Neue konzeptionelle Ansätze zur Modellierung und Simulation komplexer Systeme / New Conceptual Approaches to Modeling and Simulation of Complex Systems

 

Bewilligungen / Grants 2011

Drittes Statussymposium "New trends in simulating biological systems and soft matter"

22.02.2012 - 24.02.2012 in Potsdam

Technische Universität Berlin
Institut für Theoretische Physik
Sekr. EW 7-1
Prof. Dr. Holger Stark

Humboldt-Universität Berlin
Institut für Physik
Prof. Dr. Lutz Schimansky-Geier

Freie Universität Berlin
FB Biologie, Chemie, Pharmazie
Institut für Chemie
Prof. Dr. Ernst-Walter Knapp

Ansprechpartner:
Technische Universität Berlin
Institut für Theoretische Physik
Sekr. EW 7-1
Prof. Dr. Holger Stark
Hardenbergstr. 36
10555 Berlin
Tel.: 030 314 29623
Fax: 030 314 21130

* * * *

Hydrogen-bond network in protein folding

Bewilligung:  24.03.2011  Laufzeit:  3 Jahre

Das Vorhaben wurde am 30.06.2011 von Garching nach Berlin umgesetzt.

Detailed atomistic models of proteins can only be used to study short times, and thus can involve only very limited conformational changes in large systems. They are also of little use for understanding general principles of proteins and how complex macroscopic properties emerge from elementary interactions. Coarse grained models are used to overcome these problems. However, such models fail to describe many properties of proteins, like the phase diagram of these macromolecules and the forces favoring the formation of secondary structures. The project´s goal is to overcome these limitations and to construct a protein model that contains the correct thermodynamics and topology of real proteins, but is still simple enough to reveal general principles. The approach differs from previous attempts in the use of the hydrogen bonding network as the framework within which the protein folding problem will be formulated.

Freie Universität Berlin
Fachbereich Physik
Dr. Cristiano Dias
Arnimallee 14
14195 Berlin

Freie Universität Berlin
Fachbereich Physik
Dr. Markus Miettinen
Arnimallee 14
14195 Berlin

* * * *

Signal propagation in networks of cortical neurons

Bewilligung: 18.03.2011  Laufzeit:  3 Jahre

How do neurons within recurrent cortical networks respond to dynamical signals? Can networks of threshold neurons describe the activity of networks in vitro? These important, yet unresolved, questions will be the main focus of this fellowship. A combination of analytical and numerical techniques within the theoretical framework of correlated Gaussian processes is used for numerical network simulations with a predefined structure and interactions. During a two-year stay at the Center for Theoretical Neuroscience, Columbia University, New York, the focus is on signal processing in recurrent networks of threshold neurons. The interaction of stimulus evoked and intrinsic activity is studied in order to reveal the biophysical basis of signal relay in cortical neurons. Back in Germany, the opportunity to directly probe the theoretically predicted dynamical network responses in a closed-loop opto-genetically controlled network is seized. The expected contribution refers to the computational side. Analytical model predictions are derived for validating and refining the model assumptions.

Max-Planck-Institut für experimentelle Medizin, Göttingen
Abt. für Molekulare Biologie neuronaler Signale
Dr. Tatjana Tchumatchenko
Hermann-Rein-Straße 3
37075 Göttingen
Tel.: 0551 3899 0

* * * *

Advanced computational methods for quantum physics

Bewilligung: 18.03.2011  Laufzeit:  2 Jahre

The project aims at bridging the gap that traditionally separates algorithm development from applications in quantum physics. An improved computation of the Density Functional Theory (DFT) would allow to simulate larger and more complex materials and biocompounds. To this end, a new computational concept to solve for the sequence of generalized eigenproblems at the core of DFT methods is investigated. A second focus lies on the creation of a method for reverse simulations that will eventually allow the improvement of the current DFT mathematical modeling. The idea is to exploit patterns arising from the analysis of physical simulations to infer the shortcomings of the DFT function basis set and to eventually propose tailored improvements. A redesign of the DFT self-consistent numerical cycle is needed in order to take advantage of repetitive patterns in the eigenpencil (Hamiltonian and Overlap) matrix structure and eigenvectors evolution. High-performance libraries for standard and generalized eigenproblems will be established that exploit current and future parallel architectures. The resulting self-consistent routines are tailored for new multi-core processors.

