This project builds on a long-standing collaboration between the Jülich Supercomputing Centre (JSC), Forschungszentrum Jülich, and the RIKEN Center for Computational Science (R-CCS), under the Joint Laboratory for Extreme Scale Computing (JLESC). In the first project, "Comparison of Meshing and CFD Methods for Accurate Flow Simulations on HPC systems," the m-AIA and CUBE codes from RWTH Aachen University and RIKEN R-CCS were ported and benchmarked on Jülich and RIKEN machines. The recent project, "Deep Neural Networks for CFD Simulations," focuses on developing AI methods for engineering applications, particularly in biomedical computations, such as respiratory flow analysis for personalized medicine, part of CoE RAISE.
The first goal of this project is to develop AI-assisted automated methods for running large-scale computational fluid dynamics (CFD) simulations, including analyzing results for individual patients. The second goal is to use AI methods and novel hardware technologies at JSC and RIKEN R-CCS to speed up computations. The third goal is to provide an AI-assisted surgery planning tool for surgeons as a demonstration. The methodologies developed will also help achieve the final goal: assessing the risk of exhaled infectious viruses. All implementations will enhance the capabilities of the m-AIA and CUBE codes.