
ball falling in a space colony
#1
I was playing around this evening with simulating actual physics in a realistic space colony.  (No, this isn't something that's going to go into High Frontier.  It's the weekend, cut me some slack!)

Here's my first ever virtual space colony "ball drop" experiment.  The little yellow sphere starts out 6 meters above the deck, initially stationary with respect to the rotating colony.  (Imagine it's being held by somebody on a platform, though that somebody, and the platform itself, are invisible.)  Then we drop it, much like Galileo dropping stones from the Leaning Tower of Pisa.  But we get a behavior that Galileo never saw:


[Image: ZZLq3Tm.gif]

The camera here is lined up for optimal viewing of that slight pull to the left.  In reality, of course, there is no pull to the left... the ball is traveling in a straight line, at a constant velocity from the moment it was released, and the colony is rotating around it.

Details for the curious: the deck here has a 224-m radius and spins at 2 RPM, simulating 1G.  The white ceiling at the top of the view is about 130 m up. Those deck plates are 2 m squares, though unfortunately they don't line up perfectly with the ball's starting position — but if you can detect a slight bend in the plating, that does align with where the ball starts.  So the ball's apparent sideways motion is about a meter or so, over a 6 meter drop.

The physics simulation here is pretty simple.  On each physics timestep, we apply the ball's true (Newtonian) velocity to its position.  Then we account for how much the colony rotates around the ball.  But, for computational efficiency, we do this backwards: the colony stays put, so we counter-rotate the ball position and velocity by the same amount.  That's it.

Finally, note that there is no air here; the ball is falling as in a vacuum.  In a real colony, of course, air would apply a force in the direction of the colony's spin, reducing this Coliolis effect by some amount that depends on the aerodynamics of the object.

Joe Strout
Lead Developer, High Frontier

Reply
#2
Just a shout-out to the National Space Society, who highlighted this on their blog.

Anybody who comes by from there, welcome!  Feel free to register using the link at the top of this page, and join the discussion!

Joe Strout
Lead Developer, High Frontier

Reply
#3
Quote:Finally, note that there is no air here; the ball is falling as in a vacuum.  In a real colony, of course, air would apply a force in the direction of the colony's spin, reducing this Coliolis effect by some amount that depends on the aerodynamics of the object.

Have you done any modelling of the atmosphere in the colony? In particular I'd be curious as to what eddy currents are generated in the strong coriolis environment.
Reply
#4
(09-20-2016, 08:31 AM)William Wrote: Have you done any modelling of the atmosphere in the colony? In particular I'd be curious as to what eddy currents are generated in the strong coriolis environment.

Great question!  I wish I had a great answer.

But unfortunately, that kind of modeling is really hard.  You'd have to use CFD simulation, which is very time-consuming (tends to be run on supercomputers to get a decent answer in a reasonable time) and easy to screw up if you don't know what you're doing.  I looked into it once (reviewing a paper from a guy who attempted it) and found that this probably is a particularly tricky one to model.

On general principles, we would expect the air to be more or less stationary with respect to the shell (that is, to rotate along with it).  That would be the lowest-energy state, since if it's doing anything else, then there's wind inside that you could extract energy from.

But of course it's not going to be perfectly stationary, because it's a compressible fluid, and the individual air molecules can't just travel around in circles.  So they're going to be constantly heading toward the floor, and then knocking into other air molecules and getting pushed back up... as you said, eddy currents will certainly form.  And then too there will be thermal effects, possibly causing convection currents.

It's all quite complicated and hard to simulate in a computer.  I could imagine some interesting benchtop experiments, though.  If anybody here is (or knows) an enterprising high school student in search of a science fair topic, we should have a chat!

Joe Strout
Lead Developer, High Frontier

Reply
#5
(09-20-2016, 09:03 AM)JoeStrout Wrote: But unfortunately, that kind of modeling is really hard.  You'd have to use CFD simulation, which is very time-consuming (tends to be run on supercomputers to get a decent answer in a reasonable time) and easy to screw up if you don't know what you're doing.  I looked into it once (reviewing a paper from a guy who attempted it) and found that this probably is a particularly tricky one to model.

What's the paper? There could be some sort of modelling done just on PCs, since they do 10-100 Gflops, and there's GPU cards that do multi-Tflops.
Reply
#6
(09-23-2016, 11:35 AM)William Wrote: What's the paper? There could be some sort of modelling done just on PCs, since they do 10-100 Gflops, and there's GPU cards that do multi-Tflops.

The paper was not published. I agree that modeling could be done on PCs with enough patience; the real problem is having enough expertise to use these CFD codes properly — the boundary conditions in particular (such as where the air hits the floor of the colony) are easy to screw up.

Joe Strout
Lead Developer, High Frontier

Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)