Forschungszentrum Jülich GmbH
Institute for Advanced Simulation (IAS)
Jülich Supercomputing Center (JSC)
Edoardo Di Napoli, Ph.D.
Wilhelm-Johnen-Straße
52428 Jülich

* * * *

Extreme Ocean Gravity Waves: Analysis and Prediction on the Basis of Breather Solutions in Nonlinear Evolution Equations

Bewilligung: 04.02.2011  Laufzeit: 3 Jahre

Extreme ocean gravity waves are a mysterious phenomenon of giant waves appearing in the deep ocean seemingly from nowhere. They tend to disappear without a trace just the way they appeared. Their existence in the world ocean is a well established fact but present understanding is poor. One of the theories that may explain these extreme waves is based on breather solutions of nonlinear wave equations. Generation of breathers and their combinations can be a reason for the formation of extreme waves. Thus, modeling of such solutions may lead to understanding the nature of this phenomenon. Modeling of breather solutions of the Nonlinear Schrödinger Equations will allow to describe these waves mathematically, possibly predict their appearance and perhaps eliminate them at strategically important locations, e.g. offshore structures. These results are compared with direct numerical simulations of the fully nonlinear Euler equations and their weakly nonlinear approximations. In addition, data based tools to anticipate rogue waves are investigated.

Technische Universität Hamburg-Harburg
Institut für Mechanik und Meerestechnik
Prof. Dr. Norbert Hoffmann

Russian Academy of Sciences, Novgorod
Institute of Applied Physics
Prof. Dr. Efim Pelinovsky
RUSSLAND

Australian National University, Canberra
School of Physics and Engineering
Institute of Advanced Studies
Optical Sciences Group
Prof. Dr. Nail Akhmediev
AUSTRALIEN

Universität Oldenburg
Institut für Physik
ForWind - Zentrum für Windenergieforschung
Prof. Dr. Joachim Peinke

Ansprechpartner:
Technische Universität Hamburg-Harburg
Institut für Mechanik und Meerestechnik
Prof. Dr. Norbert Hoffmann
Eißendorfer Straße 42
21073 Hamburg
Tel.: 040 428783120
Fax: 040 428782028
Homepage: www.tu-harburg.de/mum/startseite.html

* * * *

Recurrent extreme events in spatially extended excitable systems: Mechanisms of their generation and termination

Bewilligung: 04.02.2011  Laufzeit: 3 Jahre

The aim of this project is to develop a theoretical understanding of the generation and termination of very intense but localized structures in spatially extended excitable systems with applications to two particular extreme events, namely harmful algal blooms in the ocean and epileptic seizures in the human brain. The mechanisms and conditions relevant for the generation and termination of extreme events will be evaluated depending on intrinsic local excitability properties, on spatial coupling properties, and on heterogeneities in space. Besides the usual excitable models a novel type of excitable activator-inhibitor systems is introduced in which several activators are competing with each other. Furthermore time series analysis techniques will allow a robust quantification of complex interactions in such systems even for short data sets. Real data sets are used to verify pathways of generation and termination of extreme states identified through the model.

Universität Oldenburg
Institut für Chemie und Biologie des Meeres
Theoretische Physik/Komplexe Systeme
Prof. Dr. Ulrike Feudel

Universität Oldenburg
Institut für Chemie und Biologie des Meeres
Planktologie
Prof. Dr. Helmut Hillebrand

Universität Oldenburg
Institut für Chemie und Biologie des Meeres
Planktologie
Dr. Stefanie Moorthi

Max-Planck-Institut für Physik
komplexer Systeme, Dresden
AG Nichtlineare Dynamik und Zeitreihenanalyse
Prof. Dr. Holger Kantz

Potsdam-Institut für
Klimafolgenforschung e. V.
Abt. IV:
Transdisziplinäre Konzepte und Methoden
Prof. Dr. Jürgen Kurths

Universität Bonn
Klinik für Epileptologie
AG Neurophysik
Prof. Dr. Klaus Lehnertz

Ansprechpartner:
Universität Oldenburg
Institut für Chemie und Biologie des Meeres
Theoretische Physik/Komplexe Systeme
Prof. Dr. Ulrike Feudel
Postfach 25 03
26111 Oldenburg
Tel.: 0441 7982790
Fax: 0441 7983404
Homepage: http://www.icbm.de/~feudel

* * * *

Thermal Runaway of Lithium Batteries

Bewilligung: 04.02.2011  Laufzeit: 3 Jahre

The thermal runaway of lithium batteries is an outstanding example for an extreme event in a complex technical system. Thermal runaway refers to the ignition or explosion of a battery as consequence of a self-accelerating heating process that is initiated by stochastically occurring internal or external triggers. The corresponding safety problems are today a major factor impeding the introduction of lithium batteries for electromobility or storage of energy from sustainable sources. The objective of the present study is to understand, predict and control thermal runaway in lithium batteries. A combined methodology of multi-scale deterministic modeling, stochastic simulation, and experimental analyses is proposed. This will allow probabilistic predictions of thermal runaway events and, particularly, the design of an early-alert and risk-aware heat management system for lithium batteries.

