I am trying to do an OD from TLEs plus some observational data. I want to use a numerical propagator to do this but I do not know the mass or drag cross section of the satellite that the data is for. The TLEs do however contain the B* coefficient which contains all of this information albeit combined. My question is, is there a way to configure a numerical propagator to make use of the estimated B* rather than to provide a mass and csa? I always call .setMass() on my propagator builders and also specify the csa when configuring my IsotropicDrag objects. I can’t convince myself why this is necessary as in the force models the mass and csa are combined to calculate the acceleration due to drag anyway, but I can’t see a way of configuring the drag model and thus numerical propagator without providing the mass and csa?
You have to set up a mass and CSA.
The reason is that for other forces (maneuvers, solar radiation pressure, albedo, IR), the combination would be different. Maneuvers would need the mass and no surface, radiation-related forces need a surface which may be different from drag.
Thanks for the speedy reply. That makes sense about the other forces, I guess I was only considering drag as I am interested in objects <500km.
I suppose I could get the mass to csa ratio from B* and then try orbit fitting with a range of masses with appropriately scaled csas and look at what residuals I get.
What I usually do in a Space Surveillance context when the objects are unknown is set the mass to 1kg and the drag and srp area to 1m² (or more generally S=m)
(Note that it will definitely not work if you want to add maneuvers.)
Then what you’re estimating is directly \frac{C.S}{m} where C is the drag or SRP coefficient.
The \frac{C.S}{m} is the only thing that is observable anyway.
This is the context that I was thinking of too. I will also try to use the same process. I guess my worry was that the orbit fit would be poor if I specified the wrong mass or csa but if I also estimate CS/m then I suppose this is taken care of?