Latest Results Gauss Centre for Supercomputing e.V.

LATEST RESEARCH RESULTS

Find out about the latest simulation projects run on the GCS supercomputers. For a complete overview of research projects, sorted by scientific fields, please choose from the list in the right column.

Computational and Scientific Engineering

Principal Investigator: Karine Truffin, Institut Carnot IFPEN Transports Energie, Energies Nouvelles, Rueil-Malmaison (France)

HPC Platform used: JUWELS of JSC

Local Project ID: pra102/DNS4ICE

Today, car manufacturers rely on CFD tools to design and optimise spark-ignition engines. However, current models of turbulent combustion—which are built based on the assumptions of the flamelet regime—lose their predictivity when used to simulate a highly diluted or ultra-lean combustion involving high turbulent intensities. Yet the combustion in a diluted boosted spark-ignition engine shifts from the flamelet to the thin reaction zone (TRZ) regime. This research project performed direct numerical simulations of premixed C8H18/air statistically flat flame interacting with a turbulent flow field. Results were analysed to develop a combustion model suitable for combustion in the TRZ regime based on the formalism of the coherent flame model.

Astrophysics

Principal Investigator: Minna Palmroth, Finnish Centre of Excellence in Research of Sustainable Space, Helsinki (Finland)

HPC Platform used: Hawk at HLRS

Local Project ID: SIMPLE

Space is the finest plasma laboratory one can reach, hence many of the fundamental and universal physics discoveries of to the fourth state of matter – plasma – root to space physics. The near-Earth space is the only place one can send spacecraft to study the variability of plasma ranging from meters to millions of kilometres and from milliseconds to hundreds of years. However, one can send only a few satellites on a few orbits, making near-Earth space environment modelling crucial. To model the near-Earth space accurately, one requires a good resolution for the 3D position space, and additional 3D space for particle distributions— demanding computing performance that easily can reach the limits of any available supercomputer. 

Materials Science and Chemistry

Principal Investigator: Heiko Briesen, Ekaterina Elts, Anthony Reilly, Chair of Process Systems Engineering, Technical University of Munich (Germany)

HPC Platform used: SuperMUC and SuperMUC-NG of LRZ

Local Project ID: pr58la

Solution crystallization and dissolution are of fundamental importance for science and industry. In this project, molecular dynamics simulations were used to study these processes at the molecular scale. By following the motion of molecules towards and away from the crystal surface over short periods of time the intrinsic kinetic behavior that governs the growth and dissolution can be extracted. The obtained information is then used for parametrization of other methods such as kinetic Monte Carlo and continuum simulations to study the dynamics of the crystal surface from the nanoscale up to the microscale and beyond, where the theoretical results would be industrially relevant and easily comparable to experimental results.

Astrophysics

Principal Investigator: Luciano Rezzolla, Institute for Theoretical Physics, Goethe University Frankfurt (Germany)

HPC Platform used: SuperMUC-NG of LRZ

Local Project ID: pn56bi

This ongoing project aims at investigating the long-term evolution of a merging binary system of two neutron stars. The investigation conducted within this project is well aligned with the past research conducted by the Relastro group in Frankfurt and is motivated by the gravitational-wave detection GW170817 and its electromagnetic counterpart, the
so-called kilonova. This kilonova signal is produced by the nuclear processes within the dense and neutron rich mass that is ejected during the merger. Since a lot of mass is ejected during the longterm postmerger evolution, it is crucial to investigate this part via state-of-the-art simulations in order to fully understand the observation.

Computational and Scientific Engineering

Principal Investigator: Christian Stemmer, Stefan Hickel, Technische Universität München, Fakultät für Maschinenwesen

HPC Platform used: SuperMUC and SuperMUC-NG of LRZ

Local Project ID: pr45tu

Deceleration of a supersonic flow in a channel by shocks and interaction with the turbulent boundary layer leads to the formation of a complex array of shocks, subsonic and supersonic regions, and recirculation zones. In this project, high-fidelity and well-resolved large-eddy simulations (LES) of such a fully turbulent (Reδ≈105) pseudo-shock system were performed and compared with experimental data. Particular attention is paid to the occurrence of flow instabilities (such as shock motion, shock-boundary layer interaction, and symmetry breaking of the shock system), mixing behaviour in the transonic shear layer, and a comparison with sophisticated RANS turbulence models.

