An overall strategy for numerical simulations and optimization of fluid flows is introduced. The integrative approach takes advantage of numerical simulation, high performance computing (HPC) and newly developed mathematical optimization techniques, all based on a mesoscopic model description and on Lattice Boltzmann Methods (LBM) as discretization strategies [1]. The resulting algorithms were implemented in a highly generic way in the open-source framework OpenLB [2]. In the talk, particular focus is placed on the systematic approach of facing contemporary challenges in Computational Fluid Dynamics (CFD) [3, 4]. Further, the consideration of LBM as a generic technique for the approximation of Partial Differential Equations (PDE) [5] and its implementation on heterogeneous HPC-platforms are highlighted. The presented approaches and realizations are illustrated by means of various fluid flow simulation and optimization examples, where specific aspects are discussed for the simulation of turbulent flows [6].
Biography
Mathias J. Krause studied mathematics with economics at the Universität Karlsruhe (TH) in Germany and the Cardiff University in Wales. He graduated in 2006. Afterwards, he joined the Institute of Applied and Numerical Mathematics (IANM) at the Karlsruhe Institute of Technology (KIT). In 2010, he received a doctorate with his thesis on fluid flow control and optimization. Since April 2013, he heads the interdisciplinary Lattice Boltzmann Research Group (LBRG) at KIT. His research interests are mainly dedicated to the fields of applied mathematics, applied computer science with focus on HPC, CFD and optimization under the constraints of PDE. He is initiator and main author of OpenLB. His work was honored with several prizes and a membership as WIN-Kollegiat at the Heidelberg Academy of Science.
Adrian Kummerl is currently writing his master’s thesis on high performance Lattice Boltzmann implementations at KIT, Germany. He is a professional software developer and has been working on OpenLB since joining the LBRG during his undergraduate studies in Mathematics at KIT. His research interests focus on performance optimization and HPC applications of LBM.
Stephan Simonis is a doctoral student at the IANM, KIT. His research focuses on the mathematical background and numerical aspects of discrete velocity models, relaxation systems and LBM. He is part of the LBRG and contributes to OpenLB. Prior to joining the LBRG he completed his BSc and MSc in Mathematics at KIT in Germany and received funding for an MSc exchange year in Applied and Computational Mathematics at the KTH Royal Institute of Technology in Sweden.
