High Performance Computing in Science and Engineering '09

We examine models for line-emitting gas nebulae around powerful radio galaxies in the early universe. The models assume that either the emitting gas clouds are embedded in the shocked ambient gas and driven outwards with it or that they are created and sustained by multi-phase turbulence in the jet cocoon. For this, we perform jet simulations with realistic density contrasts on large scales and on typical activity time scales of several ten million years. The employed magnetohydrodynamics code NIRVANA has been optimized for the NEC SX machines in the previous years and now runs very efficiently on the SX-6 and SX-8, allowing us simulations both in axisymmetry and full three dimensions. Future simulations on the SX-9 will benefit significantly from the increased number of shared memory processors per node for the axisymmetric runs, where an MPI-parallelization is generally inefficient for our setup due to communication overhead.

We give an overview of the problems and the current status of our two-dimensional (core collapse) supernova modelling, and present the system of equations and the algorithm for its solution that are employed in our code, and report on our continuing efforts to increase the computational efficiency of our supernova code VERTEX, which now has full MPI functionality and can be run efficiently on several nodes of the NEC SX-8. We also discuss selected results that document the progress achieved by recent simulations that are performed on the NEC SX-8 at the HLRS Stuttgart. In particular, we have studied the influence of the nuclear equation of state on the neutrino and gravitational wave signals from collapsing stellar cores and on the explosion mechanism. Another focus of our work was on the observable signatures of a subclass of supernovae originating from low-mass progenitors below about ten solar masses.

We simulate a suspension of model colloidal particles interacting via DLVO (Derjaguin, Landau, Vervey, Overbeek) potentials. The interaction potentials can be related to experimental conditions, defined by the

p

H-value, the salt concentration and the volume fraction of solid particles suspended in the acqueous solvent. Depending on these parameters, the system shows different structural properties, including cluster formation, a glass-like repulsive structure, or a stable suspension. To explore the parameter space many simulations are required. In order to reduce the computational effort and data storage requirements, we developed a steering approach to control a running simulation and to detect interesting transitions from one region in the configuration space to another. In this article we describe the implementation of the steering in the simulation program and illustrate its applicability by several example cases.

We present ab initio studies towards the adsorption of the amino acid cysteine on the Au(110) surface. By means of density functional theory calculations and using the repeated-slab supercell method, we investigate three main aspects relevant for the adsorption process. First, in order to estimate the slab width required for an accurate description of the gold surface, we calculate the surface energies for both the unreconstructed and the missing row (1×2) reconstructed Au(110) surface for varying slab widths. Then, we determine the formation energies for vacancies in the salient row on the missing-row reconstructed surface. This allows us to estimate the energy cost for a local lifting of the missing-row reconstruction upon molecular adsorption. Finally, we examine the formation of cysteine dimers via carboxyl-carboxyl hydrogen bonds and investigate the changes in the bond energy caused by intermolecular strain. We predict the cysteine dimer formation in Au surface vacancies to play an important role in the adsorption process.

We investigate structural, electronic and optical properties of colloidal IV-VI semiconductor quantum dots (QDs) using an

ab initio

pseudopotential method and a repeated super cell approximation. In particular rhombo-cubo-octahedral quantum dots (QDs) with a pseudo-hydrogen passivation shell consisting of PbSe, PbTe and SnTe are investigated for different QD sizes. The obtained dependence of the confinement energy on the QD size questions the use of 3-dimensional spherical potential well models for very small QD structures.

Large-scale

first-principles

calculations are used to rationalize the formation of well-separated (∼10 Å) molecular rows of phenylglycine upon co-adsorption of adenine and phenylglycine on Cu(110) [Chen and Richardson, Nature Materials 2, 324 (2003)]. It is found that the molecular adsorption leads to longwave oscillations of the charge density at the Cu(110) surface. The experimentally observed indirect interaction between the molecular rows is mediated by these charge fluctuations. Strain effects, in contrast, are of minor importance.

