Unable to compute Brouwer-Lyddane mean parameters

Orekit usage
1

Hello,

I am experiencing an issue when converting from an osculating to a mean orbit based on the Brouwer-Lyddane propagator. I remember having the same error before at some point but usually I can convert fine back-and-forward from mean to osculating with minimal change in the parameters and with the default epsilon and iterations.

I am setting up the spherical harmonics model as (I also tried with order = 5):

UnnormalizedSphericalHarmonicsProvider provider = GravityFieldFactory.getUnnormalizedProvider(5, 0);

And then using .computeMeanOrbit() with the osculating orbit’s date for the provider and with BrouwerLyddanePropagator.M2. I get the following error:

Exception in thread "main" org.orekit.errors.OrekitException: unable to compute Brouwer-Lyddane mean parameters after 201 iterations
        at org.orekit.propagation.conversion.osc2mean.FixedPointConverter.convertToMean(FixedPointConverter.java:254)
        at org.orekit.propagation.analytical.BrouwerLyddanePropagator.computeMeanOrbit(BrouwerLyddanePropagator.java:783)
        at org.orekit.propagation.analytical.BrouwerLyddanePropagator.computeMeanOrbit(BrouwerLyddanePropagator.java:741)
        at org.orekit.propagation.analytical.BrouwerLyddanePropagator.computeMeanOrbit(BrouwerLyddanePropagator.java:691)
        at org.orekit.propagation.analytical.BrouwerLyddanePropagator.computeMeanOrbit(BrouwerLyddanePropagator.java:657)
...

The initial orbit that I initially convert to osculating is:

Keplerian parameters: {a: 6878029.072145119; e: 0.001075536702561181; i: 97.42147320105057; pa: 89.99999999999794; raan: -131.11537426812785; v: 15.000000000002059;}

Which gives:

Keplerian parameters: {a: 6869814.7067674035; e: 0.00360140994434708; i: 97.42590626678786; pa: 85.19010782186336; raan: -131.11277925955898; v: 19.888817457681284;}

Then, when I convert back to a mean orbit I get the error stated before.

I’ve tried significantly increasing the iterations and decreasing epsilon (1e-3 runs but the orbits change quite a bit). Do you recommend using a different propagator for this type of conversion? I am curious to understand the root of the error.

Thanks a lot,

Pedro