Deutsches Zentrum für Luft-
und Raumfahrt e.V., Stuttgart
Institut für Technische Thermodynamik
Priv.-Doz. Dr. Wolfgang Georg Bessler

Universität Stuttgart
Institut für Wasserbau
Jun.-Prof. Dr.-Ing. Wolfgang Nowak

Zentrum für Sonnenenergie- und
Wasserstoff-Forschung (ZSW), Ulm
Dr. Harry Döring

Ansprechpartner:
Deutsches Zentrum für Luft-
und Raumfahrt e.V., Stuttgart
Institut für Technische Thermodynamik
Priv.-Doz. Dr. Wolfgang Georg Bessler
Pfaffenwaldring 38 - 40
70569 Stuttgart
Tel.: 0711 6862603
Fax: 0711 6862747
Homepage: www.bessler.info

* * * *

Mesoscale Weather Extremes: Theory, Spatial Modeling and Prediction (WEX-MOP)

Bewilligung:  04.02.2011  Laufzeit:  3 Jahre

Das Vorhaben wurde am 24.01.2012 von Göttingen nach Mannheim umgesetzt.

Mesoscale weather systems, such as convective cells or severe fronts, are responsible for numerous hazardous weather events that are caused by heavy rainfall and/or extreme winds. The physical understanding, modeling and prediction of their characteristics lie at the heart of this interdisciplinary project, in which atmospheric and mathematical scientists join forces with the German Weather Service. The project seeks an improved understanding of energy cascades and physical constraints for the spatial structure of mesoscale extremes. An ensemble copula coupling approach for physically consistent probabilistic forecasts across space, time and weather variables is developed as well as design verification tools for probabilistic forecasts of extreme events. Novel probabilistic models for spatial fields of extremes, as well as methodology and algorithms for conditional simulation are introduced. As an overarching theme, the project contributes to the development of a next-generation mesoscale probabilistic forecast system for extreme weather.

Universität Mannheim
Wirtschaftsinformatik und Wirtschaftsmathematik
Institut für Mathematik
Prof. Dr. Martin Schlather
Universität Bonn
Meteorologisches Institut und Interdisziplinäres
Zentrum für Komplexe Systeme
Priv.-Doz. Dr. Petra Friederichs

* * * *

Projections and predictions of Local precipitation Intensities. Advanced Downscaling using Extreme value Statistics (PLEIADES)

Bewilligung: 04.02.2011  Laufzeit: 2 Jahre

Robust knowledge on regional scales about future changes in precipitation extremes is crucial for planning adaptation to climate change. As global climate models do not provide reliable information below scales of about 200 km, dynamical or statistical postprocessing is required. However, such downscaling approaches are currently limited in their representation of precipitation extremes. The project aims to substantially improve downscaling methodologies by developing a statistical downscaling and correction method for precipitation simulated in regional and global climate models. An estimation of the full precipitation distribution, including extreme events and spatial dependence is seeked that is able to correct location biases. An ensemble of scenarios for precipitation and its extremes for Europe for the entire 21st century will be analysed in order to downscale global decadal predictions that are currently in preparation.

Leibniz-Institut für Meereswissen-
schaften an der Universität Kiel
(IFM-GEOMAR)
FB01: Ozeanzirkulation und Klimadynamik
Prof. Dr. Douglas Maraun

CNRS - Centre National de la Recherche
Scientifique, Gif-sur-Yvette
Laboratoire des Sciences du Climat et
de l'Environnement
Centre d'Etude de Saclay
Dr. Mathieu Vrac
FRANKREICH

University of Birmingham
School of Geography, Earth and Environmental
Sciences
Dr. Martin Widmann
GROSSBRITANNIEN

Ansprechpartner:
Leibniz-Institut für Meereswissenschaften an der Universität Kiel
(IFM-GEOMAR)
FB01: Ozeanzirkulation und Klimadynamik
Prof. Dr. Douglas Maraun
Düsternbrooker Weg 20
24105 Kiel
Tel.: 0431 6004057
Fax: 0431 6004052

* * * *

Robust Risk Estimation

Bewilligung: 04.02.2011  Laufzeit: 3 Jahre

The project aims at a theoretical foundation, development and application of robust procedures for risk management for complex systems in the presence of extreme events, i.e. identification, quantification, prediction and control of these risks. The elaborated method is applied to chosen reference applications. The examples consist of operational risk of a bank, unit length of stay and cost in intensive care of a university clinic, and river discharge data of Austrian rivers. Suitable parametric models for these contexts are developed. The goal is to adapt the shrinking neighborhoods approach to determine optimally-robust estimators minimizing the maximal asymptotic risk on neighborhoods about the ideal model and at the same time resulting in a high breakdown point. Furthermore, corresponding diagnostic tools to quantify and visualize the influence and outlyingness of data are developped.

