|
Atomistic simulation approaches such as molecular dynamics (MD) in conjunction with
massively-parallel supercomputers and refined interatomic potentials can now reach material length
scales directly comparable to experiment. They make it possible, for example, to study
nanomechanical behavior of entire microstructures down to the atomic level in full detail and
provide a wealth of information on the involved material defects and physical mechanisms.
However, the huge number of degrees of freedom in such large-scale simulations necessitates
advanced data filtering and analysis techniques that help us deal with the inherent complexity of
many materials processes and structures and allow us to compress the output data to be stored. The
goal of ongoing research is therefore to develop smart computational methods that can reduce the
simulation output to essential features and link atomistics with mesoscopic materials concepts such
as grains, dislocation lines, surfaces and deformation fields – in other words, high-level
descriptions that are more instrumental in gaining insights than the underlying all-atom model.
Furthermore, I will provide an overview of one of the leading data visualization tools in the
field (“OVITO”), which is being developed in my group to work with outputs from particle simulation
models.
|
