Numerical simulation is used to understand and predict the behaviour of complex systems by codifying the laws of physics from mathetical equations into computer programs. The simulations are run on computers ranging from laptops to supercomputers. My work focusses on the field of 'mesh free' methods, such as the Smoothed Particle Hydrodynamics (SPH) and Material Point Method (MPM) techniques and using them to model fluid and solid problems, alike.
Machine learning algorithms and structures like deep neural networks are excellent at relating inputs to outputs. Building a statistical mapping requires enormous amounts of data. This work focusses on using validated simulation data to teach these structures the laws of physics. This way we can accelerate prediction and estimation of physical processes.