Utilizing Supercomputing to Analyze Risks of an Emergent Large-scale Debris Field in Low Earth Orbit

David Buehler, AFRL/RV; Dane Fuller, AFIT/CSRA; John Colombi, AFIT/ENV; William Wiesel, AFIT/ENY; Dave Meyer, AFIT/ENY; Richard Cobb, AFIT/ENY

Keywords: Space debris, modeling, supercomputing, high performance computing, collision likelihood

Abstract:

The likelihood of on-orbit breakups, whether spontaneous or the result of collision, will likely continue to grow as the barriers of entry to and use of space are reduced. In all orbitalregimes, especially Low Earth Orbit (LEO), preparation to respond quickly when the next breakup occurs is critical. This research utilizes high-performance parallel computation along with Python-driven Systems Tool Kit (STK) to model a large-scale on-orbit breakup in LEO, with the goal of returning data on an operational timeline of 90 minutes or less. 90 minutes was chosen as it is the orbital period of an average LEO satellite. Data return in one orbit or less gives operators’ awareness of the situation quickly, allowing for a maneuver to be conducted if necessary. The breakup is characterized by the National Aeronautics and Space Administration(NASA) EVOLVE 4.0 breakup model and each piece is propagated for one week or until it drops below the set reentry altitude of 120 kilometers.

The debris field is analyzed over the course of one week using Gabbard plots. These Gabbard plots are then compared with the historical data of the FY-1C breakup to ensure accuracy of the model. Additionally, a modified Gabbard plot is used to analyze the inclination spread affected by the breakup. The risk posed by the breakup is determined using STK’s Advanced Close Approach Tool (ACAT) to report minimum range, minimum separation, and likelihood of collision between the debris and catalog. The minimum range values are evaluated for red zone approaches (x ? 1 kilometer), yellow zone approaches (1 kilometer < x < 2 kilometers), and green zone approach (2 kilometers ? x ? 3 kilometers), where x is the minimum range of the close approach. The field is screened for close approaches with historical TLEs for each day of the analysis week and the probability of collision is computed using multiple conjunction models (Alfano, Patera, Chan, Alfano Max) to observe how different models predict the likelihood of collision. The goal is to take steps towards preparing to respond to breakup events in the future. Over the course of the analysis week, there were over 700,000 close approaches reported within the three kilometer threshold. ~11% of these approaches were reported within 1 kilometer, ~33% were reported between 1 and 2 kilometers, and ~56% were reported between 2 and 3 kilometers. The Gabbard plot of the breakup was visually compared to that of the FY-1C breakup. The modified Gabbard plot of the breakup revealed that the average inclination of the field stayed around that of the parent and the majority of the pieces were contained within ± 1.5° inclination from the parent inclination. The average daily likelihood of at least one of the reported collisions happening was found to be 4.1% for the first three models and 22.3% for the maximum likelihood model. Within the analysis week, 240 of the debris pieces had fallen below the reentry altitude and were removed from the scenario. These results were returned within 120 minutes real time, which did not satisfy the operational timeline desired. However, there are several suggestions proposed to reduce the time to meet the desired 90-minute threshold. Date of Conference: September 11-14, 2018

Track: Poster

View Paper