Re: [Orekit Users] use of Orekit for atmospheric entry

[MAISONOBE Luc <luc.maisonobe@c-s.fr>]

Hello Martin,

Martin <martin.spel@rtech.fr> a écrit :

> We are planning on using Orekit in a tool that includes the  
> atmospheric entry.

Fine, if you need some changes or have some contribution, do not hesitate.

> Does anyone see any problem with this? We are mainly talking from 122km to
> ground. In the Cunningham gravity model there is an error message  
> regarding the
> Brouillon sphere when entering within the earth equatorial radius.

If you were only going down to about 20-30km altitude, we could  
perhaps check whereas the threshold is too high. If you need to go at  
lower altitude, there is a real problem. The Brillouin sphere is the  
sphere that circumscribes the body. For the Earth, it is a sphere that  
is tangent to the ellipsoid at equator and is above ground at all  
non-zero latitude and up to 21.3 km above poles. In fact, it is even  
larger than that since atmosphere also has a mass.

Inside the Brillouin sphere, spherical harmonics diverge, even if you  
are still above Earth surface. Expanding spherical harmonics to very  
low altitude is a well known problem and I still it is still a  
research subject. In fact, the thesis of one of my coworkers some  
years ago used ellipsoidal harmonics to limit this problem for some  
small oblate bodies like comets. So I guess if you really need to go  
so low I will have this problem with Cunningham model which uses  
spherical harmonics.

However, I think that once you are in the final part of a trajectory  
below 100km, you really don't care much about spherical harmonics. The  
major forces in this region are the central term of the potiential  
(not harmonics) and of course the atmospheric drag. So I would suggest  
two different options.

First option should be used if you want both high accuracy for initial  
trajectory and be able to continue the trajectory at very low  
altitudes. In this case, you can use a complete force model including  
Cunnungham down to a limit altitude (you can use AltitudeDetector to  
stop the propagation when you reach this altitude), then start a new  
propagation from this intermediate state with a reduced force model,  
removing Cunningham.

Second option should be used if you don't need the extra precision. It  
consists simply in avoiding completely Cunningham right from the  
beginning of the trajectory. It is feasible only if your trajectory  
initial part is not too long.

best regards,
Luc

> Thanks,
> Martin



----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.