A fuel-cell is an electrochemical equipment in which hydrogen and oxygen would be transformed into water and electric power. As a result, no carbon-containing gas pollutant would be emitted. In the past two decades, the fuel-cell technology has been advanced,[

1

,

2

,

3

] but still there are some challenging problems to be addressed, e.g. the durability of the electrode, oxygen conduction efficiency, and the limited hydrogen storage capacity. The later issue – hydrogen storage – is in the center of interest in this work. The targeted hydrogen-carrying capacity is 6.5 %wt by 2010, as designated by the US Department of Energy.[

4

]

Atomistic simulations of tensile and compressive deformation of three-dimensional nanocrystalline palladium at room temperature and different strain rates were perfomed. Detailed analysis of tensile straining has revealed almost no plasticity and an absence of dislocation activity in the grains right up to the moment of intergranular cracking. During compressive straining the sample exhibits a plastic regime brought about by the motion of extended partial dislocations emitted from the grain boundaries. At higher compressive strains the deformation mechanism changes to one that involves full dislocations and twinning.

Transport properties of strongly interacting quantum systems are a major challenge in todays condensed matter theory. In our project we apply the density matrix renormalization group (DMRG) method [

2

,

3

,

4

,

5

,

6

,

7

,

8

] to study transport properties of quantum devices attached to metallic leads.

There is no consensus in the recent literature how singular the eigenfunctions of quasirelativistic operators are close to a point-like nucleus. This question has far-reaching implications, e.g. for the convergence properties when such eigenfunctions are expanded in a basis of regular functions. For this reason the spectrum of a Douglas-Kroll operator in a large basis of spherical waves has been investigated. Such a basis sets shows a very slow but regular convergence pattern from which information on the singularity at the origin can be extracted. This calculation involves multiplications and a diagonalization of very large dense matrices, which has been performed in parallel (with up to 256 CPU cores) using functions from the

ScaLAPACK

library. For first-order Douglas-Kroll operators the eigenfunctions were known to be more singular than for the Dirac operator, and this manifests itself in our results. Second-order Douglas-Kroll and beyond, however, behaves very similar to the Dirac case, and there is ample evidence that the Douglas-Kroll eigenfunctions beyond first-order Douglas-Kroll have the same Singularity close to the nucleus than the Dirac operator. It is thus likely that an expansion of Douglas-Kroll eigenfunctions in basis sets conventionally used in relativistic quantum chemistry will show essentially the same convergence rate as found e.g. for the Dirac operator.

The multiconfiguration time-dependent Hartree (MCTDH) method is an algorithm to efficiently propagate wavepackets. Since its development it has been established as one of the most important methods to study the quantum dynamics, e.g. excitations of vibrations and rotations or scattering events, of molecular systems. The method has been developed in Heidelberg over the last 18 years and is able to treat large systems with ≳9 degrees of freedom in its single processor application. As described in a previous work [

1

] the code has been parallelized for shared memory hardware and was successfully applied to the investigation of the Zundel cation [

2

,

3

] (15 internal degrees of freedom) using up to 8 processors with speedup factors of >7. However, the Heidelberg MCTDH-code was still lacking the ability to use the more powerful distributed memory hardware of computer clusters. This drawback has been removed recently by applying the Message Passing Interface (MPI) to MCTDH, as described in this report.

Two of the most challenging tasks in molecular dynamics simulation are the simulation of long-range interactions like in electrolyte solutions and the internal degrees of freedom like in hydrogels. Both tasks lead to time consuming simulations with big systems. Therefore, massively parallel high performance computers are needed for both tasks. The first step of the present work was to test and validate different force fields for electrolyte solutions and hydrogels.

In this contribution the three-dimensional reacting turbulent flow field of a swirl-stabilized gas turbine model combustor is analyzed numerically. The investigated partially premixed and lifted CH

4

/air flame has a thermal power load of P

th

=35kW and a global equivalence ratio of

φ

=0.65. To study the reacting flow field the Scale Adaptive Simulation (SAS) turbulence model in combination with the Eddy Dissipation/Finite Rate Chemistry combustion model was applied. The simulations were performed using the commercial CFD software package ANSYS CFX-11.0. The numerically achieved time-averaged values of the velocity components and their appropriate turbulent fluctuations (RMS) are in very good agreement with the experimental values (LDA). The same excellent results were found for other flow quantities like temperature and mixture fraction. Here, the corresponding time-averaged and the appropriate RMS profiles are compared to Raman measurements. Furthermore, instantaneous flow features are discussed. The simulations have been performed on the HP XC4000 system of the High Performance Computing Centre Karlsruhe.

