20005 research outputs found
Sort by
Integrating the technical level into a model-based safety and security analysis : why it's necessary and how it can be done
No full textWe motivate why it is necessary to integrate the technical level into a model-based safety and security analysis on the dynamical system level, and show how this can be done
KI-robuste Aufgabenstellungen für Moodle - Training bei Didaktik Forum der Beruflichen Hochschule Hamburg
No full textImagine eine digital gestützte Hochschullehre, die mit modernen und vielfältigen Methoden Studierende zum kritischen Denken und Problemlösen motiviert. KI-Tools wie ChatGPT gehören längst zum Studienalltag: Rund zwei Drittel aller Studierenden nutzen sie laut Bitkom zur Prüfungsvorbereitung oder beim Schreiben von Hausarbeiten. Auch auf Lernplattformen oder in plattformbasierten Blended-Learning-Formaten wie Moodle kann der Einsatz von KI eine Herausforderung sein. Doch wo liegt die Grenze zwischen erlaubter und unerlaubter Unterstützung? Ist alles, was erlaubt ist, auch lernförderlich? Wenn das Ziel des Studiums mündige Studierende sind, ist dann die Einteilung in verbotenen und erlaubten Einsatz von KI überhaupt sinnvoll?
In diesem Training verbinden wir rechtliche Aspekte mit didaktischen Konzepten, um neue Lernräume zu schaffen. Dazu stellen wir europäisches Recht (EU AI-Act) und nationales Recht vor. Auf Basis einer Delphi-Befragung von Studierenden, Lehrenden und KI-Expert*innen zeigen wir, was Studierende motiviert, Aufgaben selbstständig zu lösen und welche Motivation sie haben, KI zu nutzen, obwohl sie eine Aufgabe ebenso gut selbstständig lösen könnten. Wir analysieren, wie Kreativität und Selbstständigkeit durch Sprach- und Aufgabenformulierungen in Moodle-basierten Selbstlernräumen gestärkt werden können, ohne KI-Unterstützung kategorisch auszuschließen.
Im Training stellen wir konkrete Methoden und Formulierungen vor, die Lernerfolge fördern und zugleich den sinnvollen Einsatz von KI ermöglichen. Damit holen wir die vielseitigen Ideen aus der Theorie in den Praxisalltag: /imagine eine Hochschullehre, in der kollaboratives Lernen, Problem-Solving und KI-Technologie Hand in Hand gehen – und dabei die rechtlichen Rahmenbedingungen respektiert werden
Optimization of oscillation dynamics in a VDS deployment
No full textUnderwater detection is a fundamental capability for many vessels, especially for naval and research ships, where precise and reliable data is crucial. To achieve this capability ships can be equipped with a variable depth sonar (VDS) system, which is generally towed behind the vessel. When such a VDS system is launched into the water, it experiences excitation induced by the ship’s motion, causing the body to oscillate. These oscillations can result in amplitudes increases and uncontrolled behavior, which can damage the VDS and reduce the vessel’s operational effectiveness. If the oscillation of the VDS, while it’s launching, can effectively be predicted, and damping due to rope motion can be modeled the potential of underwater detection of the vessel can significantly be increased. In this paper, a method is presented to calculate the oscillation of the VDS system and to optimize the winch speed, while it’s launching, to reduce the oscillation. This approach relies on the differential equation of motion for the VDS system, taking into account a variable rope length and the ship induced excitations at the deployment point of the VDS. These excitations are calculated in the time domain using “E4-ROLLS”. To enable operations in conditions with higher significant wave heights, the method integrates the principle of swing damping along with a CHWARISMI optimizer to determine the optimal damping and winch speed for every situation. Initial results indicate that this optimized VDS system can be used effectively at significant wave heights up to one meter higher than previously feasible. This advancement enhances underwater detection capabilities and increases operational safety and efficiency for vessels in challenging sea conditions
Application of electrochemical approach for in situ H2O2 generation in enzyme catalysis
No full textHydrogen peroxide (H2O2) is an environmentally safe chemical and widely employed as a co-substrate in biocatalytic processes. Among various enzymes that are capable of utilizing H2O2 as a co-substrate, the recombinant unspecific peroxygenase from the fungus Agrocybe aegerita (rAaeUPO) stands out as the favorite due to its stability and versatility. Nevertheless, the technical application and combination of rAaeUPO and H2O2 in the biotechnological field remains challenging. This is due to the toxicity of H2O2 towards biocatalysts and the inactivation of rAaeUPO at an elevated concentration of H2O2. The precise dosing of H2O2 is of great importance, yet very challenging. A number of strategies have been explored to mitigate the deactivation effect of H2O2. Thus far, they have been largely deemed unsatisfactory due to the dilution effect or formation of complex by-products. An electrochemical approach represents an attractive method that provides controllable in situ generation of H2O2. The objective of this dissertation is to develop a fully controllable system for the electrochemical