Question about the built-in Orbit and Rotation Models

The place to discuss creating, porting and modifying Celestia's source code.
Avatar
Topic author
PlutonianEmpire M
Posts: 1348
Joined: 09.09.2004
Age: 34
With us: 14 years 6 months
Location: MinneSNOWta

Question about the built-in Orbit and Rotation Models

Post #1by PlutonianEmpire » 09.11.2012, 04:14

Earlier, I mentioned my attempt to have Celestia simulate Earth's geologic history in real time.

However, I looked at the source code, and the code used for the rotation, precession, and orbit models, because I want to use SSC for when the clock goes beyond the built-in clamps. Example: I defined a test object using all the Earth's orbital parameters, and meticulously calculated a mostly correct period (1.0000174209555099247091033538672 years), and when went beyond the 4000 BCE/CE clamps for vsop87-earth, Earth drifted away from the test object, and after 500 years, it was on the opposite side of the sun as the test object. Naturally, being an extremely finicky perfectionist, I got quite angry, especially since it pretty much threw the calendar and daily clock off.

So, how do I make sense of the numbers that are used in the code for the vsop87 orbits and iau rotations, so I can have a smoother transition when Celestia crosses the 4000 BC clamp?

And please, no lectures on how accuracy decreases beyond certain points in time. Please. Don't. :roll:
Terraformed Pluto: Now with New Horizons maps! :D

Avatar
John Van Vliet
Posts: 2705
Joined: 28.08.2002
With us: 16 years 6 months

Re: Question about the built-in Orbit and Rotation Models

Post #2by John Van Vliet » 03.12.2012, 03:41

--- edit ---
Last edited by John Van Vliet on 19.10.2013, 03:59, edited 1 time in total.

Avatar
Topic author
PlutonianEmpire M
Posts: 1348
Joined: 09.09.2004
Age: 34
With us: 14 years 6 months
Location: MinneSNOWta

Re: Question about the built-in Orbit and Rotation Models

Post #3by PlutonianEmpire » 03.12.2012, 05:25

john Van Vliet wrote:you could set a timeline and use JPL-HORIZONS with a date set to the min. "BC 3000-Feb-23 "
but for 4000+ BCE ???????

that will get you "close" to the orbit

for rotational ???

and for Geological time scales
deep time and celestia do not mix well
I'm aware of the date overflow in previous versions of celestia, but in 1.6.0 and 1.6.1, I can go 2 billion years into past and future before hitting the built-in clamps for those dates with little trouble. The only bad "mix" I see is having to hav the timerate at 100x or 1000x. Were you referring to that?

Either way, what I'm looking for is what the "current" orbital and rotational parameters were at the points where the vsop87 and iau models cut off, so I don't have entire planets warp hundreds of thousands of kilometers to new positions when transitioning from the vsop87/iau models to the basic ssc parameters.
Terraformed Pluto: Now with New Horizons maps! :D

Avatar
t00fri
Posts: 8772
Joined: 29.03.2002
Age: 16
With us: 16 years 11 months
Location: Hamburg, Germany

Re: Question about the built-in Orbit and Rotation Models

Post #4by t00fri » 03.12.2012, 14:26

PlutonianEmpire wrote:...
I got quite angry, especially since it pretty much threw the calendar and daily clock off.

So, how do I make sense of the numbers that are used in the code for the vsop87 orbits and iau rotations, so I can have a smoother transition when Celestia crosses the 4000 BC clamp?

And please, no lectures on how accuracy decreases beyond certain points in time. Please. Don't. :roll:

The phenomenon of rapid orbit failure that you are observing near the limit of validity of the vsop87 approximation is unfortunately a generic behaviour that can hardly be improved.

Anyone with a solid math background about the behaviour of high-order numerical approximations will understand it right away:

vsop87 is not a basic physical theory but rather a numerical approximation of the cumulative effect of gravity (due to other bodies) on a given orbit, involving a huge number (about 400!) of polynomial terms per orbit in order to achieve the best possible accuracy around our present epoch.

Unfortunately, it is a mathematical fact that approximations in terms of high-order polynomials start to deviate more rapidly outside the proper domain of validity the higher the used polynomial degree!

Once you have understood and accepted this purely mathematical fact, there will also be no reason left for you to become angry ;-)

If one wanted to have a slower transition to non-sensical results, the only alternative would be to restrict the orbit approximations to much lower polynomial degrees. This in turn would imply a much worse ephemeridal accuracy in our epoch, which is certainly NOT what most of us want. The vsop87 approximation has been polished and fine-tuned by highly specialized experts over years, and the vsop87 code in Celestia has been copied straight from their publication site. Hence none of the devs is advised to put their hands on any of the terms in vsop87.

Fridger
Image


Return to “Development”

Who is online

Users browsing this forum: 6 guests