andrewngt / formflyt Goto Github PK

🏹 Goal: Figure out the extent of numerical integration errors using various tested numerical propagators.

We attempt to investigate a numerical propagator that should be efficient (i.e. Runge-Kutta variants) but not overly complex to the point that it makes orbit propagation slow on MATLAB, as we might need to propagate orbits for days, and for N number of satellites in an arbitrary configuration.

Current position propagation errors for a satellite with initial conditions:

a1  = 7000000;     % Semi-major axis (m)
e1  = 0.001;       % Eccentricity (unitless)
i1  = 10.00;       % Inclination (degrees)
w1  = 90.00;       % Arg of Periapsis (degrees)
R1  = 45.00;       % Right Ascension (degrees)
M1  = -90.00;      % Mean Anomaly (degrees)
Cd1 = 2.2;         % Drag coefficient
Ar1 = 0.374;       % Drag area (m^2)
Ms1 = 17.90;       % Spacecraft mass (kg)

Propagator is a modified RK4 using the 3/8 rule, and the errors are compared with the analytical two-body solution (i.e. rotational solutions of an object about an elliptical orbit). Pretty significant errors.

In library/formflyt_orbits/kepler_states.m, line 39, should the x position be calculated as: pos_X=a*(cosd(eccAnom) + e); instead?

Eccentricity should cause the focus to translate in the positive x direction rather than the negative x direction?