Simulating globular clusters over blackholes with Fortran

[Benzoate]

I suspect that the runtime simulation for a star systems containing thousands of stars , many of those stars whose mass is way beyond the solar mass to be very very long. Is it possible and cost-friendly(i.e, not expensive) to simulate a globular star system cluster over a supermassive black hole on Fortran? This is the topic I am playing around with ; I am also thinking of maybe just simulating a binary redgiant star system over a black hole . But I haven't really seen many papers on Red giant star that orbit around black holes. I am in the physics department and so have access to the appropriate computers to run such simulations.

 

[oedipa maas]

I'm not an astronomer - can you first explain why simulating a globular star cluster interacting with a supermassive black hole is interesting? Is there an idea you want to demonstrate or test?

To determine whether the calculation is feasible you first need to decided how much detail is necessary in your model. Can you treat the stars in the cluster as point particles which interact only through gravity?

If you treat the stars as point particles interacting with each other through gravity how does the number of operations required to calculate the forces on each particle scale with the number of stars in the system?

 

[Vanadium 50]

1) I don't see why having a heavy star takes any longer to calculate than a light star. The equations are the same, only the numbers are different.

2) Globular clusters do not have supermassive black holes in their centers. There is some evidence that some (okay, two) may have intermediate mass black holes, but this evidence is far from unambiguous.

3) Globular clusters have hundreds of thousands of stars, not thousands.

4) So far as I know, no simulation tries to calculate the dynamics of 10^5 objects by looking at all 10^10 two-body interactions. This is prohibitively expensive. They either use fewer stars, potential approximations, or a mix of the two.

5) I have a hard time with "I have access to the appropriate computers", because if you did, you would likely have mentioned them by name. (e.g. GRAPE-6) I don't think you quite understand what hardware this sort of calculation entails.

 

[Benzoate]

1) I don't see why having a heavy star takes any longer to calculate than a light star. The equations are the same, only the numbers are different.

2) Globular clusters do not have supermassive black holes in their centers. There is some evidence that some (okay, two) may have intermediate mass black holes, but this evidence is far from unambiguous.

But most galaxies contained at least one supermassive black hole at there very center. galaxies contained globular clusters.

3) Globular clusters have hundreds of thousands of stars, not thousands.

You were right and I was wrong. There is nothing more to say on that issue.

4) So far as I know, no simulation tries to calculate the dynamics of 10^5 objects by looking at all 10^10 two-body interactions. This is prohibitively expensive. They either use fewer stars, potential approximations, or a mix of the two.

So how would physicists confirmed that there theory on globular clusters orienting themselves in a certain manner if whether or not there theory is correct or incorrect if they do not have the experimental tools to simulate their model of how globular clusters move?

5) I have a hard time with "I have access to the appropriate computers", because if you did, you would likely have mentioned them by name. (e.g. GRAPE-6) I don't think you quite understand what hardware this sort of calculation entails.

You are correct. I really do not know much about computer modeling. I'm only a sophomore in physics trying to find an interesting simulation of a star system , that is within my reach.

 

[leon1127]

from my experience, formulate the system of differnetial equation isn't the hardest part. Perhaps you can start on that. After that you can find some numerical method to solve that system. After all that should just become a giant system of differential equations.

After you formulate it, try to explain why it might become difficult.

 

[Wallace]

What is the purpose of this project? If you are interested in the physics then I would advise against re-inventing the wheel. There are many many freely available computer codes to run N-Body simulations. I use http://www.mpa-garching.mpg.de/galform/gadget/index.shtml" [Broken].

If you are interested in the process of the simulation, then writing your own code is an interesting project. It just depends on what your focus is.

 

[Vanadium 50]

Your replies ended up in the qyote.

But most galaxies contained at least one supermassive black hole at there very center. galaxies contained globular clusters.

This is not a reason to conclude supermassive black holes are in the center of globular clusters. Kansas City is at the center of the US. The US contains other states. Therefore Kansas City is in the center of Massachusetts?

Supermassive black holes weigh as much as or more than an entire globular cluster.

So how would physicists confirmed that there theory on globular clusters orienting themselves in a certain manner if whether or not there theory is correct or incorrect if they do not have the experimental tools to simulate their model of how globular clusters move?

An excellent question, and one you should ask the faculty at your university. They can explain in great detail all the simplifications and approximations they use.