Fraunhofer-Institut für Techno- und
Wirtschaftsmathematik ITWM, Kaiserslautern
Dr. Peter Ruckdeschel

Technische Universität Kaiserslautern
Fachbereich Mathematik
AG Finanzmathematik
Prof. Dr. Ralf Korn

Hochschule Furtwangen
Hochschule für Technik
Fakultät Maschinenbau und Verfahrenstechnik
Prof. Dr. Matthias Kohl

Universität für Bodenkultur
Wien
Institut für Angewandte Statistik und EDV
Dr. Bernhard Spangl
ÖSTERREICH

Ansprechpartner:
Fraunhofer-Institut für Techno- und
Wirtschaftsmathematik ITWM, Kaiserslautern
Dr. Peter Ruckdeschel
Fraunhofer Platz 1
67663 Kaiserslautern
Tel.: 0631 316004699
Fax: 0631 316005699
Homepage: http://www.mathematik.uni-kl.de/~wwwfm/RobustRiskEstimation/

* * * *

Advanced Simulation of Coupled Earthquake and Tsunami Events (ASCETE )

Bewilligung:  04.02.2011  Laufzeit:  3 Jahre

Das Vorhaben wurde am 24.01.2012 von Stuttgart nach Garching umgesetzt.

The main objective of the ASCETE-project is the joint development of advanced simulation technology via the combination of innovative methodological approaches in earthquake and tsunami modelling with extremely efficient algorithmic concepts from computer science. Through the development of an integrative simulation environment a better understanding of the fundamental conditions of tsunami generation is envisaged, considering fully dynamic rupture processes coupled with advanced hydrodynamic models. New insight into the underlying physics of earthquakes-tsunami interaction responsible for the generation of devastating tsunamis is expected. The main methodological focus is the design of accurate and optimized solvers sharing one joint numerical approach based on the discontinuous Galerkin finite element method. The simulations will include latest models of frictional sliding, information of fault geometries and stress state derived from geodynamic models, as well as detailed ocean bathymetry. The novel modeling technology will provide realistic simulations of geohazards and can help society to mitigate earthquake and tsunami related disasters.

Technische Universität München
Informatik V - Lehrstuhl für Wissenschaftliches
Rechnen
Prof. Dr. Michael Georg Bader

Universität Hamburg
KlimaCampus-Exzellenzzentrum für Klimaforschung
Professur Numerische Methoden in den
Geowissenschaften
Prof. Dr. Jörn Behrens

Universität München
Department für Geo- und Umweltwissenschaften
Geophysik
Prof. Dr. Heiner Igel

Universität München
Department für Geo- und Umweltwissenschaften
Geophysik
Dr. Martin Andreas Käser

Eidgenössische Technische Hochschule
Zürich
Schweizer Erdbebendienst
Dr. Luis A. Dalguer
SCHWEIZ

Kontakt:
Universität Hamburg
KlimaCampus-Exzellenzzentrum für Klimaforschung
Professur Numerische Methoden in den
Geowissenschaften
Prof. Dr. Jörn Behrens
Grindelberg 5
20144 Hamburg
Tel.: 040 428387734
Fax: 040 428387712
Homepage: http://www.ascete.de

* * * *

High-pressure fluid-controlled earthquake cycles and fault networks

Bewilligung: 04.02.2011  Laufzeit: 3 Jahre

The hypothesis is tested that much of the earthquake cycle is substantially controlled by the mechanical, chemical, and time-varying hydraulic behaviour of high pressure fluids trapped at depth at the base of the seismogenic brittle crust. This hypothesized scenario is supported by numerous observations. However, numerical modeling is currently insufficient to properly explore its implications. Theoretical developments and analyses of seismic sequences are necessary to quantify evolving fracture networks, both prior to the earthquake, and of the subsequent aftershock sequences. First, a 3-dimensional numerical model that couples a pressure-dependent permeability with a deforming poro-elastic-plastic rheology is developed and used to quantify the structural and mechanical properties of the active set of faults. Next, a pattern recognition algorithm and rigorous validation schemes for fault and fracture networks are tackled. An investigation of ongoing and future seismic sequences and pattens forms the basis for an improved earthquake hazard assessment.

Universität Bonn
Steinmann-Institut für Geologie, Mineralogie und
Paläontologie
Geodynamik/Geophysik
Prof. Dr. Stephen Andrew Miller
Nußallee 8
53115 Bonn
Tel.: 0228 737430
Homepage: www.geo.uni-bonn.de

Eidgenössische Technische Hochschule Zürich
Departement Management, Technologie und Ökonomie
Chair of Entrepreneurial Risks
Prof. Dr. Didier Sornette
Kreuzplatz 5
CH-8032 Zürich
SCHWEIZ
Tel.: 004163 28917
Fax: 004163 21914
Homepage: www.er.ethz.ch

* * * *

Schriftgröße