Getting hit by space debris.

[Milkywegian]

I was watching this video

and wondered what are the chances of getting hit by a space debris? Because i think chances of getting hit by a debris while orbiting earth is pretty low.

Edited April 26, 2018 by Milkywegian

[virchau]

I'm not sure about any old orbital object, but the ISS constantly tracks anything within 25km normal and 2km radial of itself, and if the chance of collision is above 1 in 10,000, the ISS does a maneuver to go into a less risky position.

So no, the ISS is not going to get hit anytime soon.

[sevenperforce]

  On 4/26/2018 at 3:12 PM, virchau said:

I'm not sure about any old orbital object, but the ISS constantly tracks anything within 25km normal and 2km radial of itself, and if the chance of collision is above 1 in 10,000, the ISS does a maneuver to go into a less risky position.

So no, the ISS is not going to get hit anytime soon.

Expand  

Unless it's too small for the ISS to spot.

[kerbiloid]

Or has a high elliptic orbit and hits it from top.

[Nibb31]

  On 4/26/2018 at 3:59 PM, kerbiloid said:

Or has a high elliptic orbit and hits it from top.

Expand  

Anything that hits the ISS from a vertical angle isn't going to be in much of an orbit for long, which means that it wasn't in much of an orbit to begin with.

[kerbiloid]

  On 4/26/2018 at 4:09 PM, Nibb31 said:

from a vertical angle

Expand  

arctan(2/25) = 5°

Is 400x600 km orbit a vertical angle?

[Pand5461]

Use AI to compute a motion of a dynamical system when there is a myriad of numerical integration methods with proven accuracy and rate of convergence...

Probably CS students who learn things beyond buzzwords are too awkward to show in a fancy video.

[YNM]

  On 4/26/2018 at 4:17 PM, kerbiloid said:

Is 400x600 km orbit a vertical angle?

Expand  

Depends. We're talking about relative velocities here.

  On 4/26/2018 at 9:47 AM, Nafiu Sean said:

Because i think chances of getting hit by a debris while orbiting earth is pretty low.

Expand  

Given the holes, it's actually quite high.

But then we need to care of momentum, not just amount.

 

  On 4/26/2018 at 5:05 PM, Pand5461 said:

Use AI to compute a motion of a dynamical system when there is a myriad of numerical integration methods with proven accuracy and rate of convergence...

Expand  

Actually, might be a better way to obtain loads of "coefficients" rather than to use normal stuff. Count it under "solver", not "AI".

[Tullius]

  On 4/26/2018 at 4:17 PM, kerbiloid said:

arctan(2/25) = 5°

Is 400x600 km orbit a vertical angle?

Expand  

Very few things crossing the orbit of the ISS will be in such an eccentric orbit, since the atmospheric drag has the tendency to circularise orbits.

[Pand5461]

  On 4/26/2018 at 5:09 PM, YNM said:

Actually, might be a better way to obtain loads of "coefficients" rather than to use normal stuff. Count it under "solver", not "AI". 

Expand  

To fit coefficients into the known functional form for force dependence on position and velocity - yes, maybe.

Certainly not for the actual path prediction.

[YNM]

  On 4/26/2018 at 5:45 PM, Pand5461 said:

Certainly not for the actual path prediction.

Expand  

Well, a 15-thousand variable function have no problem to fit to a 3-variable function.

But it's certainly not too efficient. But what do you think computing today is, if not using as much as possible.

[Kesa]

  On 4/26/2018 at 5:05 PM, Pand5461 said:

Use AI to compute a motion of a dynamical system when there is a myriad of numerical integration methods with proven accuracy and rate of convergence...

Probably CS students who learn things beyond buzzwords are too awkward to show in a fancy video.

Expand  

It's sure you won't heard of runge kutta and whatnot in fancy videos, but I would not dismiss using neural network so easily.

The numerical integration methods are only accurate relative to the model you use, not necessarilly to reality. If we don't know precisely the gravitational field of Earth, including its anomalies, or the effect of Sun pressure over a small poorly identified debris, the integration will only make accurate prediction relative to an innacurate model.

Neural network can be a way automatically fill our ignorances about small and hardly predictible factors which weight heavily over time.

(I would expect the most efficient solution to be asking NN to predict deviations from current prediction rather than asking them for the whole prediction though)

[Xd the great]

It depends on the size of the debris and the orbit of your craft. 

Short answer: very small debris can leave small holes on your craft. They often hit the space stations. Not very dangerous, but it is like using a gun to take down an amoured vehicle. Over time, the damage will show. Larger debris is less common, but can wreck even more havoc. Try to imagine a railgun shot through the same vehicle. Some orbits are more used than others, so more debris will be in the area.

Even shorter answer: salyut expelled astronauts pee into space. The pee became crystals and hit the station. Caused a massive drop in energy generation at the end of its life as the solar panels are wrecked.