Project Orbiter Galaxy

[jedidia]

Got a bit sidetracked by Mass Effect 2, back on track now.

Database for evolutionary tracks up to 120 solar masses compiled, allthough only for one chemical composition. Let's not overdo this... I think I'll have enough bugs in the database as it is, but can't tell anything until I've implemented it.

 

[jedidia]

Database finally implemented, but evolution algorythm still pretty bug-ridden. Didn't expect anything else, really. At least I don't get that many neutron stars anymore, that's a good step in the rightdirection. But currently anything leaving the main sequence seems to be turning out a super-giant. Most probably got to do with a few wrong operations in there, digging through the mess as I write.

 

[jedidia]

sorted out most of the bugs, now the tons neutron stars are back... at least I'm getting some giants now, but not enough. After checking my Evolutionary tracks, it turns out that they only go up to the first flash in the AGB, so I'm currently skipping the whole AGB-phase of heavier mass stars. I should be able to correct that rather easy, but I'll also have some weeding out in the tracks-database to do. It's too detailed right now, increasing processing time by a rather unreasonable amount...

---------- Post added at 02:38 PM ---------- Previous post was at 10:40 AM ----------

Well, I'm coming along... the galaxy looks much better now, but still has some problems. Firstly, I have too many corpses lying around (majorly white dwarfs now, but TONS of them), and secondly, the giants are still a bit on the rare side. For now I'm concentrating my efforts on getting rid of the way too frequent dead stars. There are three possible causes remaining. Either I overestimated the overall age of the galaxy (working with 13 billion years currently, which I think should be ok according to several internet sources), or I underestimated the drop-off in age towards the rim (oldest stars in the rim currently around 6 billion years old, but I have no Idea if this is anywhere near accurate).

Or, third possibility, I'm getting too many massive stars, (since they die a lot faster). Theoretically I'm working with the initial mass function, though, and I already checked it several times and think it is ok, so I'll go about screwing around with the age parameters a bit more...

 

[Linguofreak]

sorted out most of the bugs, now the tons neutron stars are back... at least I'm getting some giants now, but not enough. After checking my Evolutionary tracks, it turns out that they only go up to the first flash in the AGB, so I'm currently skipping the whole AGB-phase of heavier mass stars. I should be able to correct that rather easy, but I'll also have some weeding out in the tracks-database to do. It's too detailed right now, increasing processing time by a rather unreasonable amount...

---------- Post added at 02:38 PM ---------- Previous post was at 10:40 AM ----------

Well, I'm coming along... the galaxy looks much better now, but still has some problems. Firstly, I have too many corpses lying around (majorly white dwarfs now, but TONS of them), and secondly, the giants are still a bit on the rare side. For now I'm concentrating my efforts on getting rid of the way too frequent dead stars. There are three possible causes remaining. Either I overestimated the overall age of the galaxy (working with 13 billion years currently, which I think should be ok according to several internet sources), or I underestimated the drop-off in age towards the rim (oldest stars in the rim currently around 6 billion years old, but I have no Idea if this is anywhere near accurate).

Or, third possibility, I'm getting too many massive stars, (since they die a lot faster). Theoretically I'm working with the initial mass function, though, and I already checked it several times and think it is ok, so I'll go about screwing around with the age parameters a bit more...

You may have overestimated the early formation rate and underestimated the late formation rate.

You may also be underestimating your corpse densities, I don't know where your getting them from. If you're getting them straight from observational data, there are strong observational biases. White dwarves are dim, you won't see them from very far away unless they have a companion that they affect graviationally. The same with neutron stars: They aren't very easily detectable. We've detected as many as we have only because they tend to go through a pulsar phase early in their lives where they send out very bright beams of radiation, but this is only in certain directions, and only lasts for 10-100 million years. According to Wikipedia (to be taken with the usual grain of salt), around 99% of neutron stars are estimate not to be pulsars anymore. There's also likely alot of Neutron stars we haven't detected because they aren't radiating in the right direction.

