Blue Venus Parasol & Reflector (Mega Structures)

[MKruer]

Preface: Blue Venus Parasol & Reflector (Mega Structures)

The idea for this project came from numerous discussions regarding a Blue Venus, and the best methods for creating another habitable world. This project focuses on the mega structure(s) that would be required to make Venus habitable using today?€™s theories and technologies. It I my hope that once this project is completed that others will revisit other Blue Venus projects.


Note: The Reflector at L2 is one ring object. For purposes of demonstration to show the planetary shadow it appears to be two objects.

The Structures (So far)

Venus Solar Deflector (14,683 100k radius hex tiles)
Diameter of ~25430
Average distance from Venus 1004392km (L1)


Venus Solar Reflector (17,664 100k radius hex tiles)
Inner diameter of ~25840km
Outer diameter ~38100km
Average distance from Venus 1004392km (L2)


References
Release: Blue Venus [v2]
Blue Venus revisited (large graphics inside)
UPDATE!! - Ultimate Terraformed Mars Texture

Blue Venus Parasol & Reflector (Mega Structures) Download

[rthorvald]

Let me know if you can help.

MKruer,
i am unsure what you want me to do, or how it relates to this subject.

I??m happy to help, but what exactly are you asking?

- rthorvald

[MKruer]

Let me back track. I am looking for help getting the Venus Solar Deflector and Reflector working in Celestia. Right now the model I have, for as simple as the design it is, is to complex to be rendered. This might be a limitation on of the 3ds format that I am using or Celestia not being able to handle one million verts over a 30,000km area, but either way it just does not show up. So I need a way to simplify everything to make it work, at the same time I would like to up the detail. This is where you might be able to help me out.

[rthorvald]

Let me back track. I am looking for help getting the Venus Solar Deflector and Reflector working in Celestia. Right now the model I have, for as simple as the design it is, is to complex to be rendered. This might be a limitation on of the 3ds format that I am using or Celestia not being able to handle one million verts over a 30,000km area, but either way it just does not show up. So I need a way to simplify everything to make it work, at the same time I would like to up the detail.

Well, first off, you should open a separate thread for this instead of hiding it in this thread. Else, chances are only those interested in a Island One will see it, which will limit any support you might get from the rest of the forum, and besides, we avoid confusing people with mixing two highly separate subjects.

In that thread, please post your data on the 3ds file, together with your SSC code so one can start by determining why your model does not show up. Then we can proceed with any other challenges.

Selden, if you are reading, perhaps you could split the thread?

- rthorvald

[selden]

split.

The most common cause of non-visibility of a model is a typo somewhere. A spelling error in the name of the Mesh will do it, for example. Check the contents of Celestia's "console log" for error messages. (Type a tilde to see it.) Size and complexity have no effect other than rendering time. Modern graphics drivers page graphics memory to and from main memory. Some 3D design programs limit the number of vertices in an individual mesh within a 3DS model file to 32K vertices, but you should get an error message from the design program when it fails to export the model.

An unexpected cause is making the albedo of the reference object too dim. If the albedo of the primary is less than some tiny value, its "children" are not drawn and no error messages are generated. (I consider this a bug and have reported it as such: the albedo of their primary ought to have no effect whatsoever on the visibility of the objects orbiting around it.)

[MKruer]

Selden, Thanks for the split. I have updated the first post with a link to the model as well as updated some information.

If you have time, you might be able to help me with a project that I can?€™t seem to get working. The idea is for a Blue Venus. In order to make this happen I need two mega structures. The first is a Venus Solar Deflector placed just outside the L1 point. The second one is a Solar Reflector Placed just inside the L2 point. The combination between radiation pressure and gravitational attraction would keep the structures in place. In addition the structure would need to act as a miniblind opening and closing them to block and reflect light, giving Venus a 24hr night and day planet wide cycle.

I need to redo the object, the tile size should be 100km in diameter not 200km, Each one of these blocks should be constructed of 1k tiles, and each of those titles 10m tiles, also the thickness of the smaller tile is paper thin, the super structure might be up to a km in thickness, and to get even more technical the entire object would have a slight cure to it of 1004392km in radius

let me restart off simple with a simple tile, but i am not sure if I will make the right. In the end this LOD will probably just become a bitmap on the larger structure.

Here is what I have so far, but I need a way to reduce the mesh greatly.

Blue Venus Parasol & Reflector (Mega Structures) Download

[selden]

Your BlueVenus parasol download works fine for me with Celestia v1.5.0pre4 after I put its files in the correct directory structure.

\Addons\BlueVenus\bluevenus.ssc
\Addons\BlueVenus\models\L1_Venus_Parasol.3ds
\Addons\BlueVenus\models\L2_Venus_Reflector.3ds


The objects aren't actually in the correct orbital positions, but that's just a matter of creating the correct orbital definitions in the SSC file.

They also cause Chris' new smooth orbit path code to crash Celestia built from cvs, but that's a separate issue.

