Tell me more ×
Physics Stack Exchange is a question and answer site for active researchers, academics and students of physics. It's 100% free, no registration required.
up vote 4 down vote favorite

Is there any general purpose stellar evolution simulation engine or software? Something to throw in properties of the star and to watch how (and why) they change along the timeline - with or without visualization (but preferably with).

The best of what I have found on the net is this site:

http://outreach.atnf.csiro.au/education/senior/astrophysics/stellarevolution_links.html

however, there is no good simulation software behind the links there (or at least I don't see one).

share|improve this question
If you want to do research-grade stellar evolution, you're going to need a supercomputer. I don't know if there are fast, crude stellar evolvers out there. Could be. – Andrew Oct 19 '11 at 1:23
A typical laptop today can do a significant fraction of what a "supercomputer" did in 1990, so it is clear that there should be some interesting codes out there. Whether anyone is going to e willing to help you get it running on your very different environment is another matter entirely. – dmckee♦ Oct 19 '11 at 2:10
@dmckee - well, I don't mind porting (or at least trying to port) software to PC environment if there is any that's in the free access. That might be an interesting task to do, actually, but I can't find any open source (or accessible via NDA) stellar evolution simulation software on the net. I'm starting to doubt there is one. – jtootf Oct 19 '11 at 7:23

1 Answer

up vote 11 down vote accepted

I'll chip in here because I'm a research student and I work with a stellar evolution code (the Cambridge STARS code) more-or-less daily. Regarding some of the comments to the question, stellar evolution is actually quite fast, depending what code you use. Certainly, it isn't like hydrodynamics or N-body simulations like those used in galaxy formation/evolution and large-scale structure simulations. I can simulate a Sun-like star reasonably accurately from before the main sequence to core He-flash in probably less than 10 minutes with a fairly standard machine (single 3GHz Intel core, about 400MB RAM is used). Here are two links for software pitched at distinct levels.

Richard Townsend of the University of Delaware created this webpage which allows you to specify initial conditions for your star and then emails you the results. I have never tried it but I believe it's based on the same code I (and many others) currently use. The output is ASCII text so you can plot it with whatever. I think the author provides some IDL scripts but its proprietary software. You could use, say, gnuplot instead.

If you want a more advanced code, your best bet is probably MESA purely because it's easiest to access. MESA is a fully-fledged research-grade code. It's quite big (code with input data is about 3GB) but very robust. I recommend going through the materials on the website if you're interested in getting it running. It's mostly maintained by Bill Paxton and makes use of much more software engineering practice than other codes. It also comes with its own visualization plugins for pgplot, if you have it.

Hope that's something like what you were looking for!

share|improve this answer
Thanks a lot! MESA seems to be exactly the thing I was looking for. – jtootf Oct 19 '11 at 9:23
I was coming at this question from the supernova simulation side, so I figured stellar evolution couldn't be that much faster. Do you have to do radiation transport for stellar? That's what eats all the time in supernovae. – Andrew Oct 19 '11 at 10:51
1  
@Andrew: In stellar codes, stars are presumed to be in local thermodynamic equilibrium. So, at each radius, the radiation field is a blackbody at the local temperature. The assumption is pretty solid until you get to the atmosphere, which is usually taken as $\tau=2/3$. From there, you start needing a proper atmosphere model to find spectral lines etc. As you rightly expect, that's a harder problem and takes much longer to calculate. But at least the 1D codes should give you a pretty accurate effective temperature. – Warrick Oct 19 '11 at 11:00
@Andrew: Also, convection is treated as a diffusive mixing process. You can look up "mixing length theory" to get the details. It says heat is transported by blobs moving over a chararcteristic "mixing length" at a related "mixing velocity". I like to think of it as a really big lava lamp. MLT is quite crude but it works sufficiently well that no-one has made any big improvements yet. – Warrick Oct 19 '11 at 11:04
Yeah, my old Stellar Astrophysics professor dumped on mixing length theory, but like you said, he didn't have anything better. Supernovae also really need full 3-d to capture anything remotely realistic. – Andrew Oct 19 '11 at 20:22
show 1 more comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.