 

[jedidia]

You may also be underestimating your corpse densities, I don't know where your getting them from.

Well, it's observation of what shows up in the generator. Currently it seems that even in the rim, where there stars are younger, I'm getting around 10% white dwarfs. The luminosity class of the star is calculated by setting an initial mass, derived from the initial mass function, and an age, depending on the position in the galaxy as well as some randomness. The star is then evolved according to its age, and stars that have outlived their approximate AGB phase turn into white dwarfs, neutron stars or black holes, according to their initial mass.

You may have overestimated the early formation rate and underestimated the late formation rate.

Most likely the case, since I never heard of either of them :lol:

None the less, I have kind of an early/late formation algorythm, which currently produces 90% of the stars in an early formation (within ten percent of the perceived maximum age at this point of the galaxy) and 10% of the star at a random age between that and 1 million years. So, if early formation rate is lower, I'm getting a too high average age. Will have to read up on this a bit.

Also, I don't quite know how the age drop-off in the galaxy really is. The core obviously formed first, and I assumed that stars in the arms formed gradually after this, the further from the core the younger. Is that right, or did all the disc form at aproximately the same time after the core?

Edit:

Just had a very interesting read... It seems like we won't know for sure until we get comprehensive data from Gaia, but it seems like a) the thin disc formed from the inside out (so for so well for my model) but b) several billion years AFTER the core. So my average stelar age in the disc is indeed too high. Alas there's not really conclusive evidence yet on when exactly it actually formed (the process is said to have taken about 7 billion years according to the two-infall theory and is still ongoing on the rim). They're waiting for Gaia to prove them right (or wrong), but I can't wait that long, so I'll just go with that. It will certainly reduce the average age in my disc by a comfortable amount...

 

[Linguofreak]

Well, it's observation of what shows up in the generator. Currently it seems that even in the rim, where there stars are younger, I'm getting around 10% white dwarfs. The luminosity class of the star is calculated by setting an initial mass, derived from the initial mass function, and an age, depending on the position in the galaxy as well as some randomness. The star is then evolved according to its age, and stars that have outlived their approximate AGB phase turn into white dwarfs, neutron stars or black holes, according to their initial mass.

Well, I meant it may be that you're expectations of how many corpses (more formally "(stellar) remnants") you'll find may be an underestimate. What are you basing your expectations on? Gut feeling? Raw observational data? Observational data corrected for luminosity and distance?

Most likely the case, since I never heard of either of them :lol:

I wasn't using this as a technical term, so you wouldn't have heard of either. I was just talking "formation that happened earlier" vs. "formation that happened later".

None the less, I have kind of an early/late formation algorythm, which currently produces 90% of the stars in an early formation (within ten percent of the perceived maximum age at this point of the galaxy) and 10% of the star at a random age between that and 1 million years. So, if early formation rate is lower, I'm getting a too high average age. Will have to read up on this a bit.

Yeah. That sounds a bit extreme, at least for the thin disk. For the bulge and halo it may actually have to much recent formation.

Also, I don't quite know how the age drop-off in the galaxy really is. The core obviously formed first, and I assumed that stars in the arms formed gradually after this, the further from the core the younger. Is that right, or did all the disc form at aproximately the same time after the core?

Well, how early/late a given part of the galaxy formed (and how much star formation in currently occuring there) actually has more to do with how much the orbits in that part of the galaxy all lie in one plane (the plane of the disk) than anything else.

The diagram in this article might help you: The oldest parts are the bulge and halo, the youngest part is the thin disk, and the thick disk is somewhere in between.

http://www.daviddarling.info/encyclopedia/D/disk_star.html

The older parts had a burst of star formation at the beginning that ended quickly.

The younger parts were technically part of the initial burst, but peaked later and higher and fell off less rapidly.

It's also best not to think of galactic regions in terms of locations, but in terms of orbital parameter regimes.

Halo and bulge stars can have any inclination or eccentricity, disk stars have low inclinations and eccentricities.

