## The 10 most recently published documents

The Lateral-Photovoltage-Scanning-Method (LPS) operates well for Si, Ge and Si_{1–x}–Ge_x for an analysis in defect regions below one part per million, where Secondary Ion Mass Spectroscopy (SIMS) or X-Ray Fluorescence (XRF) signals fall below its detection limit. Although LPS is well established since 1999, it is still poorly investigated. We used a computational simulation finite volume (FVM) approach, solving the van-Roosbroeck equations in three dimensions using a MUltifrontal Massively Parallel sparse direct Solver MUMPS. The signal transport is simulated by solving the Maxwell equations in two dimension for different sample geometries.
It could be shown that a typical LPS-measurement is distorted due to the samples geometry (except cuboid). This distortion can be simulated, understood and recalculated, as discussed for trapezoidal or cylindrical samples. Also using the signal generation simulation of this measurement technique it can be shown, that the measurement signal is convoluted depending on the inherent minority charge carrier life time reducing the local resolution. An investigation of the local resolution were made using a Gaussian function as the convolution function of this method. A comparison of simulations to real measurements was discussed on silicon samples with boron implantation pattern.
In 1955 Tauc already stated that the bulk photovoltaic effect, causative for the LPS measurement set-up, could be used detecting any quantity, which affects the band structure of a semiconductor.As strain is coupled to the conduction and valence band profiles by the deformation potential theory by van-de-Walle, we investigated the possibility to detect strain variations using LPS simulations. For an n-type Si sample with an on-top stressor stripe (silicon-nitride) the strain distribution in Si got calculated by finite elemente simulation (FEM) using solid mechanics module. By directly converting the strain profile to a single conduction and valence band, FVM LPS simulations were performed. It could be shown, that the LPS voltage can be connected to hole traps caused by the conduction and valence band profile. Therefore we can finally conclude, that the LPS measurement set-up is suitable measuring conduction and valence band variations caused by strain.

A local algebraic simulation model was developed, to determine the characteristic length scales for dispersed phases. This model includes the Ishii- Zuber drag model, the lift, the wall lubrication force and the turbulent dispersion force as well. It is based on the Algebraic Interface Area Density (AIAD) model from the Helmholtz Zentrum Dresden Rossendorf (HZDR), which provides the morphology detection and the free surface drag model. The developed model is in agreement with the current state of knowledge based on an examination of the theory and of state of science models for interface momentum transfer.
This new simulation model was tested on three different experiments. Two experiments can be found in the literature, the Fabre 1987 and the Hewitt 1987 experiment. And the third simulation is based on a steam drum experiment. This steam drum experiment is designed with ERK Eckrohrkessel GmbH internals and was developed to examine the droplet mass flow out of the turbulent separation stage.
The implementation of all models and tests was performed using Ansys CFX. The first analysis was carried out to reproduce a wavy stratified flow to examine the effects of different simulation model set-ups according to the velocity and kinetic energy profiles, as well as the pressure drop gradient and the water level measured by Fabre 1987. The second analysis was a proof on concept for reproducing the vertical flow pattern by an experiment from Hewitt 1987. The third simulation analysed the water distribution in the steam drum and feeding pipes system as well as the droplet carryover into the gas phase in the turbulent separation region of the drum.
These simulations have shown, that the accuracy of the particle distribution model in interaction with the drag and non-drag forces is able to reproduce horizontal and vertical flow patterns. Higher deviations are recognised for the liquid volume fraction close above the interface. Generally, simulations can now be performed to optimise industrial steam drum designs.

We investigate the dynamics of a Newtonian liquid layer bounded on one side by a horizontal and planar substrate and on the other side by its free and deformable surface. The system is subjected to a time-periodic gravitation field in lateral or normal direction. Based on a nonlinear coordinate transformation, which maps the time-dependent surface onto a constant domain and thus eliminates the need for tracking the interface, a finite-difference method on staggered grids is presented, allowing direct numerical simulations of the full incompressible Navier-Stokes equations in two and three dimensions. Taking into account the continuity equation, a sparse linear system for the pressure is obtained from the discretized Navier-Stokes equations whose solution satisfies the conservation of momentum and mass, so that pressure corrections can be avoided. In the case of a harmonic force perpendicular to the substrate, we find in high frequency ranges the classical square patterns oscillating subharmonically with half of the driver’s frequency. For a slow excitation, hexagonal Faraday waves emerge oscillating at the forcing frequency. Vertical two-frequency excitations lead to more complex patterns — surface waves having the shape of a square superlattice are found. In the case of a lateral excitation, the formation of coarsening droplets is observed. We show that ratchet-like forces generate a nonvanishing averaged flow rate inducing a preferred direction of motion of the drops. These results correspond well with those of a simplified model based on the lubrication approximation. Our investigations also include systems in Rayleigh-Taylor configuration, where the liquid is located on the underside of the substrate. By considering rigid walls instead of periodic boundaries, wave amplification due to resonance can be studied. The corresponding numerical results are in good agreement with experimentally obtained data.

