100 Years of Orbital Motion In The Kerbin System Simulated

[illectro]

I just took a low quality Adaptive stepsize Runge-Kutta ODE solver and ran it against a model of the Kerbin system (including the sun). The step precision was set to 10^-7 and we normalized the coordinate system to the barycenter of the Kerbin system.

Even using this primitive solver the simulation covered about 30 days of game time per second on my laptop (I\'m sick of uninformed people claiming that computing all the gravity forces is computationally impossible BTW). Using a better algorithm (say a higher order predictor-corrector method) would make this significantly faster. Somewhere down the line It\'d be nice to see a proper n-body gravity model added, there\'s no technical reason that makes this impossible, just a limit of dev resources.

As you can see the orbit of Minmus looks erratic but it\'s likely to be stable over long periods since the Hill\'s Sphere of the Kerbol-Kerbin system is about 136,000km. All those people having trouble reaching minmus are no doubt happy it\'s kept on rails

I\'ll probably make some more of these and do some commentary when I have

(BTW, DEVS, if you\'re planning on adding any new planets, I\'ll be happy to use my astro-Fu and tell you whether the proposed additions are stable).

[Talisman]

That\'s very awesome! Have you experimented with adding a dominant gas giant body in the system? Or making a full system? Would that be possible with the current kerbal sun?

[illectro]

Oh sure, we can add anything to this model, it\'s just important that new bodies be far enough away from their neighbours not to mess up the long term stability.

[maltesh]

Nifty. Was the Mun\'s orbit inclined by Minmus interaction in the simulation?

[illectro]

Yes the mun and geostationary satellite pick up a bit of inclination over time, and in turn minmus loses inclination to the mun - the system oscillates. I should really go and plot some parameters over time.

[PakledHostage]

I just took a low quality Adaptive stepsize Runge-Kutta ODE solver and ran it against a model of the Kerbin system (including the sun). The step precision was set to 10^-7 and we normalized the coordinate system to the barycenter of the Kerbin system.

Nice work! How did you pick the initial conditions for the orbits? Does Minmus\' orbit change noticeably if you use slightly different initial conditions?

[illectro]

The initial conditions were picked based on the values in the wiki, with the orbital velocity changed to account for both bodies mass.

And yes, minmus is very sensitive to initial conditions, I should do a plot that shows how the orbits vary over time if I move it by a few metres.