Computational Modelling Group

Pushing the Envelope of Planetary Formation and Evolution Simulations

1st September 2016
Research Team
Peter Bartram

A full understanding of the formation and the early evolution of the Solar System and extrasolar planetary systems ranks among natural science's grand challenges, and at present, even the dominant processes responsible for generating the observed planetary architecture remain elusive. For the last three decades, a wealth of new observations has led to renewed interest in the problem, as a dramatically increasing number of exoplanets have been identified by the COROT and KEPLER space missions.

As more reliable and complete datasets become available, more refined and sophisticated models can be proposed and compared upon observations, thus making our comprehension of planetary formation processes more insightful. These advances have been accompanied by an increase in computational power, allowing models to include richer physical descriptions, reproduce more complex dynamical features and exceed prior computational limitations. As a consequence, many theories could be recently revisited and superseded by new, more sophisticated models.

In traditional N-body simulations, the evaluation of the force over all pairs of particles scales as N squared, thus limiting these schemes to relatively modest particle numbers. As a consequence, trade-offs of accuracy for speed become necessary for increasing values of N, typically involving the use of tree codes, grid methods with pre-computed potentials, or similar approaches. Planetary formation is a very particular instance of N-body simulations, where particular dynamical features and constraints can be properly exploited to develop specifically tailored techniques that overcome the aforementioned limitations. On top of that, efficient N-body orbital propagation and close encounters need to be properly addressed.

This project is developing advanced numerical methods and techniques to improve the simultaneous orbital propagation capabilities of massive sets of planetesimals and smaller-sized particles, as well as to efficiently deal with their gravitational interactions whilst incorporating richer physical models into numerical simulations for the formation and evolution of planetary systems. These methods will provide a powerful state of the art simulation tool and a valuable asset for pushing the boundaries of our current knowledge in the mechanisms that drive planetary formation processes.


Physical Systems and Engineering simulation: Astrophysics

Algorithms and computational methods: Distributed computing, Finite differences, Finite elements, Graph Theory

Software Engineering Tools: Continuous Integration, Git, Jenkins, Mercurial, SVN, Trac

Programming languages and libraries: C, C++, CUDA, CUDA Fortran, Fortran, GPU-libs, IPython/Jupyter Notebook, Julia, Mathematica, Matlab, MPI, Python, R

Computational platforms: ARCHER, Cloud computing, GPU, Linux, VirtualBox

Transdisciplinary tags: Computer Science, Data Science, NGCM, Scientific Computing