Leibniz MMS Days 2023
April 17 - April 19, 2023
Potsdam

For download of available posters please click on the title.

Determination of air exchange rates in naturally ventilated barns - Validation of prediction Models 2.0

Ali Alai (ATB Potsdam)

We present the BeLuVa 2.0 project, which investigates the impact of boundary conditions on air exchange rates (AER) and gas dispersion patterns in individual animal occupied zones (AOZ) using simulations with a computational fluid dynamics (CFD) model. The project aims to provide a basis to design "healthy" husbandry systems with minimal pollutant and pathogen loads and to evaluate uncertainties of emission estimates based on gas balancing methods. We will discuss the three key objectives of BeLuVa 2.0. The first objective is to investigate the impact of AOZ configuration and heat transfer processes in the CFD model on the accuracy of predicted gas dispersion patterns and AER. The second objective is to broaden the understanding of the influence of wind incident angle to identify tipping points where wind direction changes induce particularly strong modifications in the flow pattern. The third objective is to refine parametric models for local AER and gas balancing accuracy derived in preceding study. We will present the methodology used to refine our AOZ model by investigating the impact of AOZ height and the ratio of lying and standing cows on resistance parameters. We will also discuss the impact of neglected heat transfer processes on local cow comfort assessment and the refinement of the heat transfer coefficient in the porous medium model that is used to parametrize processes in the AOZ. Additionally, we will discuss the dynamic source strength of different gaseous species, namely ammonia, carbon dioxide, and methane, and the injection rates of target gases to modulate the influence of gas density and source altitude on concentration fields inside the barn.

Device-Scale Simulation of a SiGe Quantum Bus for Coherent Spin-Qubit Shuttling

Lasse Ermoneit (WIAS Berlin)

Spin qubits in gate-defined semiconductor quantum dots (QDs) are a major candidate for the realization of fault-tolerant quantum computers. In recent years, isotopically purified Si/SiGe heterostructures have proven to be a promising material platform that enables exceptionally long coherence times for spin qubits. Recent concepts for scalable architectures require coherent links that interconnect separate qubit arrays to enable the transfer of quantum information over mid- to long-range distances. These interconnections are realized by a so-called quantum bus, which is a new functional element of quantum computers. The quantum bus allows to shuttle electrons along a one-dimensional channel using moving QDs implemented via pulsed control fields. We present a numerical simulation framework that combines the computation of electrostatic control fields, electron wave function propagation, and material-specific defects. We investigate optimal control schemes to maximize the transmission fidelity of the qubit in the presence of defects and disorder.

Scaling Properties of Tree-like Self-Similar Polymers

Ron Dockhorn (IPF Dresden)

In this study, regular polymeric Vicsek- and T-fractals are compared to dendrimers in means of theory and simulations. Albeit all structures exhibit exponential growth both for the number of monomers inside the structures and also for the terminal groups their structural properties differ significantly. Computer simulations with the LeMonADE-framework (https://doi.org/10.5281/zenodo.5061542) are performed to investigate the scaling properties of the tree-like self-similar polymeric fractals utilizing the Bond-Fluctuation-Model with the Metropolis method as well as with the parallelized Wang-Landau algorithm. The radius of gyration, the scattering intensity, and the θ-point of those systems is investigated to examine the coil-globule transition of the polymeric fractals. A mean field theory for the scaling exponent in different solvent regimes is applied and found in fair agreement to the simulation data. A cross-over from almost linear chain behavior to spherical shape is observed, which can be tuned by the intrinsic functionality of the building blocks. The polymeric fractals can be an alternative to dendrimers in the class of hyperbranched polymers.

Livestock-environment-interaction in naturally ventilated housing on the example of ammonia

Sabrina Hempel (ATB Potsdam)

Contemporary livestock husbandry is far from being sustainable. One of the reasons is the emission of ammonia. There are large differences in the emission rates of individual farms, as a result of complex interactions between outdoor climate, indoor microclimate and the emission source strength and gas dilution. We combine fluid dynamics and reaction-kinetics modelling to better understand these interactions, predict emission values, optimize monitoring systems, and identify and evaluate emission mitigation potentials. Simulations of a naturally ventilated dairy cattle building showed that the coupling of different (semi-)mechanistic modelling approaches to project ammonia emission dynamics has a great potential for predicting annual average ammonia emissions and the variations at the daily and subdaily time scale. However, uncertainty and unresolved processes in the submodels render emission and air quality monitoring as well as smart ventilation control challenging.