In the present paper the impact of NO

X

-emissions of aircrafts on ozone is highlighted. The chemical processes of NO

X

in the troposphere and stratosphere are explained briefly and the formation of NO

X

under the specific conditions of a scramjet combustor is discussed. Finally a model scramjet combustor with hydrogen injection is investigated numerically using the scientific in-house code TASCOM3D. The results of the simulations demonstrate the formation of NO

X

in scramjet combustors and are discussed in detail. With regard to the development of NO

X

reduction strategies further investigations are performed in terms of the variation of the inlet temperature. Variations of other relevant parameters concerning NO

X

-formation like the equivalence ratio, Mach number at combustor entrance, types of fuel injection, etc. are part of ongoing investigations but will not be discussed in this paper. Finally the performance of the scientific code TASCOM3D on the NEC SX-8 is analyzed.

Lean premixed flames in an un-confined model swirl burner are studied numerically by means of large eddy simulation. In the burner considered, swirling flow and a richer pilot flame are used to stabilize the flame. The paper investigates the occurrence of the central recirculation zone (CRZ) and the precessing vortex core (PVC) for two variants of the burner, as well as the flame response to a sinusoidal variation of the mass flow rate in the pilot jet. The corresponding isothermal swirling flows are also simulated for providing reference. Detailed comparisons with experimental data are performed for velocity statistics and generally good agreement is achieved. Additionally, coherent structures are analyzed and local power spectra computed. It is found that a very strong PVC is observed in numerical results and experiment for the non-reactive case with retracted pilot jet. In the reactive cases where the whole CRZ is enclosed in the high-temperature post-flame region, the PVC is almost completely suppressed.

In this paper we describe the recent development of the interface library

Dune

and in particular the module

Dune-Fem

. We demonstrate that the development of the software package

Dune

and in particular the module

Dune-Fem

has reached a state where

Dune

and

Dune-Fem

can be used for efficient parallel simulations including higher order discretizations and techniques like local grid adaptivity in combination with dynamic load balancing. This is shown by test problems reaching from academical up to more sophisticated problems.

Jet vortex generators have been proven to provide a mechanism to positively control boundary layer flows. The present paper illustrates a method to perform direct numerical simulations (DNS) of a jet actuator flow inside a laminar boundary layer. A structured finite difference method is used for the simulations. The numerical scheme was adapted to account for the large scale differences both in geometric and fluid dynamic aspects. Analytical mesh transformations have been implemented to resolve the jet orifice. Suitable boundary conditions are established to model the jet flow. Arising numerical instabilities have been suppressed by implementing a compact filter scheme. Test simulations are done for jet actuator configurations in laminar baseflow with jet to freestream velocity ratios of up to

R

=3.0. The computational effort on a NEC SX-8/9 is also investigated.

Direct Numerical Simulations (DNS) of flow over and heat transfer to the stagnation region of a circular cylinder have been performed at a Reynolds number of

Re

D

=140 000 - based on the far field velocity and the diameter of the cylinder

D

-. To avoid the formation of a vortex street, behind the cylinder a splitter plate was introduced. The computations were carried out on the NEC SX-8 using up to 192 processors and 746.5 million grid points. Two simulations were performed: A two-dimensional, fully laminar simulation to obtain a base-line solution and a fully three-dimensional simulation to address the influence of oncoming free-stream fluctuations on the heat transfer from the stagnation region of the heated cylinder. The incoming free-stream turbulence stems from a separate simulation of isotropic turbulence in a box. Compared to the fully laminar simulation, the addition of free-stream fluctuations at the inflow plane was found to lead to an increase in heat transfer in the stagnation region of the cylinder by 66%.

To investigate turbulent flows, direct numerical simulations (DNS) is playing more and more important roles benefitting from the modern computing technologies. Once the DNS data are obtained, according to different theory and purposes, the data analysis will be the core of the work at next stages. From the sizable capacity data for homogeneous shear turbulence, the conditional statistics along gradient trajectories have been investigated. It has been derived and also proved numerically that the two-point velocity difference structure functions along the same gradient trajectories have a linear scaling with respect to the arclength between the two points, different from the classical Kolmogorov scaling. In addition, the performance of the OpenMP parallelized code is satisfactory.