in situ generation of H2O2, designed as an optimizable platform for H2O2-dependent enzymatic reactions and to promote catalyst efficiency. To address these challenges, the All-in-One (AiO) electrode system was employed for the in situ generation of H2O2. This concept was integrated with the enzymatic hydroxylation, catalyzed by rAaeUPO, to establish a bioelectrochemical system (BES). The maximum H2O2 productivity and Faradaic efficiency achieved in the AiO electrode system were 0.87 µM min-1 cm-2 and 60%, respectively. The application of the AiO electrode within the BES yielded promising results, achieving a total turnover number (TTN) of 450,000 mol mol-1 and a turnover frequency (TOF) of 7.7 s-1. By incorporating aeration and increasing the number of electrodes, the specific surface area was enhanced, resulting in an increase in both TTN and TOF up to 700,000 mol mol-1 and 20.1 s-1, respectively, with a productivity of 1.3 g L-1 d-1. Although H2O2 productivity was identified as the limiting factor, the system demonstrated considerable potential for optimization through surface modification of the electrode. Moreover, this dissertation compared two modes of in situ H2O2 electrogeneration: the conventional galvanostatic mode and the H2O2-stat mode. The two modes were tested within the gas diffusion electrode system, employing the identical enzymatic reaction. While the galvanostatic mode demonstrated a maximum H2O2 productivity of 5.5 µM min-1 cm-2 and a productivity of 10.5 g L-1 d-1 at 6.4 mA cm-2, the H2O2-stat mode, particularly at a H2O2 concentration limit of 0.2 mM, yielded favorable outcomes with a TTN of 655,000 mol mol-1, a TOF of 80.3 s-1, and a productivity of 6.1 g L-1 d-1. The successful application of the H2O2-stat mode highlights its potential as a more effective alternative to the galvanostatic approach, significantly enhancing the process performance of rAaeUPO and advancing the field of enzymatic electrosynthesis.Wasserstoffperoxid (H2O2) findet häufig Verwendung als Co-Substrat in biokatalytischen Prozessen. Die rekombinante unspezifische Peroxygenase aus Agrocybe aegerita (rAaeUPO) weist eine hohe Stabilität sowie Vielseitigkeit auf und erscheint daher vielversprechend für technische Anwendungen in Kombination mit H2O2. Allerdings bleibt die technische Anwendung von rAaeUPO mit H2O2 bislang eine Herausforderung. Die Toxizität von H2O2 sowie die Inaktivierung von rAaeUPO bei höheren H2O2 Konzentrationen stellen wesentliche Herausforderungen dar. Bisherige Strategien zur Minderung dieser Effekte erweisen sich als unzureichend. Ein elektrochemischer Ansatz zur kontrollierten in situ Erzeugung von H2O2 könnte eine effektive Lösung bieten. Das Ziel dieser Dissertation ist es, ein vollständig kontrollierbares System zur elektrochemischen in situ Erzeugung von H2O2 zu entwickeln, das als optimierbare Plattform für H2O2-abhängige enzymatische Reaktionen konzipiert ist und die Katalysatoreffizienz steigern soll. Um die genannten Herausforderungen zu bewältigen, wurde das All-in-One (AiO) Elektroden-System für die in situ Erzeugung von H2O2 eingesetzt. Dieses Konzept wurde mit der enzymatischen Hydroxylierung, katalysiert durch die rAaeUPO, integriert, um ein bioelektrochemisches System (BES) zu etablieren. Die maximale H2O2-Produktivität und Faradaische Effizienz, die im AiO-Elektrodensystem erreicht wurden, betrugen 0,87 µM min-1 cm-2 und 60%. Der Einsatz der AiO-Elektrode im BES ergab vielversprechende Ergebnisse, mit einer total turnover number (TTN) von 450.000 mol mol-1 und einer turnover frequency (TOF) von 7,7 s-1. Durch die Integration der Begasung und die Erhöhung der Anzahl der Elektroden wurde die spezifische Oberfläche vergrößert, was zu einer Steigerung von TTN und TOF auf bis zu 700.000 mol mol-1 und 20,1 s-1 führte, bei einer Produktivität von 1,3 g L-1 d-1. Obwohl die H2O2-Produktivität als limitierender Faktor identifiziert wurde, zeigte das System ein beträchtliches Optimierungspotenzial durch Modifikation der Elektrodenoberfläche. Darüber hinaus verglich diese Dissertation zwei Modi der in situ H2O2-Elektrogenerierung: den konventionellen galvanostatischen Modus und den H2O2-stat Modus. Die beiden Modi wurden im Gasdiffusionselektrodensystem getestet, wobei die identische enzymatische Reaktion eingesetzt wurde. Während der galvanostatische Modus eine maximale H2O2-Produktivität von 5,5 µM min-1 cm-2 und eine Produktivität von 10,5 g L-1 d-1 bei 6,4 mA cm-2 aufwies, lieferte der H2O2-stat Modus, insbesondere bei einer H2O2-Konzentrationsgrenze von 0,2 mM, vorteilhaftere Ergebnisse mit einer TTN von 655.000 mol mol-1, einer TOF von 80,3 s-1 und einer Produktivität von 6,1 g L-1 d-1. Die erfolgreiche Anwendung des H2O2-stat Modus unterstreicht sein Potenzial als effektivere
Alternative zum galvanostatischen Ansatz, was die Prozessleistung der rAaeUPO erheblich verbessert und das Feld der enzymatische Elektrosynthese voranbringt.Deutsche Forschungsgemeinschaft (DFG
When heavy tails disrupt statistical inference
