Abstract:
Throughout the formation and evolution of planets, collisions with other celestial objects are common, ranging from micrometeorite impacts to major encounters with other protoplanets. This study investigates the conditions under which such events occur, examining the orbital- and physical properties of pre- and post-collision bodies and analyzing the intricate dynamics involved in these interactions. We begin by shortly summarizing key principles of planet formation, as well as introducing the computational methodologies employed in this work. Our astrophysical contributions include an analysis of minor impacts on Jupiter’s icy moons, Ganymede and Callisto, an evaluation of post-collision fragments from pairwise planetary collisions, and the compilation of a comprehensive collision catalog spanning a broad range of plausible initial conditions. Methodologically, we advance the field by developing and refining numerical simulation techniques for collision modeling and enhancing large-scale data analysis through the application of modern machine learning approaches. Our findings support established theories of planet formation. We provide refined estimates of impact velocities onto Jupiter’s icy moons and study the trajectories of post-collision fragments after major collisions. Utilizing our numerical simulation framework, we conduct extensive parameter studies of planetary collisions. The resulting data is then employed to train modern machine learning methods for fast and accurate collision handling. We conclude that combining numerical simulations with machine learning opens up new pathways in theoretical astrophysics, representing a scalable and versatile framework for elucidating the intricacies of planet formation.