The rise behavior of single air bubbles of different size in viscous liquids (water/glycerol mixtures) is investigated using a 3D Direct Numerical Simulation method. The simulations where performed by the ITLR inhouse code Free Surface 3D (FS3D) which solves the incompressible Navier-Stokes equations using a Volume-of-Fluid (VOF) technique. The computational effort is reduced by using a moving frame of reference. This moving frame allows the simulation of a large scale rising trajectory. The numerical results, which comprise the terminal rise velocity, the aspect ratio of the deformed bubbles and the resulting bubble shape are compared to experimental data from literature. The simulations were performed on the NEC SX-8 and NEC SX-9 platforms of the HLRS.

Despite the high interest in unsteady heat transfer phenomena, there is a fundamental lack of experimental and numerical work in this field. Moreover, the development of turbulence heat transfer models rely on the availability and accuracy of near wall data which are very difficult or even impossible to be measured. In recent years DNS has shown the capability to be used as a so called “numerical experiment” and has been applied to fill the gap between models and accurate data. In the present project DNS simulations were performed for a simple channel configuration with heat transfer. For the steady channel flow, the DNS results are validated against experimental data and very fine resolved DNS simulations available in the literature. Results are also shown and discussed for an unsteady pulsating flow. Neither numerical nor experimental results are available for the pulsating case involving heat transfer. In order to address the needs of turbulence modelling of unsteady near wall turbulent scalar transport the budget of the turbulent kinetic energy and its dissipation rate together with the budget of the turbulent temperature variance and its dissipation rate were determined at different time phases and for different values of the amplitude and frequency parameters. Thus, the present DNS database can be highly useful for future unsteady heat transfer closure development.

Recently, the implicit SGS modeling environment provided by the Adaptive Local Deconvolution Method (ALDM) has been extended to Large-Eddy Simulations (LES) of passive-scalar transport. The resulting adaptive advection algorithm has been described and discussed with respect to its numerical and turbulence-theoretical background by Hickel

et al.

, 2007. Results demonstrate that this method allows for reliable predictions of the turbulent transport of passive-scalars in isotropic turbulence and in turbulent channel flow for a wide range of Schmidt numbers. We now use this new method to perform LES of a confined rectangular-jet reactor and compare obtained results with experimental data available in the literature.

To analyze the fundamental physical mechanism which determines the damping effect of a riblet surface on three-dimensional transition several numerical simulations of spatial transition in a flat plate zero-pressure gradient boundary layer above a riblet wall are performed in this study. Two types of transition are investigated. The first type of transition, namely K-type transition, is induced by a dominant two-dimensional Tollmien-Schlichting (TS) wave and a weak spanwise disturbance. The second type of transition is purely excited by two oblique waves. The two-dimensional TS waves are found to be amplified by riblets, whereas three-dimensional structures, i.e., Λ-, hairpin, and streamwisely aligned vortices, are damped and their breakdown to turbulence is delayed compared to transition on a clean surface. The investigation of the near wall flow structure reveals secondary flows induced by the riblets and reduced wall normal ejections as well as a reduced downwash. Overall, especially the oblique transition is delayed by the riblets.

Impinging jets are used in a variety of engineering applications like in chemical reactors, mixing devices, drying and cooling applications. Good quality simulations of this highly complex flow field is a challenging task. In this work, the flow field and heat transfer of turbulent pulsating and swirling jet impingements are investigated by Large Eddy Simulation (LES). The benchmark case of turbulent impinging jet with out swirl and excitation, recommended by ERCOFTAC, is simulated first. In all investigations the jet’s Reynolds number (Re) is 23000 and jet outlet-to-target wall distance (H/D) is 2. The agreement between experimental data and simulation gives encouragement for further investigation of complex flow of jets impingement with excited inlet velocity profiles and swirl. The pulsating jets are investigated at four different excitations modes. Where as the swirl is introduced at two different swirl numbers (S). The correlation between the heat transfer mechanism, flow kinematics and turbulence quantities is investigated.

Lean Premixed combustion, which allows for reducing the production of thermal NOx, is prone to combustion instabilities. There is an extensive research to develop a reduced physical model, which allows - without time-consuming measurements - to calculate the resonance characteristics of a combustion system consisting of Helmholtz-resonator-type components (burner plenum, combustion chamber). For the formulation of this model numerical investigations by means of compressible Large Eddy Simulation (LES) are carried out. In these investigations the flow in the combustion chamber is isotherm, non-reacting and excited with a sinusoidal mass flow rate. The foregoing investigations concentrated on the single combustion chamber as a single resonator.