No full textHeavy tails (HT) arise in many applications and their presence can disrupt statistical inference, yet the HT statistical literature requires a theoretical background most practicing statisticians lack. We provide an overview of the influence of HT on the performance of basic statistical methods and useful theorems aimed at the practitioner encountering HT in an applied setting. Higher or even lower product moments (i.e., variance, skewness, etc.) can be infinite for some HT populations, yet all L-moments are always finite, given that the mean exists, thus, the theory of L-moments is uniquely suited to all HT distributions and data. We document how L-kurtosis, (a kurtosis measure based on the fourth L-moment) provides a general and practical heaviness index for contrasting tail heaviness across distributions and datasets and how a single L-moment diagram can document both the prevalence and impact of HT distributions and data across disciplines and datasets. Surprisingly, the theory of L-moments, an extension and evolution of probability weighted moments, has been largely overlooked by the literature on HT distributions that exhibit infinite moments. Experiments reveal L-kurtosis ranges under which various HT distributions result in mild to severe disruption to the bootstrap, the central limit theorem (CLT), and the law of large numbers, even for distributions which exhibit finite product moments
In specs we trust? Conformance-analysis of implementation to specifications in Node-RED and associated security risks
No full textLow-code development frameworks for IoT platforms offer a simple drag-and-drop mechanism to create applications for the billions of existing IoT devices without the need for extensive programming knowledge. The security of such software is crucial given the close integration of IoT devices in many highly sensitive areas such as healthcare or home automation. Node-RED is a framework to build applications from nodes that are contributed by open-source developers. Its reliance on unvetted contributions and lack of security checks raises the concern that the applications could be vulnerable to attacks, thereby posing a security risk to end users. The low-code approach could imply that users lack the technical knowledge to mitigate or even realize such security concerns. This paper focuses on hidden information flows in Node-RED nodes, meaning flows that are not captured by the specifications. They could (unknowingly or with malicious intent) cause leaks of sensitive information to unauthorized entities. We report the results of a conformance analysis of all nodes in the Node-RED framework, for which we compared the numbers of specified inputs and outputs of each node against the number of sources and sinks detected with CodeQL. The results show, that 55% of all nodes exhibit more possible flows than are specified. A risk assessment of a subset of the nodes showed, that 28% of them are associated with a high severity and 36% with a medium severity rating
Effects of laser beam shaping on Ti-6Al-4V single tracks in PBF-LB/M: A study with Aconity MIDI+
No full textThe application of beam shaping is gaining increasing interest in laser-based additive manufacturing (AM) technologies, as it revolutionizes the process by providing additional control over incident energy distribution and the resultant microstructure and mechanical properties of the manufactured part. This study provides a comprehensive analysis of how different laser beam profiles (Mode 0, Mode 3, Mode 6), applied via the Aconity MIDI+ system, in combination with varied layer thicknesses (30 µm, 60 µm, and 90 µm) and linear energy inputs (up to 1.0 J/mm), influence single-track formation, surface morphology, roughness, and melt pool behavior in the Laser Powder Bed Fusion (PBF-LB/M) process of Ti-6Al-4V, offering new insights essential for process optimization. To further elucidate these effects, a complementary FEM model was used to analyze how beam shape affects melt pool dynamics and overall process stability. The findings reveal that transitioning from a conventional Gaussian beam (Mode 0) to a ring-shaped beam (Mode 3 and Mode 6) promotes conduction-mode melting, resulting in enhanced process stability and smoother melt track formation at higher linear energy densities. These beam profiles reduce certain defects, including spattering and balling, while producing wider and more stable melt tracks. Conversely, Mode 0 generates deeper melt pools, increasing the likelihood of keyholing and tracks surface roughness at elevated energy levels. While a thinner layer (∼30 µm) facilitates stable and smoother track formation across all beam profiles, thicker layers (≥60 µm) exacerbate surface roughness and defects, especially with Mode 0. In contrast, the ring-shaped beams produce wider, smoother, and more stable melt tracks at higher energy inputs (∼0.6–1.0 J/mm) with thicker layers. These insights are particularly valuable for high-performance applications in aerospace and biomedical industries, where precise control over surface quality and defect formation in Ti-6Al-4V components is essential for meeting certification standards and ensuring production efficiency. Overall, these findings highlight the critical role of beam shaping, layer thickness, and energy input in achieving stable melt tracks and improving the consistency and reliability of the PBF-LB/M process
Poster: Can betweenness centrality mitigate greedy routing's sparse network problem?