This URL displays both the parasol and the reflector when using Celestia v1.5.0pre4 on my system:

Venus and statites



System config:
1GB 3.4GHz Pentium 4-550, WinXP Pro Sp2
256MB GF 7800 GTX, ForceWare v163.71
Celestia v1.5.0pre4

[MKruer]

Must have missed that memo to place the models in the models directory. Anyway there is room for lots of optimization. I am only getting 2-4 fps on my Nvidia 8800 GTS

Just out of curiosity do you know how to calculate the correct orbit? The only thing i got was the correct orientation of the object and the rotational speed.

cel://SyncOrbit/Sol:Venus/2008-01-11T01 ... 2054&ver=2

Code: Select all

"Venus Parasol (L1)" "Sol/Venus"
{
   Mesh   "L1_Venus_Parasol.3ds"
   Radius   12715

   EllipticalOrbit {
      Period           -224.70069
      SemiMajorAxis    1004392
      Eccentricity     0
      Inclination        0
      AscendingNode    0
      LongOfPericenter 0
       MeanLongitude    0
   }
   
   UniformRotation {
      Inclination      0
      MeridianAngle    90
      AscendingNode    0
   }
}


[selden]

You might take a look at the ssc that Grant Hutchison created for a Venus parasol.

See http://www.lepp.cornell.edu/~seb/celest ... rasol.html

[rthorvald]

Mkruer,
i am unable to open your 3ds file in my 3d editor (it crashes on both models). Perhaps you should try to export it again.

- rthorvald

[ElChristou]

Also unable to open the files...

Now looking at your screenshots, I'm wondering if doing the structure of each tile is a good idea; it's very nice ok, but I doubt people will spend their time looking closely each tiles one by one...

Using a 200ko model (instead of 6,3Mo!) + a nice 1k texture tiled on the whole model would do the job, no?

[Cham]

MKruer, your models are crashing my modeler. The meshes are buggy, and appear to be way too big (around 7 MB !) for what they are.

[chris]

Also unable to open the files...

Now looking at your screenshots, I'm wondering if doing the structure of each tile is a good idea; it's very nice ok, but I doubt people will spend their time looking closely each tiles one by one...

Using a 200ko model (instead of 6,3Mo!) + a nice 1k texture tiled on the whole model would do the job, no?

Exactly! Having geometry for each of those tiles is overkill . . . Still, I'm surprised at the low performance you're seeing. I get ~13fps with a GeForce 7900 GTX; haven't tried the 8800 yet.

--Chris

[MKruer]

Selden, That certainly is a way to do it. I am still hoping that one day celestia will just include an option to use the LaGrange points instead of this work around. (Insert shameless plug http://celestiaproject.net/forum/viewtopic.php?t=8640)

Rthorvald, I can export it into just about any format, would there be a better format then 3ds? Also its probably not that the model is bad, it just sucks up so many resources. Every modeler progrma that i use the models consumes upwards of a few hundred megabytes and into gigabytes of memory.

[rthorvald]

I can export it into just about any format, would there be a better format then 3ds? Also its probably not that the model is bad, it just sucks up so many resources. Every modeler progrma that i use the models consumes upwards of a few hundred megabytes

I don??t know; i am often working on extremely large 3ds files, and they work quite well.

Since nobody else here can open your files either, i guess there is something with your export method OR your files.

What is the native fileformat of your modeller? Maybe we can open that...

- rthorvald

[MKruer]

ElChristou Yes making a simpler model is the general idea, and it will probably be the only way to get it working in the immediate future, but to create a real texture that fits the criteria, I will have to make a detailed model of one tile and render it first.

Cham Working on it the meshes are 3d, its very big but only 1km thick, just adding the thickness increased the space required by greater then a factor of eight. The final version will probably end up being be a 2d image

Chris I think something was running in the back ground or I had a lot more stuffed turned on. I am getting 12-14FPS now.

[MKruer]

I can export it into just about any format, would there be a better format then 3ds? Also its probably not that the model is bad, it just sucks up so many resources. Every modeler progrma that i use the models consumes upwards of a few hundred megabytes

I don??t know; i am often working on extremely large 3ds files, and they work quite well.

Since nobody else here can open your files either, i guess there is something with your export method OR your files.

What is the native fileformat of your modeller? Maybe we can open that...

- rthorvald

I am using AutoCad 2008 (dwg, dwf) create the object, but for some stupid reason they took out the export to 3ds option, (they allow only import now) so I am using a program called Deep Exploration open them and export it to 3ds.

[selden]

MKruer,

In Celestia v1.5.0, defining bodies to be at fixed positions relative to a planet's position is quite easy.

Celestia v1.5.0 also allows one to define ScriptedOrbit functions which can, in principle, define the much more complicated orbits that bodies actually follow around the Lagrange points.

Someone with more knowledge than I have of the math that's involved would have to do that, however.

[rthorvald]

I am using AutoCad 2008 (dwg, dwf) create the object, but for some stupid reason they took out the export to 3ds option, (they allow only import now) so I am using a program called Deep Exploration open them and export it to 3ds

Perhaps that is the problem?

My modeller understands jas, mtl, 3dmf, rib, h, obj, fbx, dfx, 3ds, stl and sia. Maybe you are able to export to one of these.

- rthorvald

[Fenerit]

From your description seem that you use the 3d format (3ds) without to work in a pure 3d ambient. Probably when Deep Exporation convert to 3ds something goes wrong. Do you have an option inside the DE program that tell you about how many polygons do you want to be shaped the DWG/DXF? If so, try to reduce the value. I've found that your model crash either in Truespace and Cinema4D.