Materials Science and Chemistry

Principal Investigator: Alfred Kersch, Department of Applied Sciences and Mechatronics, Munich University of Applied Sciences (Germany)

HPC Platform used: SuperMUC and SuperMUC-NG of LRZ

Local Project ID: pr27su

Leveraging the computing power of HPC systems SuperMUC and SuperMUC-NG hosted at LRZ, researchers of the Munich University of Applied Sciences investigated the piezoelectric properties of ferroelectric hafnia and zirconia, which represent a novel material class based on the fluorite crystal structure. If properly doped, such thin films show large strain effects in field induced phase transitions. A large number of doped supercells were investigated with density functional theory to find the most appropriate dopants.

Computational and Scientific Engineering

Principal Investigator: Jian Fang, Scientific Computing Department, STFC Daresbury Laboratory, UK

HPC Platform used: Hawk of HLRS

Local Project ID: FlowCDR

Micro-scale directional grooves with spanwise heterogeneity can induce large-scale vortices across the boundary layer, which is of great importance to both theoretical research and industrial applications. The direct numerical simulation approach was adopted in this project to explore flow structure and control mechanism of convergent-divergent (C-D) riblets, as well as the impact of their spacing, wavelength and height. The results show that the C-D riblets produce a well-defined secondary flow motion characterised by a pair of weak large-scale counter-rotating vortices. This roll mode can play a key role in supressing separation when the flow undergoes adverse pressure gradients, but it may also lead to the increase of friction drag.

Environment and Energy

Principal Investigator: Prabhakar Shrestha, Institute of Geosciences, Meteorology Department, University of Bonn

HPC Platform used: JUWELS of JSC

Local Project ID: chbn33

Clouds and precipitation are the major source of uncertainty in numerical predictions of weather and climate. A common analysis of polarimetric radar observations and synthetic radar data from numerical simulations provides new methods to evaluate models. Using the Terrestrial Systems Modeling Platform, researchers conducted ensemble simulations for multiple summertime storms over north-western Germany. The simulated cloud processes were compared in the radar space using a forward operator with the measurements from X-band polarimetric radars. In addition, sensitivity studies were conducted using different background aerosol states and land cover types in the model to better understand land-aerosol-cloud-precipitation interactions.

Environment and Energy

Principal Investigator: Sandro Jahn, Institute of Geology and Mineralogy, University of Cologne (Germany)

HPC Platform used: JUWELS of JSC

Local Project ID: chpo15

Geological processes are generally quite complex and occur under a wide range of thermodynamic conditions. The structure and the properties of crystalline and non-crystalline phases in the Earth’s interior are often not accessible directly and must be investigated by experiments and by numerical simulations. In this project, we use predictive molecular simulation approaches to establish relations between structural properties of relevant phases, in particular oxide and silicate glasses and melts and aqueous fluids, at high temperatures and high pressures and their respective thermodynamic and physical properties.

Computational and Scientific Engineering

Principal Investigator: Harald Köstler, Chair for System Simulation, Friedrich-Alexander-Universität Erlangen-Nürnberg

HPC Platform used: SuperMUC and SuperMUC-NG of LRZ

Local Project ID: pr86ma

The open-source software framework waLBerla provides a common basis for stencil codes on structured grids with special focus on computational fluid dynamics with the lattice Boltzmann method. Other codes that build upon the waLBerla core are the particle dynamics module MESA-PD and the finite element framework HYTEG. Various contributors have used waLBerla to simulate a multitude of applications, such as multiphase fluid flows, electrokinetic flows, phase-field methods and fluid-particle interaction phenomena. The software design of waLBerla is specifically aimed to exploit massively parallel computing architectures with highest efficiency.

Life Sciences

Principal Investigator: Martin Zacharias, Department of Physics, Technical University Munich

HPC Platform used: SuperMUC and SuperMUC-NG of LRZ

Local Project ID: pr27za

Most biological functions are mediated by conformational changes and specific association of protein molecules. Atomistic simulations are ideal to study the molecular details of such systems. However, often the associated timescales are beyond the maximum simulation times that can be reached even on supercomputers. In this project, researchers developed and tested advanced sampling simulations to accelerate protein domain motions and association of partner molecules. These techniques allow to study domain motions and association of protein molecules on currently accessible time scales. They were successfully applied to study the Hsp90 chaperone protein and to several protein-protein and protein-peptide systems of biological importance.

Life Sciences

Principal Investigator: Christine Peter, Computational and Theoretical Chemistry, University of Konstanz

HPC Platform used: JUWELS of JSC

Local Project ID: chkn01

The conformations of ubiquitin chains are crucial for the so-called ubiquitin code, i.e. the selective signaling of ubiquitylated proteins for different fates in the eukaryotic cellular system. Extensive molecular dynamics simulations at two resolution levels were carried out for ubiquitin di-, tri- and tetramers of all possible linkage types. Analyzing the resulting, exceedingly large high-dimensional data sets was made possible by combining highly efficient neural network based dimensionality reduction with density based clustering and a metric to compare conformational spaces. The so obtained conformational characteristics of ubiquitin chains could be correlated with linkage-type and chain-length dependent experimental observations.

Environment and Energy

Principal Investigator: Michael Bader(1), Alice-Agnes Gabriel(2), (1)Technical University of Munich, (2)Ludwig-Maximilians-Universität München

HPC Platform used: SuperMUC and SuperMUC-NG of LRZ

Local Project ID: pr45fi

In the framework of the ASCETE (Advanced Simulation of Coupled Earthquake and Tsunami Events) project, the computational seismology group of LMU Munich and the high performance computing group of TUM jointly used the SuperMUC HPC infrastructures for running large-scale modeling of earthquake rupture dynamics and tsunami propagation and inundation, to gain insight into earthquake physics and to better understand the fundamental conditions of tsunami generation. The project merges a variety of methods and topics, of which we highlight selected results and impacts in the following sections.

Computational and Scientific Engineering

Principal Investigator: Stefan Platzer, Institute of Helicopter Technology, Technical University of Munich

HPC Platform used: SuperMUC-NG of LRZ

Local Project ID: pn56lu

Rotorcraft are regularly operating in ground effect over moving ship decks or on hillsides. However, only a very limited amount of research has been done to investigate the complex three-dimensional flow fields in these flight conditions and the resulting changes in rotor performance. Therefore, a hovering rotor in non-parallel ground effect was simulated in this project. URANS CFD simulations were made using various turbulence models to gain insight into the three-dimensional flow field, the rotor tip vortex evolution and the velocity distribution in the rotor plane. Best agreement with available experimental data was seen with a Reynolds stress model. Overall, the flow field was most affected close to the rotor hub and on the uphill side.

Environment and Energy

Principal Investigator: Michael Bader(1), Alice-Agnes Gabriel(2), (1)Technical University of Munich, (2)Ludwig-Maximilians-Universität München

HPC Platform used: SuperMUC and SuperMUC-NG of LRZ

Local Project ID: pr48ma

The ExaHyPE SuperMUC-NG project accompanied the corresponding Horizon 2020 project to develop the ExaHyPE engine, a software package to solve hyperbolic systems of partial differential equations (PDEs) using high-order discontinuous Galerkin (DG) discretisation on tree-structured adaptive Cartesian meshes. Hyperbolic conservation laws model a wide range of phenomena and processes in science and engineering – together with a suite of example models, an international multi-institutional research team developed two large demonstrator applications that tackle grand challenge scenarios from earthquake simulation and from relativistic astrophysics.

Elementary Particle Physics

Principal Investigator: Francesco Knechtli, Bergische Universität Wuppertal

HPC Platform used: SuperMUC and SuperMUC-NG of LRZ

Local Project ID: pn56fo

Quantum Chromodynamics (QCD) is the theory of strong interactions. It explains how quarks and gluons form the composite particles called hadrons which are observed in nature. Hadrons can be studied by means of computer simulations of QCD discretized on a Euclidean lattice. This project focuses on hadrons formed by heavy quarks. The question addressed is the relevance of including virtual charm-quark effects in lattice QCD simulations. This dynamics is challenging since it requires small values of the lattice spacing for reliable extrapolations to zero lattice spacing. It is found that its effects are at the sub-percent level even for quantities like the decay constants of charmonium at an energy scale of about half of the proton mass.

Astrophysics

Principal Investigator: Daniel Seifried, I. Physikalisches Institut, Universität zu Köln

HPC Platform used: SuperMUC and SuperMUC-NG of LRZ

Local Project ID: pr94du

Researchers investigated the formation and evolution of molecular clouds, i.e. the nurseries of star formation, by means of 3D magneto-hydrodynamical simulations. These molecular clouds, which are embedded in a galactic disk like our Milky Way, were modelled with a high spatial resolution using a smart zoom-in approach relying on the adaptive mesh refinement technique. As the modelled molecular clouds were embedded in a realistic astrophysical environment, it was possible to study their detailed evolution, e.g. the impact of supernova explosions and radiation from nearby massive stars. Moreover, the research team modelled the chemical evolution of these clouds as well as their dynamics and complex internal structure.

Elementary Particle Physics

Principal Investigator: Kálmán Szabó, University of Wuppertal and Forschungszentrum Jülich, Germany

HPC Platform used: JUWELS of JSC, SuperMUC and SuperMUC-NG of LRZ

Local Project ID: chfz04, pn56bu

Today, large-scale computations of lattice QCD can easily reach a precision of 1% and below—a level at which it is necessary to factor in isospin breaking arising from 1. the presence of the electromagnetic interaction, and 2. the mass difference between up and down quarks. The most prominent consequence of these effects is the mass difference of the neutron and the proton as its numerical value influences the stability of matter: were this difference a bit different from what is measured in experiments, matter would become unstable so that no atoms, molecules and more complex structures could be formed. It was successfully demonstrated that this mass difference can be computed in a common lattice framework of a full QCD + QED calculation.

Computational and Scientific Engineering

Principal Investigator: Philip Ströer, Anthony D. Gardner, Kurt Kaufmann, Institute of Aerodynamics and Flow Technology, German Aerospace Center (DLR), Göttingen

HPC Platform used: SuperMUC and SuperMUC-NG of LRZ

Local Project ID: pr83su

Investigations of different approaches to transition modelling on rotors were undertaken, including comparison to experimental data and results of other European CFD codes. For flows at Reynolds numbers below 500,000 the transition transport models predict unphysically large areas of laminar flow compared to the experimental data. A new boundary layer transition model was developed to improve the transition prediction for a wide range of parameters crucial to external aerodynamics. The new model was implemented into the DLR TAU code and works on either structured or unstructured grids. The agreement of the new model with the experimental data is significantly improved compared to the results of the basic transition transport model.

Computational and Scientific Engineering

Principal Investigator: Sahin Yigit, Josef Hasslberger, Markus Klein, Numerical Methods in Aerospace Engineering, Bundeswehr University Munich

HPC Platform used: SuperMUC and SuperMUC-NG of LRZ

Local Project ID: pn56di

This project focuses on the modelling and physical understanding of 3D turbulent natural convection of non-Newtonian fluids in enclosures. This topic has wide relevance in engineering applications such as preservation of canned foods, polymer and chemical processing, bio-chemical synthesis, solar and nuclear energy, thermal energy storages. Different aspects of non-Newtonian fluids have been analysed in the course of this work: The behaviour of yield stress fluids in cubical enclosures, 2D and 3D Rayleigh-Bénard convection of power-law fluids in cylindrical and annular enclosures and finally the investigation of Prandtl number (Pr) effects near active walls on the velocity gradient and flow topologies.

For a complete list of projects run on GCS systems, go to top of page and select the scientific domain of interest in the right column.