Research Group of Prof. Dr. M. Griebel
Institute for Numerical Simulation

MULTIMAT - Multiscale modeling and characterization of phase transformation in advanced materials

Many applications of so-called smart materials are based on changes in their atomic, nano- and microstructure brought about by phase transformations. In order to further develop and apply these materials we need a better and fundamental understanding of the underlying principles of these processes. MULTIMAT aims to continue existing and fruitful collaborations between strong theoretical and experimental research groups active in this exciting and promising field of advanced materials.

MULTIMAT is a Marie Curie Research Training Network (RTN) funded by the European Union. Many teams mainly from Europe contribute to the MULTIMAT network, see the team list on the homepage of the MULTIMAT network. The team from the University of Bonn consists of Within our research group, Prof. Dr. Michael Griebel and Dr. Marcel Arndt are involved in the following projects of the MULTIMAT network:

The macroscopic behavior of material is governed by effects on the atomic level and even finer length scales. Therefore a precise description of the material behavior needs to be targeted on the full hierarchy of scales. To this end, it is essential to study the relationship of the different length scales. It is necessary to develop both analytical and numerical techniques to bridge the different length scales. Several aspects of this topic are addressed by our two projects within the MULTIMAT network. As an application, we study the behavior of shape memory alloys, which exhibit an interesting multiscale behavior.

Project "Multiscale numerical algorithms: Coupling of different physical models on the micro scale and the macro scale"

The goal of this project is to develop numerical techniques in which models from different length scales are coupled numerically, that means to deal with different models on different scales simultaneously within a single simulation. Instead of using a single model on a single length scale, different models on different length scales are integrated in one simulation. The underlying principle is to use the macro model whereever it is applicable, and to fall back to the more precise but computationally costly micro model in regions where the macro model fails or where its accuracy is insufficient. This way, numerical simulations can be efficiently performed which are too costly on the microscale and which involve effects which cannot be described on the macroscale.

Project "Computational modeling of multiscale materials: Derivation of continuum mechanical models from atomistic models"

In this project, analytical and computational techniques are developed to derive models on larger length scales from models on finer length scales. The rigorous derivation allows for models which are more accurate than conventional phenomenological models. Especially we focus on deriving high-accuracy continuum mechanical models from atomistic models. Additionally, we compare the achieved results with experimental data to verify the newly developed analytical and numerical techniques.