Eldorado - Repositorium der TU Dortmund

Ressourcen aus und für Forschung, Lehre und Studium

Bei diesem Service handelt es sich um das Institutionelle Repositorium der Technischen Universität Dortmund. Hier werden Ressourcen aus und für Lehre, Studium und Forschung gespeichert, erschlossen und der Öffentlichkeit zugänglich gemacht.

Hauptbereiche in Eldorado

Wählen Sie einen Bereich, um dessen Inhalt anzusehen.

Gerade angezeigt 1 - 5 von 9

Fach- und Universitätsübergreifendes 1319
Fakultäten 13829
Graduate School und Graduiertenkollegs 12
Hochschulverwaltung 1248
Sonderforschungsbereiche 1594

Aktuellste Veröffentlichungen

Characterization of protein structure and dynamics using solution- and solid-state NMR
(2026) Kavaleuskaya, Hanna; Linser, Rasmus; Wiegand, Thomas
Proteins carry out biological function through a combination of structure and dynamics, often sampling multiple conformations and including intrinsically disordered regions. Obtaining residue-specific information for large, poorly soluble, heterogeneous, or highly dynamic systems remains challenging, because many established methods lose sensitivity or resolution under these conditions. This thesis develops approaches to study such systems using solution and magic-angle-spinning (MAS) solid-state NMR, supported where useful by computational and evolutionary analyses. The work focuses on three connected objectives. First, it addresses limited H/D back exchange in 1H-detected MAS NMR, which can obscure solvent-protected amide sites, by testing cell extract-based selective deuteration as a way to improve amide protonation while preserving spectral quality. Second, it establishes quantitative criteria for MAS sample preparation by comparing solution, sedimented, and microcrystalline sample states in terms of resolution, sensitivity, and stability, and uses these results to develop a practical framework for backbone assignment from combined solution and MAS datasets. Third, it applies residue-resolved solution NMR to the intrinsically disordered N-terminal domain of cGAS to characterize its conformational dynamics in the apo state and upon binding different DNA molecules, and to interpret these findings in an evolutionary framework. Together, these studies show that labeling strategy, sample state, and analysis approach strongly shape the information obtainable from NMR in challenging biomolecular systems. The results expand the practical toolbox for studying protein structure and dynamics by NMR.
Hybrid code generation
(2026) Busch, Daniel; Steffen, Bernhard; Lee, Edward A.
Language-Driven Engineering (LDE) aims to provide the most suitable modeling language for every purpose and stakeholder. As a concept for Low-Code/No-Code (LC/NC) en- vironments, LDE aims to create an easy-to-use development environment for everyone. Besides textual modeling languages, graphical modeling languages are particularly well suited for this purpose. They are often easy for humans to understand and easy to learn. Large Language Models (LLMs) are similarly easy to use. Natural language provides a universal interface that humans are accustomed to using every day. Consequently, LLMs have become ubiquitous in recent times. They are used in various fields, including text, image, and video generation. They are also highly popular for code generation. LLMs offer a high degree of flexibility, not only because of their natural language input, but also because of their output. They can produce arbitrary outputs, making them more versatile in code generation tasks than conventional code generators. However, they have downsides regarding controllability and reliability. For example, they are not deterministic and can produce different outputs when given identical inputs. Additionally, they are difficult to control reliably. Their output may not meet user expectations, and it may contain errors. This uncertainty is particularly undesirable in environments like LDE, where code is generated from formalized models. Nevertheless, LDE environments may still want to benefit from LLMs’ flexibility for code generation. Furthermore, natural languages allow for the simplicity of expression that Domain-Specific Languages (DSLs) seek to provide. LC/NC approaches aim to provide the same code generation experience as LLMs. These approaches enable even non-expert users to easily articulate their needs and generate code. Therefore, combining LDE and LLMs would enable great synergies. The flexibility of LLM code generation could be incorporated into LDE. At the same time, LDE could provide mechanisms to incorporate the control of conventional code generation from formalized models into LLM code generation. This dissertation presents a hybrid code generation approach that combines LLM- supported and conventional code generation in the context of LDE. Specifically, it combines a two-step generation approach and an extension to Template-based Code Generation (TBCG) for LLMs. This two-step process leverages the flexibility of LLMs within LDE, while maintaining control. To achieve this, it intertwines DSLs and natural language into Domain-Specific Natural Languages (DSNLs). This enables synergies between conventional and LLM code generation. The extension to TBCG makes it easily usable as an LLM tool. Instead of outputting code without guidance, it constrains LLM code generation output. This makes LLM code generation more controllable. The hybrid code generation approach takes advantage of the flexibility of LLMs. At the same time, it establishes three layers of control over the use of LLMs: 1. Contextualization: The presented approach generates contextualization for the LLM, so users can interact with it using a DSNL instead of natural language alone. DSNLs define the domain in which LLMs operate and enable referencing of other models in the LDE environment. DSNLs shift the responsibility of good prompting from users to LDE developers. This makes it easier to achieve good code generation results. Thus, this layer controls the input for LLMs. 2. Validation by Design: System-level validation is applied to observe whether the generated output meets user expectations. Exploiting control over code generation ensures that Active Automata Learning (AAL) can infer behavioral automata for validation purposes. This enables quality control of the output at the semantic level. 3. Output Constraints: Extending TBCG for easy use with LLMs guides them regarding their generation capabilities. Rather than allowing the LLMs to output arbitrary and potentially unwanted code, the TBCG tooling constrains their output options. These constraints control the LLMs during code generation.
