Re: [Orekit Users] Problem when doing propagation from different point of views

[Luc Maisonobe <Luc.Maisonobe@c-s.fr>]

Le 31/03/2014 22:32, Quentin Nénon a écrit :
> Hello,

Hi Quentin,

I still want to have a look at your problem, hopefully, I will be able
to do it in the next few days, sorry for the delay.

> 
> Thanks for the suggestion. I think that you refer to the paper updated
> every three year by the IAU Working group on cartographic coordinates
> and rotational elements. In this paper, the IAU publishes with a high
> precision information such as shape, size, pole position and rotation
> axis and rates for the celestial bodies in the solar system.
> 
> I am still a beginner with Orekit and I do not know if all these
> parameters are taken into account in the definition of the celestial
> bodies of the CelestialBodyFactory.

Yes, IAU pole and prime meridians are taken into account if you use
CelestialBodyFactory to build the celestial bodies for you. This factory
relies on JPLEphemerisLoader, which loads the ephemeris (any type you
want, in the DExxx or INPOP families) and then delegates to
IAUPoleFactory the construction of an IAUPole adapted to the body.

> I am quite sure that these ones are
> taken into account for the Earth as the EOP (Earth Orientation
> Parameters) are implemented and may be provided by the IMCCE database
> for instance.

For Earth, the way to handle it is often more complex as we do have more
accurate frames, and a lot of them (GCRF, EME2000, MOD, TOD, ...), so
the basic IAU pole and prime meridian are often not used for this
specific body (despite it is of course possible to use it).

If you want to use your own frame for some celestial body, you can add
it, but this is somewhat an expert use and I doubt you will try this.

> 
> But in the way I see Orekit and in the way I am trying to use it, these
> parameters are not important. I am only working with newtonian
> attractions from the different celestial bodies. I therefore do the
> approximation that there is an homogeneous distribution of the mass in a
> spherical body and that the attraction of each body can be reduced to
> the one produced by an infinitisimal point with a mass.

If you use only the Keplerian hypothesis, then the celestial body as
built by CelestialBodyFactory contains the relevant information (i.e.
the central attraction coefficient. This coefficient is retrieved from
the JPL/INPOP ephemeris.

> If the
> orientation parameters and the flattening of the different bodies are
> taken into account for the computation of the central body attraction
> and are not for the ThirdBodyAttraction, then a difference may appear
> due to the fact that you are on the "big side" or the "light side" of
> the planet. My mistake would therefore be to change from a central body
> attraction dealing with the orientation parameters and mass repartition
> corrections to a simple newtonian attraction. This may explain the
> results I have, even if I am surprised of the magnitude of the errors
> due to this effect...
> 
> I may be wrong but I think that Orekit allows for motion about any
> arbitrary central body. I understand that when you want define an orbit,
> this one should be defined in an inertial frame as the concepts of
> frames and referentials seem to be combined in Orekit. Then, the
> computation of the acceleration of the spacecraft is done using the
> inertial frame/referential of the central body. In Orekit, each
> celestial body has a function that returns an inertially oriented frame
> that is oriented like the Earth EME2000 frame (meaning with an
> inclination of about 23.4° with the ecliptic plane). I therefore think
> that Orekit is designed for what I want but, once again, I may be wrong.

You are right, this is the proper way to do it and Orekit was designed
for this. What Orekit can't do yet (I hope to add it in the future) is
compute an orbit without a central attraction (i.e. using only other
force models to describe all attracting bodies, like for example
trajectories about L2 point). For now, you must have one central body,
and this central body must be at the origin of the inertial frame.
Another way to say the same thing is that currently the central
attraction force is always computed and is always pointing towards
inertial frame origin.

> 
> Thank you very much again for your answer, I will create my own
> celestial bodies with only a spherical shape, a simple mass model and a
> position provider given by JPL ephemerides. I then hope to avoid any
> problem due to orientation and mass corrections.

I'm not sure this will help, I guess the error is elsewhere (maybe in
Orekit, maybe in your code), but its only a wild guess.

I'll try to have a look in the next few days.

best regards,
Luc

> 
> I keep you informed very soon,
> 
> Quentin
> 
> Le 31/03/2014 21:23, paulcefo a écrit :
>> Hello.
>>
>> Is Orekit designed to allow for motion about an arbitrary cental body?
>>
>> If so, what characteristics can this central body have?
>>
>> I am thinking of the paper that Ken Seidelmann updates for the IAU
>> every couple of years.
>>
>> Paul
>>
>