With parallelism becoming the standard in computer design, research on parallel constraint solving technique is of vital importance for enhancing the performance of constraint solving. In this dissertation, we reviewed the literature on exploiting parallelism in constraint solving to help gain insight into the rationale of different types of parallel constraint solving approaches. On this basis, we analyzed the effectiveness of parallel constraint solving, with the focus on obtaining a first solution when solving computationally hard combinatorial problems. We have shown that a well-designed search space splitting method and constraint programming model can enable the embarrassingly parallel search (EPS) to solve some open instances of the social golfer problem that have not been solved by a sequential algorithm. We also observed superlinear speedups when solving these instances, which confirms our theoretical analysis. Besides, we examined two practical constraint optimization problems, including the traveling tournament problem with predefined venues and the talent scheduling problem. Our proposed constraint models outperformed the existing models on the same instances, and the EPS approach could always attain better feasible solutions in terms of the optimal objective value by using more parallel processors. To explore the use of massively parallel processing, we proposed the parallel stochastic portfolio search, which is a simple and non-intrusive way to parallelize different incarnations of a sequential solver. When comparing the existing portfolio to our portfolio approach by solving the same constraint satisfaction problems using the same constraint models, our technique could solve harder and larger instances. The successes of our new parallel approaches are attributed to early diversity; i.e., some diversity early in the search introduced by parallelism can offset early mistakes caused by weak heuristic choices. Unlike the other techniques (e.g., limited discrepancy search) used to overcome early mistakes, the studied two parallel constraint solving approaches not only can explore more nodes simultaneously but also does not sacrifice the guarantee of completeness.
We also presented a hypertree decomposition method that builds a degenerate decomposition tree for a given constraint network, in which each node of the decomposition tree possesses and executes a subset of constraints of the given constraint network. The usefulness of our proposed parallel techniquedepends on whether we can find an efficient way to join the results of each node.

The increasing demand for competitive, whilst also environment-friendly airplane travel, compels the design of highly efficient engines in the aeronautical field. A potential for improvement of traditional polycrystalline Ni-based superalloys, aiming higher creep resistance, was investigated. The approach adopted the concept of metal matrix composites (MMCs) to incorporate a rigid discontinuous phase, in the form of particles, to a γ’-strengthened Ni-based superalloy. In order to make the concept feasible, different microstructures resulting from diverse manufacturing techniques were investigated. By using distinct mixing and sintering methods, powders of Inconel X-750 and TiC were combined to form composites containing 15 vol.% of reinforcing particles. Powders were prepared with low and high energy milling processes, and formed by uniaxial pressure sintering and spark plasma sintering methods. Non-reinforced variants and composites had microstructures thoroughly examined at their initial state and after long isothermal aging treatments. Selected variants were further submitted to tensile and compression creep tests at temperatures between 700 and 800 °C, in the stress range of 200 to 500 MPa.
A comprehensive analysis was conducted using techniques such as EBSD, XRD Rietveld refinement, EDS and TEM to evaluate the development of γ’, η and TiC phases, determining the achievable microstructures with each fabrication method and establishing their evolution after aging treatments over times up to 1000 h. Likewise, creep properties were analyzed by obtaining parameters such as creep exponents, threshold stresses and activation energies. A creep life estimation was conducted with the use of a Monkman-Grant relationship and a Larson-Miller parametrization. Lastly, the potential for a reduction in creep strain rates in a working turbine blade, considering the density of investigated materials as a parameter, was evaluated.
All produced composites presented power law creep, with dislocations surpassing γ’ particles by climb. The variant produced by high energy ball milling and spark plasma sintering exhibited the highest creep rates, resulting from intense diffusion through grain boundaries. It also presented η phase after long isothermal aging, which affects negatively the creep resistance. Contrastingly, in the variant produced by low mixing combined with pressure sintering the lowest creep rates were observed. It was proposed that reinforcing TiC particles effectively acted as a load bearing phase, counterbalancing the adverse effects of the intergranular diffusion in the refined microstructure. Furthermore, a higher microstructural stability was observed in this variant, resulting from limited interaction between TiC particles and the matrix during fabrication.

