Offenburg University of Applied Sciences
Hochschulschriftenserver der Hochschule OffenburgNot a member yet
6641 research outputs found
Sort by
Kaltes Wärmenetz und Wärmepumpe: Ein Ansatz zur Effizienzsteigerung mithilfe von Abwärmenutzung
No full textDieser Fachartikel beschreibt ein zukunftsweisendes Energiekonzept für den Sanitärherstellers Hansgrohe. Das Unternehmen strebt eine vollständige Dekarbonisierung der Wärmeversorgung am Standort Offenburg bis 2030 an. In Zusammenarbeit mit der Hochschule Offenburg wird untersucht, wie industrielle Abwärme aus der Produktion über ein kaltes Nahwärmenetz mithilfe von Wärmepumpen effizient genutzt werden kann. Der Text vergleicht dieses System mit konventionellen Luft-Wasser-Wärmepumpen und nutzt numerische Simulationen, um die Potenziale zur Steigerung der Energieeffizienz zu quantifizieren. Ein zentraler Aspekt ist dabei die Integration von Photovoltaik und thermischen Speichern, um den Eigenstromverbrauch zu optimieren und fossile Brennstoffe zu ersetzen. Durch die systematische Erfassung von Lastprofilen wird aufgezeigt, wie Abwärmequellen aus der Galvanik und Druckluftaufbereitung zur nachhaltigen Wärme- und Kältebereitstellung beitragen können
Multi-cycle high-throughput growth media optimization using batch Bayesian optimization
No full textBACKGROUND
The optimization of growth media is a crucial step in enhancing microbial growth and product formation in industrial bioprocesses. While Design of Experiments (DoE) is widely used for this purpose, it has inherent limitations, particularly in capturing complex interactions between media components and managing the high number of required experiments without overlooking important factors or interactions. This study evaluates batch Bayesian optimization (BBO) as an alternative approach by applying it in simulations and in wet-lab experiments, using Sporosarcina pasteurii as a model organism, to demonstrate its effectiveness.
RESULTS
In silico, BBO significantly outperformed DoE on average (+14%), although it was not superior in all test cases, highlighting the need for further investigation. In automated high-throughput microbioreactor experiments with S. pasteurii, BBO led to a 28% increase in maximum backscatter value, indicating a higher biomass titer compared to a DoE-optimized medium. BBO enabled a more targeted optimization of media components, allowing broader exploration when needed and focusing on specific adjustments where appropriate. The BBO-optimized growth medium, designed in small-scale experiments, also outperformed the DoE-optimized medium at a 2 L bioreactor scale.
CONCLUSION
In our use cases, BBO offers a more effective and targeted approach for growth media optimization than traditional DoE. Placing our results in the context of other studies, BBO shows promise for outperforming DoE in bioprocess optimization tasks. The ability of BBO to streamline the optimization process could enhance yields and reduce costs in industrial bioprocesse
Comprehensive Review of NIST Lightweight Cryptography Algorithms for IoT: Performance Evaluation, Attacks, Optimizations, and Protocol Integration
No full textThe Internet of Things (IoT) is rapidly expanding into critical healthcare, industrial, and commercial do- mains, yet its resource-constrained devices remain vulnerable to cyberattacks. IoT devices have resource constraints, making it challenging to execute standard security algorithms. To address these limitations, the National Institute of Standards and Technology (NIST) selected 10 Lightweight Cryptographic (LWC) finalist algorithms in 2023 to provide suitable confidentiality for constrained environments. This review focuses exclusively on these finalists and highlights their importance in modern IoT security. A systematic search was conducted across IEEE Xplore, ScienceDirect, Springer, and the Cryptology ePrint Archive, using the PRISMA methodology, defined keywords, and strict inclusion/exclusion criteria. In the initial 2118 retrieved studies, 40 high-quality contributions were selected after title, abstract, and full-text screening. The selected works were categorized into four themes: performance evaluation across hardware and software platforms, cryptanalytic and security assessments, algorithmic optimization, and integration of LWC algorithms into existing systems and communication protocols. Performance analysis research indicates that TinyJambu is the most energy-efficient among the NIST block-cipher-based algorithms. Xoodyak and ASCON demonstrated the best energy efficiency among permutation-based algorithms. On the other hand, the set of Elephant, ISAP, and Grain128-AEAD was the least energy-efficient, consuming up to 10 to 25 times more energy than the most efficient set, TinyJambu. In particular, the first reported cryptanalytic break of the 7-round Xoodyak, presented in a recent article, substantially expands the threat model for the NIST LWC finalist. Some experimental reports indicate a full key-recovery attack with success rates exceeding 90%. In contrast, adapted variants of the attack have proven effective against multiple Elephant-family ciphers, illustrating the importance of updated security assessments and implementation countermeasures. Finally, this study identifies critical research gaps that require further investigation, emphasizing the importance of addressing these challenges through targeted research efforts and developing adaptive solutions in future studies
Unified multi abstraction layer testbed & functional testing methodology for spatially distributed wireless networks in IoT systems
No full textWith the increasing trend of digitalization and the proliferation of IoT applications in domains such as smart cities, industrial automation, environmental monitoring, smart metering and many others have has accelerated the development and deployment of Spatially Distributed Wireless Networks (SDWNs). These networks form the backbone of wireless IoT systems, however, functional testing of these complex spatially distributed wireless communication networks remains a significant challenge due to their diverse topologies, heterogeneous technologies, dynamic wireless channel characteristics, and operational environments. This thesis addresses the important research gap and address challenges by proposing a unified multi abstraction layer testbed and functional testing methodology tailored for SDWN systems, with a strong focus on IoT-oriented use cases.
