Years ago I wrote a program that could generate more realistic stars and planetary systems (ostensibly for the Traveller tabletop RPG but it evolved into something beyond that, which would work for any realistic scifi setting). I used published stellar evolution grids to calculate the star evolution and properties for the calculated age of the star, created and migrated jovians, and then created terrestrial worlds in the orbits that they could survive in (if they were within the jovians' "gravitational reach" then they'd be ejected from the system or turned into asteroid belts) and calculated their blackbody temperatures and determined their atmospheres. I've always wanted to try to import them into a visualisation program to see what they actually looked like (rather than as tables on a page) and recently realised that Space Engine could be a perfect way to do this. So here it is!

[center][/center]

The NST (New STar) Cluster is a 500x500x500 ly cube of stars, located 50,000 ly above the galactic plane of the Milky Way (it shouldn't clash with any existing real or proc-gen stars out there). The screenshot above shows the stars with a limiting magnitude of 15.00 so you can see most of them. Installation instructions are at the end of this post. 

The Cluster itself is consists of 244,589 solo star systems, containing a total of 638,318 planets and 22,046 asteroid belts (represented by a single asteroid here, adding more asteroids in the belts would have really killed the loading time!). There are white dwarfs, neutron stars and a stellar-mass Black Hole, and most other star classes and types are represented. My program can generate the stars for multiple systems (but not the planets) and I've figured out a way to add binary systems and may add more complex systems later on - however if I do then that would change all the existing planets (they have randomised elements filling some of the gaps in the conversion code that would be regenerated) - so be warned that subsequent updates will change the Cluster! Also there are no moons since the star solver doesn't generate them for imported planets (I may be able to figure out a way to add those later).

Here's a sample system from the text output:
There are seven worlds orbiting a 4.371 billion year old F7 V star in this system - four rocky planets, and three gas giants. The third and fourth planet are both ringed habitable worlds with water on their surfaces, and the fifth and sixth planets are gas giants that are more massive than Jupiter. And this is what the inner habitable planet looks like: 

[center][/center]

The SE Star Solver calculates some extra things like tidal heating and greenhouse effect (and rings) so the temperatures are sometimes different to what I calculated, which means there might be a few odd worlds (e.g. that have water on their surface but should be too hot to allow that to exist). I also added things like a hydrosphere and pre-set atmosphere for the habitable planets during the conversion process.

Have fun exploring the cluster, and let me know if you find anything interesting! And thanks to everyone here and on the discord who has helped me get this up and running!


INSTALLATION: To install the NST Cluster add-on, download this file:   NSTCluster Pak file
and then put that in your Steam\steamapps\common\SpaceEngine\addons\catalogs folder (which will be wherever steam is installed on your machine - you may need to create the addons/catalogs folder manually). If you want to try it on an older/non-steam version of SE then you can put it in your SpaceEngine\addons\catalogs folder - but be warned that I haven't tested it on older versions though. 

NOTE: The add-on takes 1-2 minutes to load on my machine, but it uses about 14 GB of RAM (I have a NVIDIA GeForce 980 Ti 6GB video card) so you will need a relatively beefy machine to run it. I've only tried this on the steam version of SE (0.990) but it seems to work there.

UPDATES: 
Version1: Solo stars only
11Feb2020 am: Updated the pak file to fix eccentricity and class errors.
11Feb2020 pm: Updated the pak file to fix ocean height errors. Now it only takes a minute or so to load!
12Feb2020 am: Updated the pak file to fix a few minor issues (implemented neutron stars/black hole properly, fixed some incorrectly zeroed ages). 

Version2: Binaries!
17Feb2020 pm: HUGE UPDATE - replaced pak file with a new version of the Cluster that now includes Binary Systems (the stars/orbital configuration generated by my code, with worlds placed by me but generated by SpaceEngine), life, and many other small tweaks. The download is now 45MB and requires about 14 GB RAM (still only takes a minute to load though). 

The classes "Ice Giant" and "Gas Giant" do not exist.
It's called "IceGiant" and "GasGiant". Without space.
Furthermore there is no class "DwarfPlanet".
A dwarf planet has "DwarfPlanet" as the object name. So: In principle you can set any class as 'Class', although it makes little sense to give the 'Class "GasGiant"' to a dwarf planet. You can also leave out the 'Class', SE will come up with one.

And there are thousands of such error messages in the 'se.log': By the way, you still have a lot of negative values for ocean depth. Of course, that doesn't work at all.
[hr]
To avoid the many (thousands) error messages in the 'se.log' file, you can put 'LogLevel 0' at the beginning of the script. Of course, other errors will also not be shown if they appear in your script.