Beschleunigung der Verdichterkennfeldberechnung mithilfe von Methoden des maschinellen Lernens
(2021)

In der heutigen Triebwerksentwicklung ist die Verwendung komplexer und zeitaufwändiger numerischer Strömungssimulationsverfahren (3D-CFD) unerlässlich. Dies gilt auch und insbesondere für den Bereich der Verdichterkennfeldberechnung, welcher viele zeitintensive 3D-CFD Berechnungen benötigt. Dabei sind zur qualitativen Beurteilung eines Verdichterentwurfs sowohl Betriebspunkte wie, Reiseflug, Start und Landung, hinreichend genau abzubilden, als auch die kritischen, den Verdichterarbeitsbereich limitierenden Betriebsgrenzen Pumpen und Sperren zu detektieren. Bisherige Arbeiten zur automatisierten Verdichterkennfeldberechnung basieren auf strukturierten Berechnungen von verschiedenen Drehzahllinien, auf welchen jeweils isoliert Pump- und Sperrgrenze gesucht werden. Durch die Beschränkung auf einzelne Drehzahlen wird jedoch nicht der gesamte Charakter des Kennfeldes erfasst, so dass unbekannte Betriebsbereiche aus linearer Interpolation abgeleitet werden müssen. Ein zusätzlicher Nachteil solcher auf einzelne Drehzahllinien fixierten Methoden ist ihre geringe Parallelisierbarkeit.
Der Fokus dieser Arbeit liegt daher auf der Entwicklung eines effizienten Verfahrens zur Erfassung des gesamten Verdichterkennfeldes. Die zwei wesentlichen Anforderungen an das Verfahren sind erstens die Reduktion der Anzahl der notwendigen CFD-Berechnungen zur hinreichend genauen Beschreibung des Verdichterkennfeldes sowie zweitens die Beschleunigung jeder einzelnen 3D-CFD-Berechnung. Zu diesem Zweck wird zur Kennfeldberechnung eine Strategie vorgeschlagen, welche sich von der üblichen strukturierten Berechnung einzelner Drehzahllinien löst und stattdessen mit unstrukturierten, zufällig bestimmte Stützstellen arbeitet. Dabei wird ein zweiphasiges Verfahren entwickelt, bei dem zunächst die Pump- und Sperrlinien in ihrer Gesamtheit mit einer iterativen, hoch parallelisierbaren, auf Support-Vector-Machine beruhenden Strategie bestimmt werden. Als nächster Schritt wird mit Methoden der statistischen Versuchsplanung eine ausreichende Dichte von Stützstellen innerhalb der Betriebsgrenzen des Verdichters generiert. Abschließend werden auf Basis aller verwendeten Stützstellen Antwortflächen für Verdichterdruckverhältnis, Wirkungsgrad und Eintrittsmassenstrom aufgebaut.
Zur Reduktion der Rechenzeit jeder einzelnen 3D-CFD Rechnung werden unterschiedliche Methoden zur Erzeugung von Startlösungen betrachtet. In diesem Rahmen werden Initialisierungsansätze aus reduzierten Strömungsmodellen und aus der Superposition von bereits bekannten Strömungslösungen auf Basis der Methode der Proper-Orthogonal-Decomposition (POD) untersucht.
Als Validierung wird abschließend das entwickelte Verfahren zur Kennfeldberechnung in Kombination mit dem POD-Initialisierungsansatz erfolgreich auf die Analyse eines 4.5- stufigen Forschungsverdichters angewendet.

Machine learning is a field that has been the object of study of many researchers around the globe during the last decades. Very often to solve machine learning challenges like classification problems for example, one needs to train an artificial neural network. To train this network a certain loss function has to be minimized. There is a ubiquitous approach to achieve this which consists of using variants of the stochastic gradient descent combined with the backpropagation algorithm. In our work, we aimed at testing a rather non-conventional scheme consisting of making use of the solvers a software called AMPL offers.

Die vorliegende Publikation präsentiert Ergebnisse von 20 Studierenden des Masterstudiengangs Stadtplanung aus dem Projekt „Plattform Innenstadt - Zurück zu multifunktionalen Zentren“ im Sommersemester 2020 am Fachgebiet Stadtmanagement der BTU Cottbus-Senftenberg.
Das Projekt befasste sich mit zehn ausgewählten Lausitzer Kleinstädten. Der Fokus lag dabei auf strukturellem Wandel in der Region sowie im Einzelhandel im Besonderen und den daraus resultierenden Wirkungen auf die Innenstädte.
Eingebunden war das Projekt in einen bundesweiten Wettbewerb des Wissensnetzwerks Stadt und Handel e.V. (WSH). Daran beteiligten sich insgesamt vier Hochschulen mit unterschiedlichen Fachrichtungen und Themen, zum Beispiel Städtebau,Stadtplanung oder Betriebswirtschaft.
Eine Vielzahl an Lösungsansätzen und eine große Bandbreite von Ergebnissen wurden in Austausch gebracht. Dabei konnten die Studierenden Erfahrungen sammeln und die Untersuchungsgegenstände auf neue Art methodisch und didaktisch angehen.