The principal scientific concept underlying this thesis is the formulation of a unified testbed environment and testing methodology that enables identical test case description, execution, and analysis across multiple abstraction layers—spanning simulations, emulations, and field testing. This methodology leverages a unified abstract test description and abstracted test runner with unified interfaces to the system under test and unified test results interfaces to ensure that test goal semantics are preserved while unifying the test description and execution methodology, and results remain comparable and reproducible, regardless of the testing platform/ abstraction levels. The thesis builds on the foundation that bridging the gaps and non-uniformity in testcase descriptions between existing simulation, emulation, and field test environments can lead to significant improvements in functional testing and performance efficiency, comparison, reproducibility and scalability.
A key academic contribution is the conceptualization and implementation of unified testbed and testing methodology with “press-one-button” execution capability, which allows the same test case to be reused, executed, and evaluated across all supported abstraction levels — streamlining the testing workflow significantly.
The results of this work demonstrate a significant improvement in SDWN functional testing and performance measurements efficiency and effectiveness, reducing development costs and time while improving deployment readiness. The unified methodology and prototype implementation offers a unified and systematic environment for developers, researchers and practitioners SDWNs, paving the way for systematic and efficient functional testing for future advancements in SDWN
Multifaceted Analysis of Deep Convolutional Neural Networks and Novel Fourier Modules
Get PDFThe increasing reliance on neural networks in everyday applications underscores a critical challenge: ensuring their robustness and reliability beyond idealized conditions. Evaluating vision classification models solely through clean accuracy and spatial perspectives is insufficient. Thus, we employ a multifaceted analysis of robust models and leverage Fourier theory to enhance robustness, efficiency, and our fundamental understanding of convolutional neural networks. This thesis explores the interplay between adversarial robustness, confidence calibration, efficiency and sampling artifacts through the lens of Fourier Theory in convolutional neural networks. We first demonstrate that adversarially robust models exhibit significantly lower overconfidence compared to their non-robust counterparts. We demonstrate that even subtle modifications to fundamental network components can significantly improve confidence calibration, highlighting the power of architectural design. Building on this, we investigate aliasing effects in robust models, revealing that they downsample more effectively than standard models, leading to reduced aliasing. To quantify this phenomenon, we introduce a novel aliasing measure and show its connection to catastrophic overfitting in FGSM adversarial training, inspiring an early stopping criterion based on our aliasing measure. Leveraging these discoveries, we present Frequency Low Cut (FLC) Pooling and Aliasing and Sinc Artifact-free Pooling (ASAP), novel Fourier-domain downsampling methods designed to be inherently aliasing-free. These techniques contribute to enhanced native robustness and improved adversarial training stability, effectively addressing catastrophic overfitting in FGSM adversarial training. Building upon our Fourier-domain investigations, we present Neural Implicit Frequency Filters (NIFFs), enabling efficient large convolutions. By leveraging neural implicit functions for weight learning and efficient Fourier-domain convolution, NIFFs provide a feasible and fair comparison to large spatial convolutions. Using NIFFs, we analyse learned kernel size preferences, revealing insights that facilitate more efficient network design. We demonstrate that optimal feature extraction often requires kernels larger than the typical 3 ×3, with 9 ×9 kernels being predominately learned by the network, especially on ImageNet-1k. Our multifaceted analysis and findings contribute to a deeper understanding of adversarial robustness, confidence calibration, sampling artifacts and the role of Fourier theory in convolutional neural network design, paving the way for more robust and efficient deep learning models
Chancen und Risiken der Finanzierung von Gewerbeimmobilien am Beispiel Hamborner REIT AG
No full textDie wissenschaftliche Arbeit bietet einen umfassenden Einblick in die Thematik der Gewerbeimmobilienfinanzierung, wobei sowohl die potenziellen Chancen als auch die eigenen Risiken beleuchtet werden. Zu Beginn der Arbeit werden die grundlegenden Aspekte der Immobilienfinanzierung dargelegt, einschließlich der verschiedenen Finanzierungsinstrumente sowie der spezifischen Eigenschaften von Immobilien als Anlageklasse. Ein besonderes Augenmerk wird hierbei auf die Analyse der Beziehungen zwischen wirtschaftlichen, zinsbedingten und konjunkturellen Faktoren und deren jeweilige Auswirkungen auf die Immobilienfinanzierung gelegt.