There are also a lot of these errors in the script. This means that after the end of the 'Orbit' tag, a second 'Eccentricity' line appears. With a superfluous closing bracket.
This can also lead to problems.

Thanks - Sigh. Dammit, I thought I caught those. I've fixed them and reuploaded a new version that shouldn't have those Class and eccentricity errors now.

The ocean depth errors are deliberate though - I want varying depths of oceans and setting the Height to a randomised value between -1 and a value up to the BumpHeight seems to result in oceans that range from no ocean (for the negative values) to isolated spots of water (for heights < 1 km) to nearly global coverage (for heights close to the BumpHeight). It may give errors in SE but it seems to have the effect I intended. (EDIT: I suppose I could just randomise the bumpoffset instead and set the Ocean Height to equal that? That should have the same effect without the errors then?)

Also the error file seems to think that some of the star classes are wrong when they're not:  Which is nonsense since there's no way a K3 V could ever have an intensity of 10 Sols. So ignore those errors (they don't change the star sizes in the scripts from what they're intended to be).

It looks like SE uses new descriptors like Ferria, Carbonia and Aquaria that aren't described in the manual. Are those Planet Classes? Are they usable in scripts? Would it be better to use those here if they are? 

New planet classes are 'Terra, Titan, Aquaria, Jupiter, Neptune, Ferria, Carbonia'.
(Don't know if these are all new classes.)

The things with the spectral classes are not error messages, just hints.

yeah I'm not sure if there are any ferria or carbonia generated in the Cluster. Maybe because I specified that the worlds should be Selena, Terra, Oceania etc, or maybe because ferria and carbonia are determined internally some other way?

I've updated the file now so that it has no errors (beyond the stellar class suggestions). It now only takes a minute to load - I guess the errors must have been slowing that down, though it still takes up about 11.5 GB of RAM.

If you want some interesting things to look for: 

- there are some gas giants with VERY eccentric orbits around their primary star. Some aren't the only worlds in their system either.
- there are in fact a few neutron stars and a black hole in the Cluster - my system generator actually attempted to make them, but they were only partially made (I guess I hadn't completed the creation procedure in the code), but I've fixed them for the Cluster now!
- there's at least one very oblate star with some very close, entirely molten worlds!

Is this a fictional cluster you've created inside the Milky Way galaxy?  I want to install it, but I read about the loading time and it already takes me a minute to load SE with the 5 million star Gaia catalog added (and it uses 11.7 GB of memory.)  I want to install this though and check it out and compare it to the other clusters already in the program.

It's a fictional "cluster" (it's obviously a cube of stars rather than a realistic-looking cluster, but the point was to make the stars and planets with my code, then I figured out how to visualise them in SE afterwards). It's near the Milky Way, I put it above the disk so it'd be out of the way of everything else.

The object was to make realistic systems with my code - and as far as it goes it looks similar in its realism level to SE's own procedural generator, but I think there are some more interesting/different orbital configurations of planets generated in my code and I wanted to see what those actually looked like.

The loading time is significantly reduced now that I've got rid of the errors, though it still uses about 12GB of RAM. Maybe temporarily disable your other catalogs and see if you can load this one? 

Incidentally, I may do a short stream on my twitch channel this weekend to explore the cluster a bit, if anyone's interested in finding out more about how and why it was made and what's in it. I'll post details later!

It's a fictional "cluster" (it's obviously a cube of stars rather than a realistic-looking cluster, but the point was to make the stars and planets with my code, then I figured out how to visualise them in SE afterwards). It's near the Milky Way, I put it above the disk so it'd be out of the way of everything else.

The object was to make realistic systems with my code - and as far as it goes it looks similar in its realism level to SE's own procedural generator, but I think there are some more interesting/different orbital configurations of planets generated in my code and I wanted to see what those actually looked like.

The loading time is significantly reduced now that I've got rid of the errors, though it still uses about 12GB of RAM. Maybe temporarily disable your other catalogs and see if you can load this one? 

Thanks, I wonder if it has a midsized black hole at its core like some of the larger clusters are theorized to have!  I guess I will have to find out and see!  Is your final revision on the same link you posted before?

No, it's not got anything specific at its core, it's just a cubical volume of space populated by solo stars and their planets (no nebulae or multiple stars). Though there is one black hole in it somewhere now . 

The final version (at least of this incarnation - all fixed with no errors) is at the NST Clusters pak file link I posted in the original post. 

Thanks!  I'm excited about trying it out

Cheers! Check out my stream in the weekend if you want to learn more about how I made it and what's in it!