Machine Learning for Simulation Intelligence in Composite Process Design

Denis Korolev (WIAS Berlin)

Accurate modelling of composites can be challenging due to the complex geometries of composite media and the presence of multiple scales with different physics. This requires computational grids with very fine resolution and imposes severe computational constraints on standard numerical solvers designed to solve the underlying equations (PDEs). Homogenization techniques are often used to overcome such difficulties and allow efficient extraction or upscaling of material properties from the microscale level for further use in macroscale simulations. In addition, mesh-free physics-informed neural networks can be used as the PDE solvers. In this poster, we discuss the applications of physics-informed machine learning to the composite process design and homogenization, the possibilities of combining neural networks with standard numerical solvers, and the possible advantages, disadvantages and future perspectives of our machine learning approach in the context of materials modelling.

Thermo-hydraulic modelling and simulation of borehole heat exchangers for an optimized shallow geothermal energy utilization in urban areas in Germany

Ernesto Meneses Rioseco (Georg-August-Universität Göttingen)

Ground coupled heat pumps are widely recognized as exceedingly energy efficient heating and cooling arrangements for a broad range of building applications. Properly capturing and implementing the geological settings as well as the thermal and hydraulic conditions at the local- and regional- scale of the first 400 m depth under the Earth surface remains a challenging task for geoscientists. We illustrate in this work the current modelling setup and numerical efforts taken to simulate a variety of multiple borehole heat exchangers (closed loops) and shallow geothermal wells (open loops) in different geological settings typical of the first 400 m depth under the Earth surface in Germany.

A computational framework for sustainable geothermal energy production in fracture-controlled reservoir based on well placement optimization

Ondrej Partl (WIAS Berlin)

We model numerically the transient non-isothermal fluid flow in a fracture-dominated geothermal reservoir in the context of sustainable geothermal energy production. Our goal is to find optimal placements of multi-well geothermal facilities using gradient-based optimization algorithms. The fractured rock is modeled as a 3D layered porous medium containing discrete fracture networks represented by 2D manifolds. The spatial discretization is carried out using the finite element method. Sustainable and optimized geothermal energy production in complex geological settings is a subject of ongoing research. In this work, we present our latest modeling and simulation results.

Learning Stiff Atmospheric Chemical Kinetics: The Regional Atmospheric Chemistry Mechanism (RACM)

Levin Rug (TROPOS Leipzig)

Incorporating complex atmospheric chemistry into Chemical Transport Models (CTMs) is crucial to improve accuracy and reliability of the results of CTMs. Due to the high dimension and stiffness of the resulting systems of ordinary differential equations (ODEs), it remains challenging to ensure accurate results, while preventing prohibitive computational cost from implicit solvers for such ODEs. Neural networks are in comparison cheap to evaluate, but suffer from error accumulation in long-term predictions. We present a neural network approach to emulate the solver of a realistic system (RACM), a handling of the curse of dimensionality and remarks on quality of data. The range of training data is supposed to cover typical atmospheric conditions, which is important, since we have no guarantees for accurate results outside this range. We achieve magnitudes of order speed-up for simulating chemistry, with stable and low-error long-term predictions in the testing data set.

Simulation of elastic strain in electron shuttling devices

Ignatii Zaitsev (IHP Frankfurt/Oder)

Characterization and engineering of elastic strain in semiconductor quantum devices is essential to manage band energies and transport properties of the device. We use x-ray diffraction microscopy to map the strain tensor in the Si quantum well between plastically relaxed SiGe buffer layers under arrays of TiN gate electrodes of the same design, but different parameters of the physical vapour deposition process that cause the electrodes to exert different amounts of stress. We compare these values with those simulated in COMSOL Multiphysics, and estimate their effect on the band behaviour using the k·p model. The extracted strain is confirmed by the simulation to be caused by the stress in the gate electrodes, and its modulation is shown to be close to ~4*10-4 found by Park et al. Additionally, the simulation allows to extrapolate the strain field to the temperature range of T ~ 20 mK the qubit is meant to be operating at, showing that would cause further tensile stress in the electrodes. The strain- derived band edge variation is expected to be at -1.5 meV, which is comparable in magnitude to the charging energy of electrostatic quantum dots.