Im weiteren Verlauf der Arbeit werden die zentralen Chancen der Immobilienfinanzierung herausgestellt. Hierzu zählen insbesondere die Möglichkeit der Steuerersparnis, der Inflationsschutz sowie die Diversifizierung des Anlageportfolios. Im nachfolgenden Kapitel werden die potenziellen Risiken der Immobilienfinanzierung detailliert analysiert. Neben den finanziellen Risiken, wie beispielsweise Zinsänderungsrisiken, Leerstandrisiken und Liquiditätsrisiken, werden auch marktbezogene und umweltbedingte Risiken in den Blick genommen.
Einen wesentlichen Bestandteil der Arbeit bildet die Analyse der Hamborner REIT AG, die es ermöglicht, die zuvor erörterten theoretischen Konzepte in einem realen Kontext zu betrachten. Durch die Untersuchung der spezifischen Herausforderungen und Chancen, denen sich die Hamborner REIT AG gegenübersieht, wird die praktische Relevanz der theoretischen Ausführungen untermauert.
Die Arbeit schließt mit einer Zusammenfassung der zentralen Erkenntnisse sowie einer kritischen Auseinandersetzung mit den Ergebnissen. Darüber hinaus wird ein Ausblick auf mögliche zukünftige Forschungsfragen gegeben
Umkonzeptionierung und prototypische Implementierung eines webbasierten Transportmoduls
No full textDie bestehende Auftragserfassung der Disposition im Unternehmen weist verschiedene Schwachstellen auf, die die Benutzungsfreundlichkeit beeinträchtigen. Ziel dieser Arbeit ist die Entwicklung eines webbasierten Transportmoduls, das besonders auf eine benutzungsfreundliche Gestaltung ausgerichtet ist.
Die Arbeit beginnt mit einer Einführung in die Thematik, in der die Forschungsfrage und die Zielstellung dargestellt werden. Anschließend werden die theoretischen Grundlagen erläutert und die Methodik beschrieben. Danach folgt eine Analyse der bestehenden Systeme und der notwendigen Anforderungen mit ihren Funktionen. Auf Basis der ersten Entwürfe wurde ein Prototyp entwickelt. Abschließend wurde eine systematische Evaluation des Prototyps durchgeführt, um dessen Funktionalität und Benutzungsfreundlichkeit zu überprüfen. Die Ergebnisse zeigen, dass das entwickelte Transportmodul tendenziell benutzungsfreundlich ist und auf dieser Grundlage zukünftig weiterentwickelt werden kann
Exploring the antibacterial effects of novel phage ORFans against multidrug-resistent Acinetobacter baumannii by heterologous cloning
No full textAcinetobacter baumannii is a pressing antimicrobial resistance threat. Phage ORFans represent an novel source of antibacterials. This thesis evaluates four ORFans (ORF3, ORF6, ORF9, ORF36) from A. baumannii phage HMGU1 via in-silico triage, recombinant expression, purification, and first-pass activity tests. Predictions prioritized ORF6 as holin-like, ORF3 as a JAMM-family–like metalloprotease, ORF36 as a secreted/disordered candidate, and indicated low confidence for ORF9. The genes were cloned into pET-26b(+) and expressed in E. coli BL21(DE3); inducibility and toxicity were quantified by growth curves, and proteins were purified by FPLC. ORF3 yielded a 13 kDa product of high purity, whereas the other ORFs were recalcitrant under the tested conditions. Induction reduced host growth without lysis signatures, and purified ORF3 produced no clearance zones against A. baumannii, consistent with non-lytic or intracellular activity. Overall, the work establishes a reproducible pipeline from “hypothetical” phage genes to purified proteins and standardized readouts, nominating ORF3 for deeper mechanism and synergy studies
KI-gestützte Inhaltsanalysen im Social-Media-Marketing
No full textAngesichts der großen Menge an Inhalten auf Social-Media-Plattformen wie Instagram, TikTok oder YouTube stehen Unternehmen vor der Herausforderung, relevante Informationen aus meist unstrukturierten Datenformaten wie Text, Bild und Video zu extrahieren, zu analysieren und zu nutzen. Dabei rücken KI-gestützte Verfahren immer stärker in den Fokus. Der Beitrag systematisiert zentrale Analysemethoden und stellt ein strukturiertes Vorgehensmodell zur Inhaltsanalyse vor. Die einzelnen Kapitel geben einen detaillierten Einblick in die Analyse von Text-, Bild-, Video-und multimodalen Inhalten. Ziel ist es, einen fundierten Überblick über aktuelle Entwicklungen und methodische Ansätze zu geben, um Social-Media-Inhalte im Marketing effektiv und effizient analysieren und nutzen zu können