Using single sided 1H nuclear magnetic resonance to investigate water absorption and stability of earth-based building materials
(2026-01-17) Soßna, Florian; Schulte Holthausen, Robert; Orlowsky, Jeanette
There is an increasing interest in the use of earth-based building materials to improve the environmental footprint of our built environment. One of the crucial features currently limiting its use is the inferior resistance of the clay binder to water absorption and erosion. Furthermore, there is a lack of meaningful quantitative measurement methods for characterising earthen material as building material. Here, we study the water absorption into earth-based materials with 1H single-sided nuclear magnetic resonance. Samples are prepared using an elaborate extrusion process and preconditioned under several typical moisture conditions. We demonstrate the use of the 1H NMR technique to visualize the water absorption and to quantify the capillary water absorption. The latter increases with a more water-reducing preconditioning. Furthermore, the technique allows for the visualisation of the internal swelling of the clay binder, illustrating the strong influence of sample preconditioning.
Enhancing edge integrity of high-strength steels by high-speed blanking to achieve improved crashworthiness
(Elsevier BV, 2026-01-30) Schrage, Olaf; Lingbeek, Roald; Peetsalu, Priidu; Hahn, Marlon; Dardaei Joghan, Hamed; Korkolis, Yannis P.; Tekkaya, A. Erman
The increasing use of ultra-high-strength steels (UHSS) in automotive safety components is driven by stricter crash safety requirements, vehicle weight reduction, and ecological goals in production and service. The application of UHSS requires adaptations in the manufacturing process chain, as conventional slow-speed blanking (SSB) used in mass production is challenging due to tool wear. Another aspect is crashworthiness: The interaction between material properties and blanking-induced defects—such as surface irregularities, microvoids, and microcracks—promotes crack initiation at free edges and limits edge formability. Local plastic deformation without breakage is a precondition for a stable break load of safety components, wherefore edge stretchability serves as an indicator for crashworthiness. High-speed blanking (HSB) of three steels with ultimate tensile strengths in the range of 1500 MPa—martensitic Docol 1500M, press-hardened (PH) 22MnB5, and carbon steel C60—is examined. Blanking trials are followed by central-hole tensile tests (CHTT) to assess edge stretchability. HSB produces edges with high geometric accuracy and homogeneous fracture surfaces, exhibiting roughness values comparable to wire-eroded surfaces. The shear-affected zone is confined to a narrow band of less than 2% of the sheet thickness, which is four times smaller than those observed in SSB. CHTT results show that HSB edges retain the same load-bearing capacity and edge fracture strain as wire-eroded edges, showing that edge integrity has not been compromised by HSB. In contrast, SSB triggers premature crack initiation reducing the achievable fracture strain by nearly half.
Chasing prompt muons
(2026) Gutjahr, Pascal Thomas; Rhode, Wolfgang; Kröninger, Kevin
The atmospheric muon flux provides a powerful probe of TeV-to-PeV hadronic interactions in air showers and offers unique access to particle production in the forward kinematic region that is not accessible at collider experiments. While the conventional flux from pion and kaon decays is comparatively well understood, the prompt component, dominated by decays of charmed hadrons, lacks precise experimental constraints at the highest energies. The atmospheric muon flux is also a dominant background for astroparticle analyses, in particular searches for astrophysical neutrino sources. This thesis presents a measurement of the atmospheric muon flux, including its prompt component, using 12.12 years of data from the IceCube Neutrino Observatory. The muon energy spectrum is unfolded to surface level in the range 10 TeV to 15 PeV, extending previous IceCube measurements by roughly an order of magnitude in energy. A key contribution of this work is a machine learning-based reconstruction of the energy of the leading muon (the most energetic muon within a bundle), increasing sensitivity to the spectral differences between conventional and prompt components. New CORSIKA simulations were produced to tag muons of prompt or conventional origin, and a dedicated event selection was developed for this analysis. The unfolded flux is compared with MCEq and CORSIKA predictions for several combinations of primary cosmic-ray and hadronic interaction models. Among the models considered here, the MCEq prediction using the GlobalSplineFit primary cosmic-ray model together with SIBYLL 2.3c provides the best agreement with the high-energy unfolded flux. For this prediction, the prompt atmospheric muon flux is observed with a significance of 4.4 sigma. Several models have been tested and prompt and conventional normalizations are determined, with most fits exceeding 5 sigma. The results provide new experimental constraints on forward charm production in air showers and will contribute to ongoing investigations of the muon deficit in air shower experiments.