The disks of spiral galaxies are blue because most of the light is coming from (very) recently formed stars. The bulges are red/yellow because most of the light is coming from older stars.

 

[jedidia]

Well, I meant it may be that you're expectations of how many corpses (more formally "(stellar) remnants") you'll find may be an underestimate. What are you basing your expectations on? Gut feeling? Raw observational data? Observational data corrected for luminosity and distance?

Well, not much more than how much white dwarfs we have in the solar vicinity, but that estimate might be a bit too low. Other than that, it's just gut-feeling...

Yeah. That sounds a bit extreme, at least for the thin disk. For the bulge and halo it may actually have to much recent formation.

actually, I don't have the halo in yet... still thinking about wheather it's worth adding or not...

The younger parts were technically part of the initial burst, but peaked later and higher and fell off less rapidly.

Well, I'll just tune around on those parameters until I find something agreeable, then. Thanks.

 

[Linguofreak]

Well, not much more than how much white dwarfs we have in the solar vicinity, but that estimate might be a bit too low. Other than that, it's just gut-feeling...



actually, I don't have the halo in yet... still thinking about wheather it's worth adding or not...

Should be fairly simple. It' should have roughly the same age and composition as the core, just less density.

 

[jedidia]

screwed around another bit with the age parameters, corrected the duration of the AGB phases a bit, I'm getting quite nice results now. The class IV stars are maybe a bit on the rare side...

However, none of my giants seems to turn out RED. most seem to be in the G and F range. I think that's because I don't have any data about the AGB-phase, so the radiuses are currently not increasing during that phase. Will probably add some ball-park estimate just to bring them into the K- and M-range, since I don't really want to simulate the merry jumping around they do in reality.

Anyways, work has gone well this week, unfortunately the next buisness trip starts tomorrow already, so it'll have to wait a bit more. As soon as I get those giants into the red I'll add the thick disc and the halo, and then it's StarGen-revision time again...

 

[eveningsky339]

Here's a little something...

Have you ever heard of Noctis IV?

http://en.wikipedia.org/wiki/Noctis

It's a 2000-era DOS game that generated an entire galaxy (~78 billion stars). The program weighs in at just under 1 MB, because the conent is generated on the fly.

I heard somewhere that the source code was released, but I can't seem to find it. It also my very well prove useless considering the decade-old DOS platform.

 

[jedidia]

yeah, I played a bit of Noctis back in the time...

I very much doubt however that the star-generation code would fit well into what I already have now, which is neaarly finished... some tweaking to do still, especially with the AGB-phase of giants as well as the systems, but I'm almost there. At least I hope. I thought I was almost there 3 months ago...

 

[eveningsky339]

yeah, I played a bit of Noctis back in the time...

I very much doubt however that the star-generation code would fit well into what I already have now, which is neaarly finished... some tweaking to do still, especially with the AGB-phase of giants as well as the systems, but I'm almost there. At least I hope. I thought I was almost there 3 months ago...

Haha, things never really go according to plan in this life. At least you're making progress! This was a significant undertaking. :tiphat:

 

[jedidia]

I thought a little update might be in order here...

I weeded out my evolution tracks database to make the whole thing a bit faster. Doing so I also convinced myself that the final temperatures for the beginning of the AGB-phase were all in the M and K spectra, so the problem why my Giants didn't turn red couldn't be caused by the data.

After looking at my code a bit I noticed that I forgott to type a rather important line... I only adjusted the spectral fraction of my stars, but not the spectral class itself. Corrected that, and all my Giants are turning out nice and red now. This last step pretty much concludes the stellar evolution model, except a few tweaks here and there. Next focus is on the evolution of the planetary system itself once the major body starts loosing mass and heating up. This should be a rather interesting experiment within the stargen code...

(the reason why it took me so long for this rather small update was a) work and b) the close advent of the low thrust trajectory planner... I just have to get a usable eject calculator working for that low thrust engines...)