In this paper the results of the numerical investigations of a coupled system will be presented. The reduced physical model was extended for the coupled system of burner and combustion chamber. By means of numerical simulation and the physical model the resonance characteristics of a combustion system can be predicted already during the design phase. In order to predict the resonant characteristics of the coupled system and to provide an insight into the flow mechanics 10 compressible LES were carried out. The results are in very good agreement with the experimental investigations.

The turbulent flow in two asymmetric diffusers with complex three-dimensional separation was computed employing Large-Eddy Simulations (LES) and Reynolds-Averaged Navier–Stokes (RANS) calculations. The computational setup matches existing experiments in the literature. The objective of the present study is to obtain reference data to be used for assessing the performance of newly developed hybrid LES/RANS techniques.

The largest known direct numerical simulations of grid-generated turbulence were performed, on one hand resolving the grid placed in an incompressible fluid flow while on the other hand following reliably the evolution of flow structure both in the direction of the mean flow and across it well into a region of self-similar behaviour. The excellent scaling of the employed lattice Boltzmann algorithm allowed to perform a series of DNS runs and quantify several parametric effects. Several major qualitaitve results were obtained, as well, having immediate impact on turbulence modeling: A new algebraic law (of Kolmogorv type) was discovered for the decay of turbulent kinetic energy in low-intensity grid-generated turbulence. Furthermore, the assumption of the existence of a single decay rate for the anisotropy of weak “homogeneous” turbulence was shown to be qualitatively inappropriate, while the assumption of an “inertial range” for the spatial development of such turbulence – on which the former assumption should be based – appears to be valid.

For the numerical simulation of journal bearings, current software solutions use the Reynolds differential equation where inertia terms are not included. The Finite Difference Element Method (FDEM) is a black-box solver for nonlinear systems of elliptic and parabolic partial differential equations (PDEs). Based on the general black-box we implement the Reynolds equation with inertia terms for the simulation of a journal bearing. We can easily implement different models for the turbulence factors and the dynamic viscosity, and we also consider cavitation. We give results for different Reynolds numbers, and we also give a global error estimate for each of the cases. This shows the quality of the numerical solution and is a unique feature of FDEM.

A combined experimental as well as computational analysis of a complete scramjet demonstrator model has been initiated. The experimental tests will take place under real flight conditions at a hypersonic wind tunnel. Prior to those tests, a numerical analysis of the performance of two possible demonstrator geometries is conducted. In the current paper, the results of the performance analysis for two different two- and three-dimensional intakes employing single and double outer compression ramps as well as side wall compression are discussed. It is shown that the intakes, in combination with a subsequent isolator, are able to generate flow conditions required for stable supersonic combustion using a central strut injector. Furthermore, the effect of boundary layer transition is also discussed.

In this paper preliminary numerical investigations are presented which have been conducted at the Chair of Mechanics (LFM) of RWTH Aachen University during the course of the new experimental/numerical project Aero-Structural Dynamics Methods for Airplane Design (ASDMAD). The goal was to study the dynamic vibrational decay behaviour of the HIRENASD wing modified by different wingtip devices. After a short introduction to the project specifications the numerical method and model is described. The calculated results are compared with those of the original HIRENASD model, and conclusions are drawn with respect to the differences between the two models.

In its Fourth Assessment Report (AR4) the Intergovernmental Panel on Climate Change (IPCC) states that it is very likely that many land regions, especially on the Northern Hemisphere, will warm during the 21st century due to the general climate change, which itself is caused by the increase of anthropogenic greenhouse gases [

7

]. On the global scale the IPCC estimates that for the next two decades a warming of about 0.2

○

C per decade can be expected on the basis of the SRES (Special Reports on Emission Scenarios, [

11

]) emission scenarios. Even if the concentrations of all greenhouse gases were kept at the year 2000 level, a further warming of about 0.1

○

C per decade can be expected. For many land regions of the globe it is estimated that the annual mean temperature increase will be higher than the global mean. This variability of climatic change persists over all scales, from global to regional, and more or less for all climatological observables. On the spatial scale of global climate models (about 200 km) for example, the largest warming in Europe is likely to happen in the northern part in winter and in the Mediterranean area in summer [

2

]. Annual precipitation is very likely to increase in most of northern Europe and decrease in most of the Mediterranean area. In Central Europe, precipitation is likely to increase in winter but decrease in summer, but the agreement between the results of various models is quite low there. Extremes of daily precipitation are very likely to increase throughout Europe [

1

]; compared to Northern and Southern Europe, however, climatic change for Central Europe is more difficult to assess due to sometimes conflicting tendencies.

