CSIRO Computational Modelling and Simulation

Sharen Cummins
Seminar

CSIRO, Australia’s national scientific industrial research laboratory, has had a rich and sustained history in computational modelling of complex industrial, environmental and biophysical processes. In this presentation an overview of CSIRO’s computational  modelling capability will be given. The capability is primarily particle-based; Smoothed Particle Hydrodynamics (SPH) and Discrete Element Method (DEM) being the main numerical methods. However ever increasing commercial requirements to simulate complex multi-physics and multi-scale problems is driving a need to combine these particle based methods with grid based methods like FEM/FV and novel mesh-based algorithms. I will present progress towards our multi-scale and multi-physics modelling capability and discuss three driver applications - comminution, additive manufacturing and biomechanical modelling. I will also present CSIRO’s scientific workflow and application development framework called Workspace. Workspace is used to create licensed customised packages of our solvers  for our industrial clients. It is also used across CSIRO to deliver in other scientific applications.

Short Bio:
Sharen Cummins is a senior research scientist in the Computational Modelling group at CSIRO. She completed her PhD at the Mathematics at Monash University in 2000, specialising in modelling incompressible flows using SPH. She then worked at Los Alamos National Laboratory as a post-doctoral associate and research scientist where she developed numerical algorithms and software to model granular and fluid flows, crystallisation and casting processes. In 2008 she joined the Computational Modelling group at CSIRO as a software and algorithm developer where her work supports the accurate, efficient and robust simulation of industrial, biological and geophysical applications. Her career interests are in the development of particle-based methods (SPH and DEM) and grid based methods  (Finite Volume and Finite Element); in particular their coupling to model multiple physics and their efficient implementation in large commercial applications.