ROSETTA Lander Philae – Touch-down Reconstruction

The landing of the ROSETTA-mission lander Philae on November 12th 2014 on Comet 67P/Churyumov-Gerasimenko was planned as a descent with passive landing and anchoring by harpoons at touch-down. Actually the lander was not fixed at touch-down to the ground due to failing harpoons. The lander internal damper was actuated at touch-down for 42.6 mm with a speed of 0.08 m/s while the lander touch-down speed was 1 m/s. The kinetic energy before touch-down was 50 J, 45 J were dissipated by the lander internal damper and by ground penetration at touch-down, and 5 J kinetic energy are left after touch-down (0.325 m/s speed). Most kinetic energy was dissipated by ground penetration (41 J) while only 4 J are dissipated by the lander internal damper. Based on these data, a value for a constant compressive soil-strength of less than 1.7 kPa is calculated. This paper focuses on the reconstruction of the touch-down at Agilkia over a period of around 20 s from first ground contact to lift-off again. After rebound Philae left a strange pattern on ground documented by the OSIRIS Narrow Angle Camera (NAC). The analysis shows, that the touch-down was not just a simple damped reflection on the surface. Instead the lander had repeated contacts with the surface over a period of about 20 s +/-10 s. This paper discusses scenarios for the reconstruction of the landing sequence based on the data available and on computer simulations. Simulations are performed with a dedicated mechanical multi-body model of the lander, which was validated previously in numerous ground tests. The SIMPACK simulation software was used, including the option to set forces at the feet to the ground. The outgoing velocity vector is mostly influenced by the timing of the ground contact of the different feet. It turns out that ground friction during damping has strong impact on the lander outgoing velocity, on its rotation, and on its nutation. After the end of damping, the attitude of the lander can be strongly changed by the additional ground contacts even with the flywheel still running inside the lander. The simulation shows that the outbound velocity vector and the lander rotation were formed immediately at touch-down during the first 1.5 s. The outbound velocity vector is found to be formed by the ground slope and the lander damping characteristic, especially the nearly horizontal flight out