Space simulation

[Jacques]

Speculation

OK lets go!

It is a kind of pseudocode, but don't wory you won't need to code to understand what I mean.

Take [math]10^{100}[/math] identical sticky spheres with a 1 unit radius and 1 unit mass.

Place them randomly in a finite space, give them random speed from 0 to .999 c and random directions.

Defenitions:

-Sticky: When two spheres toutch each other:

--If the difference in speed is less than .001 c they will stick

--else (>.001c) it is an elastic collision.

-Momentum is conserved when 2 spheres stick together or bounced of each other.

 

 

Let run the simulation! We cannot run a program with [math]10^{100}[/math] on all the computers on earth, but if you use smaller numbers you will be able to see that grouping occur and look like the expansion of space: take a group for reference, you will see almost all the other group going away from you and the further they are the faster they will go.

No need of a bigbang with extreme density and temperature, to explain the hubble flow.

Does it make some sense for you ?

PS excuse my english I am doing my best

[iNow]

Does it make some sense for you ?

PS excuse my english I am doing my best

Pas de probleme, monsieur. Avez-vous des examples?

[timo]

Your english is fine, don't worry. I like the idea of such a little simulation but hope you're not getting too excited about having found the sage's stone of cosmology. There's several issues with your simulation, one being that I am pretty sure that you were using non-relativistic calculations and only restricted your initial conditions to be v<c in some frame of reference.

One thing you could try out: Does the clustering have any influence on the velocity distribution at all? You can get (using a non-relativistic model) get the result of relative velocity being proportional to distance very easily: Start all points (radius=0 here) at the same point in space with some random velocity and no colissions. This will trivially exactly reproduce Hubble's Law. I think your results come from these conditions (which are relatively similar to your program).

[Jacques]

This will trivially exactly reproduce Hubble's Law.

Yes I see, but there is no need to start all points at the same point. The points can be spreaded in large volume of space and after some time the Hubble flow will emerge.

Thanks for your answer. And yes I thought I was on something...

[timo]

As soon as the volume in which the particles are spread out is much larger than the area in which the particles were randomly distributed, the original position ditribution can be ignored.

What I find more interesting is the question to what extent switching on elastic collisions drives the effect. With collisions you should get an effective pressure from the inside to the outside (where there are no particles that could bounce you back). So I'd think that switching on the collisions would quicken the "Hubble's Law"-effect. It should even be possible to make thermodynamical predictions that could then be tested by your simulation.

[Jacques]

As soon as the volume in which the particles are spread out is much larger than the area in which the particles were randomly distributed, the original position ditribution can be ignored.

That is true without collisions. Collisions are necessary to create structures.

If the volume in which the particles are spread out is 0 then no collision will occure. I think that why there is an need for quantum fluctuation in the BB model