Capability Development for Modeling Planetary Impacts, 15-9089Printer Friendly Version
Inclusive Dates: 08/01/98 - 01/01/00
Background - A fundamental question in planetary science is how the planets in our solar system formed, and, in particular, what processes led to the formation of a habitable Earth. Recent work indicates that the final stage of planet formation was dominated by hypervelocity collisions between planet-sized bodies. Such impacts were so large that they determined a planet's spin rate, the tilt of its rotational pole, and if it ended up with a moon. The specifics of the last impacts that a planet experienced determine many of its end characteristics, including its ability to sustain a life-supporting climate. Many of the planets in our system show evidence of such impacts (e.g., Mercury, Uranus, Earth/Moon, and Pluto/Charon). Despite their fundamental importance, modeling of planet-scale impacts has been difficult, and previous computer models have been hindered by slow computational speeds. The main challenges in modeling large impact events are: 1) their physical scale precludes extrapolation from laboratory experiments, and 2) energies involved are so large that both gravitational dynamics and thermodynamics of the involved materials must be explicitly tracked.
Approach - In this project, the research team developed expertise in the use of a numerical impact simulation using smoothed-particle hydrodynamics, or SPH. In SPH, objects are represented by a collection of spherical particles whose evolution is tracked as a function of time. The SPH method follows the history of each particle as it moves in space, making it well suited to problems involving deformation and debris ejection. An SPH run's thermodynamic output is the internal energy of each particle, as well as locally computed pressure and density; the dynamic outputs are the mass, position, and velocity of each particle. To gain proficiency in the use of the SPH method (which had not been previously utilized by our group), the SwRI team involved an outside collaborator who is expert in SPH methods, Dr. Erik Asphaug. With the simulation utilized in this project, the research team can achieve greater speeds (and therefore higher numerical resolutions) than previously obtained.
Accomplishments - Serial and parallel versions of the SPH code were ported onto the team's cluster of 16 DEC Alpha nodes. Testing was completed to ensure that the simulations reproduced results obtained by other SPH codes and those expected for idealized cases. Comparison of simulations completed at several numerical resolutions indicates that the resolutions obtainable by the SwRI model are sufficiently high to provide convergence of numerical results. In conjunction with this code-development effort, the team identified a new analytic scaling relationship that appears to generally predict the results of the SPH simulations of planet-scale impacts done to date as a function of impact parameters.