No full textGeographic greedy routing in sparse L-band Digital Aeronautical Communications System (LDACS) Air-to-Air (A2A) networks, typical during the system's gradual deployment, often becomes stuck in local minima, requiring backup mechanisms. Instead of relying on such reactive measures, we propose combining the geographic advance metric with local Betweenness Centrality (BC). While global BC is infeasible due to its reliance on full network knowledge, our results show that local BC strongly correlates with its global counterpart. We further develop Subset Betweenness Centrality Routing (S-BCR), a protocol that combines local BC and geographic advance. We demonstrate through proof-of-concept simulations that it achieves higher routing success rates than either metric alone, all within tight beacon-size limits
In silico testing of a multimaterial scaffold for mandibular reconstruction
No full textIntroduction: Mandibular reconstruction following segmental resection is a challenging procedure. The implantation of scaffolds as an alternative for microsurgical free flaps appears as a promising strategy; however, there is still a lack of understanding of how such scaffolds should be designed to support bone regeneration. This study investigates the influence of scaffold design and its mechanical properties on the biomechanical conditions induced in mandibular reconstruction. Methods: A 3D finite element model of the human mandible was developed, and a large bone defect scenario was simulated, with physiological post-operative loading and boundary conditions. The large defect was bridged with a scaffold, supported by a titanium mesh, and stabilized with a load-bearing titanium fixation plate. To study the effect of the fixation device stiffness on the induced biomechanical conditions within the scaffold pores, two different materials were tested for the fixation device, namely, a Ti-6Al-4V titanium alloy and a polylactic acid (PLA). In addition, three different strut-based scaffold architectures were investigated with different strut orientations, while keeping the same strut diameter and similar overall porosity. Two types of material distributions through the scaffold were also studied. The first type was a hydrogel-based scaffold, whereas the second type was a multimaterial type where the scaffold was divided into three equal volume parts: in the center, a hydrogel material was employed, and in the extremities, a ceramic material. These combinations of two fixation materials and three scaffold architectures with two combination materials resulted in 12 experimental groups. Results and Discussion: No failure was predicted in the fixation devices for any of the configurations investigated. The PLA fixation device induced higher strains within the healing region than the titanium fixation device. Differences in scaffold architecture did not influence the strain levels within the healing region. Changes in the scaffold material distribution led to considerable differences in the mechanical strains within the scaffold pores. The multimaterial scaffold induced higher strains within the healing region than the only hydrogel scaffold, which might be beneficial to promote bone healing in the defect. Thus, a multimaterial scaffold seems to be able to provide a more suitable biomechanical environment to support bone regeneration, especially in large segmental defects. Future studies should focus on the mechanobiological optimization of the scaffold design and its fixation system in different clinical scenarios.European Unio
On linearization and uniqueness of preduals
No full textWe study strong linearizations and the uniqueness of preduals of locally convex Hausdorff spaces of scalar-valued functions. Strong linearizations are special preduals. A locally convex Hausdorff space (Formula presented.) of scalar-valued functions on a nonempty set (Formula presented.) is said to admit a strong linearization if there are a locally convex Hausdorff space (Formula presented.), a map (Formula presented.), and a topological isomorphism (Formula presented.) such that (Formula presented.) for all (Formula presented.). We give sufficient conditions that allow us to lift strong linearizations from the scalar-valued to the vector-valued case, covering many previous results on linearizations, and use them to characterize the bornological spaces (Formula presented.) with (strongly) unique predual in certain classes of locally convex Hausdorff spaces