Currently, over 45,000 human-made objects are orbiting our planet, forming a tangle of trajectories that seems increasingly difficult to secure.
A team from the Lawrence Livermore National Laboratory in California has developed a new approach aimed at modeling trajectories in cislunar space. This region, between Earth and the Moon, is a passage point for space missions.
One of the cislunar orbits calculated by the researchers. The lunar orbit appears in light gray, while the spacecraft follows the colored trace over six years of simulation.
Credit: Dan Herchek
With this in mind, scientists simulated one million orbits over six years, using a public database and the capacity of their supercomputers. This technique allows for an overview of possible movements, simplifying examination through machine learning tools to spot irregularities or estimate the stability of trajectories.
The results indicate that approximately half of the modeled orbits remain stable for at least one year, while less than ten percent maintain this stability over the entire six years. This information provides useful clues for anticipating satellite behavior over a long period, where traditional equations fail to rigorously predict future positions.
This computing operation required 1.6 million hours of processor time, which would correspond to over 182 years on a conventional computer. However, thanks to the Quartz and Ruby machines at the laboratory, these simulations were completed in just three days.
This work could help identify areas of high and low orbital density and improve satellite launches. With the continuous growth of space launches, such modeling provides valuable information to coordinate activities in the absence of binding international agreements.
Scientists believe that their method will help reduce collision risks for the coming decades.