Propagation is inconsistent sometimes

[astrojuanlu]

For example:

op = StaticOrbitPlotter()

op.plot(
    Orbit.from_vectors(
        Sun,
        [0.36137779, 0.84843643, 0.36772403] * u.au,
        [-0.01849308, -0.00788258, -0.00339819] * u.au / u.day,
    ),
    trail=True,
)

[astrojuanlu]

Interestingly, this does not happen only with trail plotting. Take this snippet:

from astropy import units as u
from astropy.time import Time

import matplotlib.pyplot as plt

from poliastro.bodies import Sun, Earth, Mars
from poliastro.twobody import Orbit
from poliastro.twobody.propagation import propagate
from poliastro.maneuver import Maneuver
from poliastro.plotting.static import StaticOrbitPlotter
from poliastro.util import time_range

ss = Orbit.from_vectors(
    Sun,
    [0.36137779, 0.84843643, 0.36772403] * u.au,
    [-0.01849308, -0.00788258, -0.00339819] * (u.au / u.day)
)

coordinates = propagate(
    ss,
    time_range(ss.epoch, periods=200, end=ss.epoch + 10 * u.day),
)

fig, ax = plt.subplots()
op = StaticOrbitPlotter(ax=ax)
op.set_attractor(Sun)
op.set_frame(*ss.pqw())

op.plot(ss, color="r");
op.plot_trajectory(coordinates, color="k")

ax.set_xlim(-459279554.4666631, 80427317.98609273)
ax.set_ylim(-171462462.17020893, 171486927.37946865)

Explanation:

• In red, the fully sampled orbit, and a big red dot where the actual position is at the given epoch
• I am taking that Orbit, called ss, and calling the propagate function on it on a time range that starts at ss.epoch and starts a bit later to see if the small trace matches the position of the big red dot

...and it's not happening. There's something weird going on here.

[astrojuanlu]

This can be observed in poliastro 0.13.1.

[priyanshurohilla]

I think this is happening because you have not propagated the epoch . You have extracted the coordinates but did not propagated the epoch and i think maybe that's why the red marker has not changed the position.

[Meuge]

Hi @priyanshurohilla ! I’ve recently talked with @astrojuanlu about the issue, and it seems the problem is with the transpose in the plot coordinates.
Thanks for the suggestion :). I’ll also consider it while solving the issue.

[Meuge]

It seems the problem comes from the propagate method as there is a discrepancy in the position values. As you may notice from the example below:

from poliastro.bodies import Earth
from poliastro.twobody import Orbit
from poliastro.twobody.propagation import propagate

position = [-6045, -3490, 2500] * u.km
velocity = [-3.457, 6.618, 2.533] * u.km / u.s

ss = Orbit.from_vectors(Earth, position, velocity)

coordinates = propagate(ss,ss.epoch.reshape(1))
print(coordinates)
print(ss.rv()[0]) #The one define in position

Output

 
[(-7010.52954739, 1276.67143012, 3581.24784291)] km
[-6045. -3490.  2500.] km

[astrojuanlu]

Thanks @Meuge! This has nothing to do with plotting then, so changing the title accordingly.

[astrojuanlu]

Another way to reproduce this issue:

from astropy import units as u
from astropy.time import Time

import matplotlib.pyplot as plt

from poliastro.bodies import Sun
from poliastro.twobody import Orbit
from poliastro.plotting.static import StaticOrbitPlotter

orb = Orbit.from_vectors(
    Sun,
    [0.36137779, 0.84843643, 0.36772403] * u.au,
    [-0.01849308, -0.00788258, -0.00339819] * u.au / u.day,
)
op = StaticOrbitPlotter()

op.plot(orb)
op.plot_trajectory(orb.sample(min_anomaly=orb.nu, max_anomaly=orb.nu + 5 * u.deg), color="k")

plt.show()

Probably because:

In [11]: orb.r.to(u.km)
Out[11]: <Quantity [5.40613479e+07, 1.26924283e+08, 5.50107319e+07] km>

In [12]: orb.sample(1, min_anomaly=orb.nu, max_anomaly=orb.nu + 1e-3 * u.deg).get_xyz(xyz_axis=1)[0].to(u.km)
Out[12]: <Quantity [2.84137792e+08, 1.48650679e+08, 6.41613016e+07] km>

[astrojuanlu]

Original figure after merging #926:

✔️