Welcome, Guest

Author Topic: Solar system gravity  (Read 4772 times)

terzer

  • *
  • Posts: 5
Solar system gravity
« on: December 04, 2010, 04:45:10 PM »
Hello,
yesterday l found this program, its amazing! Just wonderful!
I started playing and l noticed few things:
When l try to create a solar system (just add one sun and add rings with 100000 particles in 10-20 AU, simulating accretion disc), the rings just expand outwards forever, there are no planets forming. l checked the collision mode, also lagrange points, accuracy mode, hill spheres... l waited several million years, the solar system is now 5ly wide (only rings). Its not realistic if the standard theory for solar system formation is true. The colliding particles should create bigger/heavier bodies, which eventually will form asteroids, then planets and should "clear" the sol. system.
I also noticed this, when l try to merge (collide) galaxies. After millions/billions of years, there is no single galaxy, the particles are expanding outwards...  
Is there a dark matter and dark energy simulation, or the software physics are just based on Newtonian physics?
l'm l missing something?
Thanks!

Naru523

  • Universe Sandbox 1 Beta Team
  • *****
  • Posts: 1295
  • let's walk the true path of life
Re: Solar system gravity
« Reply #1 on: December 04, 2010, 04:59:06 PM »
Dust doesn't collide. You'll have to add lots of asteroids, but it'll slow down your computer.
For galaxy collision: Galaxies do not collide (black holes) yet for now, however Dan Dixon says he will make it collide in the near future.

Currently there's no dark matter or energy physics/simulations in this. It's only base on Newtonian physics.

terzer

  • *
  • Posts: 5
Re: Solar system gravity
« Reply #2 on: December 04, 2010, 05:11:25 PM »
Thanks!
The dust should collide. Its still matter, it cannot escape gravity.
Wonderful software!

edit: l believe l have found the problem:
When l accelerate the time lapse, the orbits become unstable. For example if l open Earth and moon simulation and then accelerate the time lapse, in few years the moon escapes the Earth orbit...
« Last Edit: December 04, 2010, 06:01:54 PM by terzer »

Bla

  • Global Moderator
  • *****
  • Posts: 1013
  • The stars died so you can live.
Re: Solar system gravity
« Reply #3 on: December 05, 2010, 02:49:30 AM »
The problem with dust is just that if you add thousands of particles, if they all have mass it'll probably take forever to calculate, and slow the simulation down a lot (imagine replacing the 100.000 particles with 100.000 bodies). I agree there should be an option to select whether particles should have mass or not, if you use particles in a protoplanetary disc you might want them to have mass, but if they're a ring orbiting a planet in a star system, the mass might just slow down the simulation.
Also, I think particles don't have any radius right now, they're just particles which scale depending on how far away you view them, so they wouldn't be able to collide either.
edit: l believe l have found the problem:
When l accelerate the time lapse, the orbits become unstable. For example if l open Earth and moon simulation and then accelerate the time lapse, in few years the moon escapes the Earth orbit...
This is true, the problem will probably always be there though, and the only solution is a faster CPU, which can calculate each time step faster, so the simulation will run faster on slower timesteps. The time step is how often the system is calculated, so if the time step is too big compared to the time it takes for a body to orbit another (like, if it's 1/8 of the time it takes the Moon to orbit Earth, it will only calculate the position of the Moon 8 times during the orbit, which might not be enough to sustain the orbit, or at least would make it very unstable).
« Last Edit: December 05, 2010, 02:53:49 AM by Bla »

MrLawbreaker

  • *
  • Posts: 3
Re: Solar system gravity
« Reply #4 on: December 05, 2010, 06:50:11 AM »
This is true, the problem will probably always be there though, and the only solution is a faster CPU, which can calculate each time step faster, so the simulation will run faster on slower timesteps. The time step is how often the system is calculated, so if the time step is too big compared to the time it takes for a body to orbit another (like, if it's 1/8 of the time it takes the Moon to orbit Earth, it will only calculate the position of the Moon 8 times during the orbit, which might not be enough to sustain the orbit, or at least would make it very unstable).

How does it come that if my simulation is stuttering my CPU is only at maybe 60%.
I got a Dual Core with 2x3.0GHz.
Is it the programm that does not want to use the other 40%?

Bla

  • Global Moderator
  • *****
  • Posts: 1013
  • The stars died so you can live.
Re: Solar system gravity
« Reply #5 on: December 05, 2010, 07:23:37 AM »
How does it come that if my simulation is stuttering my CPU is only at maybe 60%.
I got a Dual Core with 2x3.0GHz.
Is it the programm that does not want to use the other 40%?
Exactly because it's dual core I think. It has two cores, but I think Universe Sandbox can only use one core, and that it'll be able to use multiple cores when .NET 4 is released (or whatever it was).

terzer

  • *
  • Posts: 5
Re: Solar system gravity
« Reply #6 on: December 05, 2010, 11:17:29 AM »
Thanks, it makes sense. However this could be a problem with galaxy (or cluster) colliding. The step must be in million years in order to have noticeable movement...

Chuck14

  • ***
  • Posts: 23
Re: Solar system gravity
« Reply #7 on: December 05, 2010, 12:47:05 PM »
what is the game called?