 

[Bloodworth]

Dude, keep up the good work, this is a monumental undertaking and I can't wait to see the result

Why Low thrust? On a galactic scale, anything below ftl capability and you are talking about generation ships (which I haven't seen any of in the orbit hangar).

 

[Izack]

Dude, keep up the good work, this is a monumental undertaking and I can't wait to see the result

Why Low thrust? On a galactic scale, anything below ftl capability and you are talking about generation ships (which I haven't seen any of in the orbit hangar).

Those engines would have also have a higher ISP (specific impulse). :tiphat:

 

[Bloodworth]

I might be able to make it in the flamberge, I have the thrust of the gods and can hit relativistic speeds in relatively short order, but low thrust?

 

[jedidia]

engines with low thrust, but high ISP might take a bit longer to get up to speed, but their amount of Delta-V is magnitudes higher than that of engines with higher thrust but less ISP.

Engines combining both high thrust and high ISP have a serious heat problem, because of the total energy they release (read: the engine would melt itself to slack during the first seconds unless you have some incredibly efficient way of getting rid of the heat).

Currently, most advanced engine concepts (except Orion) are low thrust (e.g. no one really expects to ever be able to get much more than 60000 newtons of thrust out of any fusion drive, unless you build the core incredibly large to be able to absorb more heat).
My interest in seeing these engines fly in orbiter is not really linkied with Orbiter Galaxy, I just want to fly these beasties (Jupiter in six months with an engine that lies within realistic expectations is just a thrill to have...)

 

[jedidia]

I found another evening to code on... Evolution of the solar systems seems to come along nicely, yet also has left some questions.

currently, when The star goes into the giant phase and turns up its luminosity drastically, many of the gas giants get stripped naked (i.e. their layers get blown away and only the core remains). I know Stargen well enough by now, so I fully expected that. Yet I'm wondering how realistic that is... First, I can perfectly understand why a Gas Giant can't FORM when it's too hot, but orbital migration shows clearly that once they're formed, they can travel into regions where they couldn't have formed. The same should also apply to the situation when the star gets hotter, on the other hand it starts to throw stuff around and produces some serious solar winds that might be very disruptive for gas giants...

What would you say: When the Sun enters the AGB phase, will Jupiter be stripped naked? if yes, I can comfortably leave the current model standing, if no, I'll have some rewriting to do.

Also, even IF Jupiter were stripped, I'd expect the remaining core to be highly compressed, and therefore smaller in radius than a normally formed planet of that mass would be. Stargen of course doesn't calculate this yet, and I wonder how I would best go about making an estimate?

 

[jedidia]

#@}$* pointers!! :hunter:

3 hours of wild goose chase. At least it seems to have fixed pretty much any bug that was still persistant, and one hell of a memory leak too (practically a third of all my memory was just flowing right out of a*se of the code, but I knew that. The trouble was finding the pointers that pointed right into nirvana, effectively preventing me from deleting all those pointer-chains savely.)

I'm still not quite decided yet what to do about the gas giants. Decided to let the matter rest and calculate the new orbits first, but I happened upon a slight problem... when the central body looses mass gradually over time, I would expect the semimajor axis of the orbit to grow gradually, without much significant change in the excentricity, right?

The problem is just, how do I calculate the increase in the semimajor axis without running the thing through iterations? It's slow enough already, if I run every planet of a giant through several iterations that won't help the matter... any suggenstions? (formulae prefered :lol

 

[jedidia]

Evolution of stellar systems concluded for now (I'm pretty certain I'll have to work on that a bit more once a demo is out and people can look at the stuff...)

Next up is backwards-evolution of catalogue stars outside the main sequence, to get some believable systems created for them. I got everything I need for that, but I've not decided yet what would be the most efficient way to do it code-wise.

After that it's finishing up the interface, and the demo of the generator will be ready. I *might* just make it until the end of summer vacation, but since I DON'T have any vacation in that time, it's highly uncertain. There shouldn't be any significant backdrops anymore, but then again, they always come completely unexpected... :lol: