Statistics Group

GRK 1932: Stochastic models for innovations in the engineering sciences

The Deutsche Forschungsgemeinschaft (DFG) has decided to implement a new research training group (RTG) "Stochastic Models for Innovations in the Engineering Sciences" at the University of Kaiserslautern for the period April 2014 -September 2018.

The RTG focuses on innovative interdisciplinary projects which aim at the development, implementation and validation of new stochastic models, algorithms  and methods for applications in the engineering sciences. Possible applications are in the areas of stochastic production processes, the system-on-chip-design, the hardware acceleration of Monte Carlo mehtods with applications in financial mathematics, image analysis for caharacterization of multi-phase materials as well as stochastic modelling and analysis of fiber-reinforced concrete.

DFG Project : Bubble distributions as markers for deformation in polar ice

The analysis of air entrapped in bubbles in polar ice cores yields valuable information on the composition of the atmosphere during the last centuries. However, exploitation of this source of information is hampered by the lack of absolute dating tools linking the ice and the observed air to a certain period in time. Recent dating approaches rely on models which require a measure for the cumulated deformation of the ice as input. Up to now, there is no means for obtaining this information directly from an observed ice core. In the proposed project, we suggest using the locations of bubble centres as markers for the deformation. The main goal is to establish methods for the direct estimation of deformation parameters from the point pattern of bubble centres. To this end, our interdisciplinary research group will combine approaches from glaciology and point process statistics.The locations of bubble centres will be measured by means of core-scale X-ray-microfocus computer tomography on ice samples from different deep ice cores (EDC, EDML, Renland). The main idea of our analysis is based on the assumption that deformation introduces an anisotropy in the bubble centre patterns. Hence, estimates for the main axes and the degree of anisotropy may provide information on the main deformation directions and the strength of deformation. Although point process statistics is an active field of current research, suitably general methods for the anisotropy analysis of 3D point patterns are not available up to now. Hence, the glaciological problems posed in this project ask for advancements in the field of spatial statistics. In particular, methods for the estimation of preferred directions and deformation parameters in 3D point processes have to be found. Additionally, the presence of noise bubbles forming during relaxation of the ice must be taken into account. Using the estimated deformation parameters, models from glaciology for prediction of deformation parameters and dating of the ice will be validated and improved. The measured cumulated strain will give insights to the potential effect of microstructure (grain size, texture and fabric, impurity content) on the deformation behaviour of ice. The analysis of bubbly ice from the Renland ice core will provide information on deformation close to bedrock of a polar icecap for the first time.



Project partner:

Dr. Johannes Freitag, Alfred-Wegener-Institut für Polar- und Meeresforschung, Bremerhaven


The separation of cells in cell suspensions is e.g. important in stem cell therapy and the diagnosis or therapy of blood cancer. Previously established techniques such as centrifugation and fluorescence-based flow cytometry are limited in their throughput and in part mechanically stressful for the cells. In addition, they can not be influenced by specific ligands in their selectivity. Cell chromatography is a promising alternative that allows different cell types in a suspension to be separated by a relatively gentle process. As part of a therapy based on cell chromatography could be taken from the patient blood and returned immediately after the chromatographic filtering. An important step in the development of chromatography-based therapies is the selection of suitable filter media. An experimental investigation of a variety of different porous media with regard to their chromatographic efficiency is associated with high technological complexity. A virtual design of the filters to achieve high chromatographic efficiency could significantly accelerate the development of chromatographic filters. The aim of this project is therefore to obtain an understanding of relationships between the geometric microstructure of a filter medium and its chromatographic efficiency by combining stochastic microstructure modeling with methods for flow simulation.

 Project partner:

Completed projects

Junior Endowed Professorship "Statistics of Spatial Structures for Innovations in Engineering Disciplines" of the Carl Zeiss Foundation, 2013-2017

• DFG-Project Stochastic modeling and calculation of the probability of failure of metal foams on the mesoscale 2009-2011

• DFG-Project Multi-scale stochastic calculation of the natural frequencies of metal foam beams on the basis of stochastic geometry models 2013-2015

• BMBF project "Stochastic Models for the Analysis of Highly Porous Micro- and Nanostructures (AMiNa)", 2010-2013

• BMBF project "Analysis of Low-Dimensional Structures in Three-Dimensional Image Data (AniS)", 2013-2016

• (CM) ² Project "Image Processing in Civil Engineering", 2008-2013


 . Project "Applied System Modeling for multi scale materials" as part of the innovation centre  "Applied System Modeling", 2010-2013

Zum Seitenanfang