DFG - scientific network
Scale-bridging simulation methods based
on order-reduction and co-simulation (CoSiMOR)
What is the aim of CoSiMOR?
Who contributes to CoSiMOR?
Former Members of CoSiMOR
- joint research: comparison of Hyper-Reduction (HR) and Discrete Empirical Interpolation Method (DEIM)
joint paper in preparation
- short course on Model Reduction in nonlinear dynamics of fluids and structures
Jan 25-29, 2016. Mines ParisTech, France
(www incl. material) (David Ryckelynck)
- Data-driven Model Order Reduction and Machine Learning workshop (MORML2016)
Mar 29-Apr 01, 2016. University of Stuttgart, Germany
(www) (Bernard Haasdonk)
- 3rd International Workshop on Order Reduction Methods (WORM2016)
Aug 29-31, 2016. Bad Herrenalb, Germany
(www) (Felix Fritzen, David Ryckelynck)
- Model Order Reduction seminar
continuous event, University of Stuttgart, Germany
(www) (Bernard Haasdonk, Jörg Fehr, Felix Fritzen)
- invited minisymposium Model reduction for coupled problems in material science at the ECCOMAS Thematic Conference Coupled Systems 2017
(David Ryckelynck, Felix Fritzen)
CoSiMOR is the acronym for the scientific network Scale bridging simulation methods based on order-reduction and co-simulation. The intention of the CoSiMOR network is to provide an environment for the interdisciplinary exchange of advanced computational methods and possible fields of application. The focus is on multiscale problems, e.g., in the sense of different time or length scales. From a methodological point of view, methods belonging to the class of co-simulation (CoSi) and model order-reduction (MOR) are investigated by the members of the network. Examples for such problems are found in electronic circuit simulation, where individual devices are replaced by fine resolution finite element models usually living at different time-scales. A possible solution technique for this problem class is co-simulation.
Another example is given by (thermo-)mechanical multi-scale problems in the context of mechanics of microstructured materials. These simulations exhibit a complex geometry and non-linear constitutive models. Therefore, the numerical solution to these problems is associated with high computational cost. A massive reduction of the CPU time and memory requirements is possible by means of model order reduction (MOR), often in the sense of reduced basis (RB) methods.
While the development of methods originates mainly from the mathematics community, realistic applications can only be chosen when considering actual engineering applications. These require a high amount of knowledge from a modeling point of view (constitutive modeling; geometrical modeling; experimental validation; …). A major goal to combine the knowledge of both fields is, thus, the stimulation of an open, interdisciplinary discussion of computational strategies in CoSiMOR.
Currently, six people contribute actively to CoSiMOR:
- Dr.-Ing. Felix Fritzen (www) (email) (applied mathematics; mechanics)
Chair of Continuum Mechanics, Institute of Applied Mechanics (CE), University of Stuttgart, Germany
activities: Reduced basis methods for nonlinear mechanical problems; focus on multiscale problems; applications to twoscale FE simulations, topology optimization and material modeling
- Prof. Dr. rer. nat. Bernard Haasdonk (www) (email) (applied mathematics)
Institute for Applied Analysis and Numerical Simulation, University of Stuttgart, Germany
activities: Research group Numerical Mathematics; Model reduction in particular reduced basis methods for parametric, linear, nonlinear, coupled, multiscale and inequality constraint problems; Machine learning and data-based modelling
- Dr.-Ing. Ralf Jänicke (www) (email) (mechanics)
Chair of Continuum Mechanics, Institute for Computational Engineering, Ruhr-Universität Bochum, Germany
activities: diffusion processes in fluid-saturated porous and fractured rocks; reduced order computational homogenization; identification of substitute material models
- Dr.-Ir. Varvara Kouznetsova (www) (email) (mechanics)
Mechanics of Materials, TU Eindhoven, The Netherlands
- Prof. Dr.-Ing. David Ryckelynck (www) (email) (numerical mechanics; mechanics of materials)
Centre des Matériaux, Ecole des Mines de Paris (Mines ParisTech), France
activities: POD and Hyper-reduction methods in elastoplasticity, viscoplasticity, damage mechanics and finite strain. Acceleration of inverse problem solution for parameter calibration by fitting experimental results.
- Dr. rer. nat. Sebastian Schöps (www) (email) (applied mathematics; electrical engineering)
Graduate School Computational Engineering, TU Darmstadt, Germany
activities: Model order reduction for high and low frequency electromagnetic problems, coupled "multiphysics" simulations, in particular time stepping (co-simulation).
- Dr.-Ing. Annika Radermacher (www) (email) (mechanics)
former institution: Institut für Angewandte Mechanik, RWTH Aachen, Germany
- Dipl.-Math. Judith Schneider (mathematics)
former institution: Max Planck Institute for the Dynamics of Complex Technical Systems, Magdeburg, Germany
The financial support of the CoSiMOR scientific network by the German Research Foundation (DFG) under grants DFG-FR2702/4, DFG-FR2702/7 is highly acknowledged.