The dramatic change in the region of the West African Monsoon (WAM) from wet conditions in the 50s and 60s to much drier conditions from the 70s to the 90s represents one of the strongest inter–decadal signals on the planet in the 20th century. Marked inter-annual variations in the recent decades have resulted in extremely dry years with devastating environmental and socio-economic impacts. Vulnerability of West African societies to climate variability is likely to increase in the next decades as demands on resources increase due to the rapidly growing population. The situation may be exacerbated by the effects of climate change, land degradation, water pollution and biomass burning. Furthermore, the WAM has an impact on the downstream tropical Atlantic by providing the seedling disturbances for the majority of Atlantic tropical cyclones and on the global climate as one of the world’s largest source regions of mineral dust and of fire aerosol. Motivated by the need to develop strategies to reduce the socio–economic impacts of climate variability and change in the WAM, the integrated European project African Monsoon Multidisciplinary Analysis (AMMA) aims to improve our knowledge and understanding of the WAM on daily to interannual timescales and thus improve our ability to forecast the weather and climate in the West African region.

A major challenge for molecular modeling consists in optimizing the unlike interaction potentials. A broad study on fluid mixtures [

1

] recently showed that among the variety of combination rules that were proposed in the past, none is clearly superior. In many cases, all are suboptimal when accurate predictions of properties like the mixture vapor pressure are needed. The well known Lorentz-Berthelot rule performs quite well and can be used as a starting point. If more accurate results are required, it is often advisable to adjust the dispersive interaction energy parameter which leads to very favorable results [

1

,

2

,

3

,

4

,

5

].

We present the basic conception of a new dynamical model of the thermal and chemical evolution of Mars. Therefore new enlargements of the code Terra are necessary which allow to improve the solutions of the convection differential equations with strongly varying viscosity. These enlargements have been partly tested already. We describe considerations on the chronology of the early evolution of Mars and on magma ocean solidification since they lead to a

structural

model of the

early

Mars. This is important as a starting presupposition for a

dynamical

solution of the martian evolution similar to [

122

] which derives the essential features of the Earth’s mantle’s history. At present there is no PREM[39]-analogon neither for the present time nor for the start of the solid-state creep in the martian mantle. Mars has not only a topographical and crustal dichotomy but also a chemical dichotomy. We discuss different mechanisms which could generate not only these stuctures but also an early strong magnetic dipole field that vanishes after 500 Ma at the latest. Section 7 presents recent and future numerical improvements of the code Terra. Section 8 gives results on performance and scalability.

This contribution addresses computational considerations in the framework of the main inverse problem in geodetic research, i.e., the determination of the terrestrial gravitational potential. The computational considerations comprise strategies for both the solution of the underlying systems of equations in order to resolve the model parameters and the estimation of the parameter variance-covariance information.

Over the past several years, there has been significant interest and progress in using wireless communication technologies for vehicular environments in order to increase traffic safety and efficiency. Due to the fact that these systems are still under development and large-scale tests based on real hardware are difficult to manage, simulations are a widely-used and cost-efficient method to explore such scenarios. Furthermore, simulations provide a possibility to look at specific aspects individually and to identify major influencing effects out of a wide range of configurations. In this context, we use the HP XC4000 for an extensive and detailed sensitivity analysis in order to evaluate the robustness and performance of communication protocols as well as to capture the complex characteristics of such systems in terms of an empirical model.

When buildings are not longer used, controlled deconstruction is necessary. Besides systematic dismounting with heavy machines, a very efficient way is to use controlled explosives. The main load-carrying structural parts of the building are weakened by an explosive charge, which leads to a specific collapse kinematics. This has to be predicted reliably, in order to consider the boundary conditions such as neighboring buildings and traffic loaded streets. When planning such a collapse event, it is desirable to have a reliable simulation of the complete collapse process, also considering the uncertainty of primary parameters influencing e.g. the resistance of structural elements of a building. The ‘Research Unit 500’ [

1

], funded by the German Research Foundation (Deutsche Forschungsgemeinschaft – DFG) develops a special simulation concept by subdividing the collapse analysis into several – problem specific – analyses.