Model order reduction methods for efficient full scale selective laser melting simulation

This PhD aims to provide novel solutions to simulate the 3D printing process for selective laser melting. As these 3D printing methods have gained broader acceptance within industry, improved process control is of paramount importance. Advanced numerical simulation tools have the potential to provide strongly improved insight in the printing process, especially when it comes to the propagation of deformation defects from a local to a full printed component level. Current methodologies are however too expensive to enable full scale simulation in a more practical engineering setting. Recently a range of model order reduction methodologies have been developed which enable accelerated simulations both for nonlinear and multiscale problems. This project aims to leverage these techniques in order to bring the computational load of the 3D printing process simulation down to a level where they can be exploited in engineering practice. This research will develop novel model order reduction schemes tailored to the specific application of selective laser melting. Starting from a wide range of existing methodologies it is investigated which are most suitable for this coupled thermo-mechanical problem, based on physical insight they are adapted to achieve the best performance, and evaluate how the methodologies can be interfaced with commercial software tools.