Hi Laurent,
Attention: The Moon TLE (2line) it's not usefull for the "SGP4/8 two lines format ephemerid model " (InstantTrack, SatPC32 etc. ) the format is only a compress readable parameters
The moon:
Giving Keplerian elements for the moon is much more difficult. The main problem is that the moon's orbital plane has a fixed inclination with reference to the ecliptic plane. The moon's orbital plane wobbles around there and turns once every 18.6 years. Seen from the earth, that gives an changing inclination of about 18 to 28 degrees. The inclination of true earth satellites, on the other hand, remains fixed and the tracking program assumes that the inclination is not changing. The only way to overcome that problem is to give sets of "Keplerian elements" every 50 to 100 days. In that period the change of the inclination is not too big and the error stays small.
A second problem with "Keplerian elements" for the moon is that the moon's orbit is severely perturbed in a way a ordinary satellite tracking program cannot handle. Those perturbations come from the sun, the earth (not being a exact sphere) and from major planets. To compute those perturbations exactly, we would need several hundred corrective terms. For the accuracy needed here (1-3 degrees for the usual amateur satellite antenna), it has been sufficient to correct only for the varying eccentricity.
73, de Michael