DSST mean conversion : increasing potential degree does not seem to enhance mean semi major axis

1

Hi all,

I am facing a quite weird behaviour while trying to convert osculating to mean orbits using DSST. For a given osculating orbit (LEO one), I propagate it numerically using a geopotential only force model with a given time step (1 minute) over roughly one orbital period (2h) and convert each osculating state into mean using the same degree/order for the geopotential.
I was expecting to get a very approximate result in each case (the force model being the same between the osculating propagated orbit and the mean one). This is not the case (cf. below) and more annoyingly, increasing the degree/order seems to worsen the mean semi major-axis (less flat for 16x16 than for 2x2). Does this mean that DSST reduction is less efficient for high geopotential degree/order ?

dsst_mean_conversion_comparison
dsst_mean_conversion_comparison999×665 65.4 KB

Please find attached the code which generated the plots.

Thanks in advance for your feedback.

François

compare_dsst_mean_conversion_forum.py (7.1 KB)

2

Just an update on the orekit and hipparchus versions used in the script. In the previous one, these were old versions. Please find attached with the latest versions of both orekit and hipparchus.
Result still holds.

dsst_mean_conversion_comparison_orekit13.1
dsst_mean_conversion_comparison_orekit13.11603×845 90.7 KB

compare_dsst_mean_conversion_forum.py (7.1 KB)

3

Hi François,

You might be seeing yet another manifestation of issue 848 issue 1491

Cheers,
Romain

4

Hi Romain,

Thank you for your answer. However, I am a somewhat skeptical. Issue 858 deals with TLE / SGP4, not DSST. Have I missed something ?
Thanks,

François

5

Sorry wrong copy paste

6

Thank you Romain. Indeed there seems to be quite similarities in the issues.