I'm trying to propagate the orbit of a sun-synchronous satellite. The initial data is defined by the TLE.

The orbit of the satellite is controlled by the operator by means of maneuvers. That's to say, the effects of perturbations are compensated, the orbit of the satellite always remains sun-synchronous.

Is there a way to propagate the orbit, considering the above conditions, making the calculations on the kepletian elements?

I tried the SGP4 model, but it's not appropriate in this case.

  • $\begingroup$ Can you add the source of the images you posted? $\endgroup$
    – Chris
    Commented Mar 21, 2018 at 19:10

2 Answers 2


TLE's already considers drag up to some point (check if BSTAR is negative, if it is, then the TLE is considering that maneuvers have been performed). A positive BSTAR reduces (mean) semi-major axis over time, an effect of drag. J2, J3 and J4 are each partially considered (think taylor series approximaiton with $J_2^2$ set to zero).

If you have a sun-synchronous orbit, then at least J2 needs to be considered, otherwise there could be no such orbit. All other perturbations are needed only on a basis of propagation duration and simulation fidelity: How long the simulation runs and how accurate should it be. 2 weeks and low accuracy means SGP4 output is fine.

I believe SGP8 will consider sun and moon perturbations, but not SGP4.

  • $\begingroup$ These days the umbrella "SGP4" is inclusive of SDP4 and so it has a limited ability to approximate perturbations from the Sun and Moon. See “Deep space” corrections in SGP4; how does it account for the Sun's and Moon's gravity? $\endgroup$
    – uhoh
    Commented Mar 12, 2018 at 7:38
  • $\begingroup$ Looks like I should consider only Earth and J2. So, SGP4 is not good for this task. Is there a way to propagate on the basis of orbital elements? $\endgroup$ Commented Mar 12, 2018 at 7:50
  • $\begingroup$ @Mefitico can you consider updating your answer to adjust your statement that SGP4 does not consider Sun and Moon perturbations. It didn't at one time (SDP4 did) but I think today any modern source for SGP4 does include these perturbations by incorporating SDP4. $\endgroup$
    – uhoh
    Commented Mar 21, 2018 at 3:10
  • $\begingroup$ @uhoh Anyway, SGP4 doesn't comply with the stated conditions. Please, consider my equations $\endgroup$ Commented Mar 21, 2018 at 13:44
  • $\begingroup$ @TarlanMammadzada : I recognize the equations you presented in (1) for the rates of the elements are Kozai's equaitons, and they are should be used with mean elements to update mean elements, and the output needs to be converted to osculating elements before using the equations you present in (2). See as a reference Schaub H. Junkins, J. Analytical Mechanics of Space Systems 2nd edition AIAA Education Series 2009 $\endgroup$
    – Mefitico
    Commented Mar 21, 2018 at 18:21

With respect to what orbit propagator you should use, it depends. If you are designing a space mission that will operate in a Sun-syncrhounous orbit, then usually a J2 is better. In those cases, the mission usually have ways to keep itself on the designed orbit, avoiding the all perturbations but J2. Hence, a J2 propagator is better. On the other hand, if you are propagating a real satellite using a real TLE, then SGP4 is much more accurate.

If you use Julia and SatelliteToolbox.jl, then you can very easily select the propagator. For example, for a pure J2 propagator:

julia> orbp = init_orbit_propagator(Val{:J2}, Orbit(0.0,7130982.0,0.001111,98.405*pi/180,pi/2,0.0,0.0))
julia> (o,r,v) = propagate!(orbp, collect(0:3:24)*60*60)
julia> r
9-element Array{Array{T,1} where T,1}:
 [5.30372e-7, 7.12306e6, 3.58655e-6]
 [-9.98335e5, 2.14179e6, -6.72549e6]
 [-5.75909e5, -5.83674e6, -4.06734e6]
 [6.65317e5, -5.69201e6, 4.2545e6]
 [9.62557e5, 2.37418e6, 6.65228e6]
 [-1.10605e5, 7.11845e6, -231186.0]
 [-1.02813e6, 1.90664e6, -6.79145e6]
 [-4.82921e5, -5.97389e6, -3.87579e6]
 [750898.0, -5.53993e6, 4.43709e6]

This propagated the orbit for 24h and stored the information every 3 hours. For SGP4, then you can do the following:

julia> orbp = init_orbit_propagator(Val{:sgp4}, Orbit(0.0,7130982.0,0.001111,98.405*pi/180,pi/2,0.0,0.0))
julia> (o,r,v) = propagate!(orbp, collect(0:3:24)*60*60)
julia> r
9-element Array{Array{T,1} where T,1}:
 [-2159.7, 7.13166e6, -14607.2]
 [-1.00096e6, 2.1411e6, -6.73899e6]
 [-5.78906e5, -5.83897e6, -4.08451e6]
 [6.64614e5, -5.70129e6, 4.24735e6]
 [9.6287e5, 2.37768e6, 6.64987e6]
 [-1.12629e5, 7.12679e6, -2.45705e5]
 [-1.03066e6, 1.90639e6, -6.80469e6]
 [-4.86132e5, -5.97626e6, -3.89338e6]
 [7.5014e5, -5.54932e6, 4.42998e6]

If your orbit is defined in the file sat.tle, then you can propagate the orbit using:

julia> tle = read_tle("sat.tle")
1-element Array{SatelliteToolbox.TLE,1}:
                            Name: Test
                Satellite number: 6251
        International designator: 62025E
                    Epoch (Year): 6
                     Epoch (Day): 176.82412014
              Element set number: 398
                     Inclination:  58.05790000 deg
                            RAAN:  54.04250000 deg
             Argument of perigee: 139.15680000 deg
                    Mean anomaly: 221.18540000 deg
                 Mean motion (n):  15.56387291 revs/day
               Revolution number: 677

                              B*: 0.000000 1/[er]

                        1   d
                       ---.--- n: 0.000000 rev/day²
                        2  dt

                        1   d²
                       ---.--- n: 0.000000 rev/day³
                        6  dt²

julia> orbp = init_orbit_propagator(Val{:sgp4}, tle[1])
julia> (o,r,v) = propagate!(orbp, collect(0:3:24)*60*60)
julia> r
9-element Array{Array{T,1} where T,1}:
 [3.98829e6, 5.49898e6, 928.983]
 [4.77232e6, 4.45027e6, -1.90764e6]
 [4.99362e6, 2.89039e6, -3.60054e6]
 [4.62067e6, 1.00698e6, -4.88731e6]
 [3.69211e6, -9.77058e5, -5.62356e6]
 [2.31243e6, -2.82934e6, -5.72624e6]
 [6.40617e5, -4.33414e6, -5.18251e6]
 [-1.1272e6, -5.31664e6, -4.05161e6]
 [-2.77972e6, -5.66325e6, -2.45952e6]

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.