Skip to content

utilities.elements_from_state_vector: fix nan values for f#49

Open
mgiuca wants to merge 6 commits into
RazerM:masterfrom
mgiuca:from_state_vector-clamp-acos
Open

utilities.elements_from_state_vector: fix nan values for f#49
mgiuca wants to merge 6 commits into
RazerM:masterfrom
mgiuca:from_state_vector-clamp-acos

Conversation

@mgiuca
Copy link
Copy Markdown
Contributor

@mgiuca mgiuca commented Sep 16, 2025

This was caused by the argument to acos being slightly higher than 1.0 due to rounding errors, resulting in nan (when it should just be 0).

Fixes #40.

This is intended to be based on top of PR #48.

Adds a missing check for the case where the eccentricity vector is in the third
or fourth quadrant, requiring arg_pe to be negated.

Note that this causes an edge case where arg_pe should be 0° but it comes out as
360° because we are not using modulo, but rather adding 2*pi. I didn't want to
fix that in this commit because all the calculations in from_state_vector behave
the same way.

Fixes RazerM#39.
When negating an angle, consistently use mod 2*pi instead of adding 2*pi. This
causes 0 to remain as 0 instead of becoming 2*pi.
This was caused by missing check for the case where e ~= 0 when orbits are
inclined, thus dividing by the eccentricity resulted in a divide-by-zero.

Fixes RazerM#38.
…orbits.

This was calculating both raan and f as nan, since i=180° was going into the
inclined orbit code path, when it is actually non-inclined. When invoked via
KeplerianElements.from_state_vector, the nan value would fail an assertion.

This also adds an important case needed to get the correct value of arg_pe
for retrograde non-inclined orbits: if i=180°, arg_pe needs to be inverted since
the orbit itself is upside down. Added two regression test cases for this case.

Fixes RazerM#18 and RazerM#25.
…tion.

This abstracts the repeated code that finds the unsigned minimum angle between
two vectors using acos to a standalone function.
This was caused by the argument to acos being slightly higher than 1.0 due to
rounding errors, resulting in nan (when it should just be 0).

Fixes RazerM#40.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

from_state_vector sometimes crashes with f=nan at the periapsis

1 participant