Vorarlberg University of Applied Sciences
Online Publication Server University of Applied Science VorarlbergNot a member yet
2106 research outputs found
Sort by
Sequential linear optimization method with fairness considerations for PV hosting capacity quantification in low-voltage grids
No full textHosting capacity quantification intends to provide the maximum aggregated capacity of distributed generation in low-voltage grids. Given the nonlinear nature of this problem, finding a solution in a simple and effective manner remains challenging. Moreover, previous studies have shown that solutions often display high disparities in hosting capacities among individual nodes, penalizing some and favoring others. Such disparities raise fairness concerns. In this paper, we propose a sequential linear optimization method to quantify the PV hosting capacity in low-voltage grids to address these issues. The linear formulation is achieved by using sensitivity matrices that approximate linear relations between changes in voltage and current magnitudes due to changes of PV capacities. To mitigate inaccuracies due to the linear formulation, multiple optimization problems are solved sequentially, updating the grid state at each stage with an AC power flow to realign the solution with the nonlinear problem. The performance of the proposed method is showcased using the CIGRE European low-voltage grid and by comparison with other established methods. To draw conclusions on a large and realistic dataset, the hosting capacities of 5196 low-voltage feeders from Austria are quantified under varying levels of disparity allowances
FFG Leitprojekt TeleCareHub – Digitale Services für die Pflege und Betreuung zu Hause
No full textIn Austria, much care is provided informally by relatives. To support these carers, the TeleCareHub project developed a digital platform with seven telecare services for informal dementia carers. Five services are currently evaluated in a one-year, multi-center pilot study: knowledge base, training modules, video counselling, self-help group, and stress check. The study assesses impact on caregiver burden, preparedness, relationship quality, and acceptance using questionnaires, focus groups, and usage data. By June 2025, 46 participants (78% female, average age 51) enrolled. Early results show barriers to support contact as well as digital registration issues. These findings guide platform development and integration into dementia care
Data-driven modeling for privacy-preserving energy prediction of industrial robots
No full textThis study presents a data-driven, privacy-preserving framework for predicting the energy consumption of industrial robots. First, we compare different machine-learning based approaches for the modeling problem and show that high accuracy can be achieved. Second, to protect sensitive data in collaborative workflows, privacy-preserving machine learning (ppML) techniques based on multi-party computation (MPC) are applied to the trained models. The approach enables accurate energy modeling while maintaining data confidentiality, which is critical in industrial settings where intellectual property protection is essential, thus promoting safe and efficient energy optimization in collaborative workflows
Suppression of magnetohydrodynamic interfacial wave instabilities by means of parametric anti-resonance
Full text linkThe high electrolytic currents in aluminum reduction cells can provoke magnetohydrodynamic wave instabilities in the liquid-liquid cryolite-Al interface. Most critical for the safe operation of aluminum smelters is the metal pad roll (MPR) instability, which can set the interface into self-growing rotational wave motions. Such interface instabilities are commonly averted by keeping the poorly conducting cryolite layer sufficiently thick, but at the detriment of the energy efficiency. Mohammad et al. (in: Eskin (ed) Light Metals 2022. The Minerals, Metals & Materials Series, Springer, Cham, 2022; JOM 74:1908–1915, 2022) have recently demonstrated that the cryolite layer can be markedly reduced when adding an oscillating component to the electrolytic current, inhibiting exponential growth of the MPR instability. We dedicate this paper to the investigation of the underlying physics behind this new suppression technique. We analyze the MPR stability using a simplified mechanical model, which reduces the mathematical problem to a set of two coupled Mathieu’s differential equations. The state of stability is calculated both numerically using Floquet theory and analytically by applying the complexification-averaging method. Our analysis reveals that observed stability patterns can essentially be attributed to a simultaneous occurrence of parametric resonance and anti-resonance. We identify ideal system parameters and show ways to verify the rather exotic phenomenon of parametric anti-resonance in MPR experiments
A comparative simulation study of single and hybrid battery energy storage systems for peak reduction and valley filling using norm-2 optimization
No full textThe objective of this study is to address the power imbalance between supply and demand caused by the adoption of electric vehicles and renewable energy sources. Due to power imbalance at the point of common coupling, additional peaks and valleys will be created. The novelty of this work lies in proposing a hybrid energy storage system that combines power-dense and energy-dense batteries, optimized using a Norm-2 approach, to mitigate these imbalances effectively. The methodology involves a simulation study using realistic distribution grid load curves, focusing on two case studies. The results of this study reveal that, with an optimally sized energy storage system, power-dense batteries reduce the peak power demand by 15% and valley filling by 9.8%, while energy-dense batteries fill the valleys by 15% and improve the peak power demand by 9.3%. Furthermore, a hybrid energy storage system outperforms and is useful for multiple grid applications when compared with a single type of energy storage system. The study identifies an optimal capacity share of 40% power-dense and 60% energy-dense batteries as providing an effective balance between power and energy requirements. The findings highlight the proposed system successfully manages not only the highest peaks and valleys, but also intermediate fluctuations caused by renewable energy and electric vehicle integration. During a one-year simulation using a hybrid energy storage system, peak power demand decreased by 11%, peak-to-average ratio improved by 12%, and power variance was reduced by 29%, indicating more stable and efficient grid performance compared to without any storage system
Entscheidungen meistern
No full textEinführende Darstellung vieler theoretisch begründeter und zugleich praxisbewährter Entscheidungsmethode
Vorverarbeitung zur Extraktion von Gaskonzentrationen aus Spektrometerdaten basierend auf neuronalen Netzen
Full text linkDas Erkennen unterschiedlicher Gaskonzentrationen in Spektrometerdaten ist ein gängiges Problem in technischen Anwendungen. Für einzelne Konzentrationen beziehungsweise für eine begrenzte Anzahl an Konzentrationen konnten bereits Techniken etabliert werden, die beispielsweise auch mit der Hilfe von neuronalen Netzen und Deep Learning arbeiten. Bisher konnten für die Identifikation von multiplen Konzentrationen stark unterschiedlicher Gase in dem für unser Projekt geforderten Anwendungsbereich jedoch noch keine genauen Methodiken etabliert werden. Ziel dieser Masterarbeit ist es, unterschiedliche Konzepte zu erstellen, die mithilfe von Deep Learning und neuronalen Netzen bis zu elf Gaskonzentrationen aus einem Spektrometer-Output approximieren sollen. Zunächst soll eine Recherche zu unterschiedlichen Methoden zum Erkennen von Gaskonzentrationen aus Spektrometerdaten unter Nutzung neuronaler Netze durchgeführt werden. Auf Basis dieser Recherche sollen Konzepte erstellt werden, die anschließend auf reale und simulierte Trainingsdaten trainiert werden. Die Ergebnisse sollen evaluiert werden, um das für einen praktischen Einsatz am besten geeignete Konzept zu ermitteln. Das Konzept soll anschließend mit praktischen Spektrometerdaten validiert werden, die das reale Einsatzgebiet direkt widerspiegeln. So soll bestätigt werden, ob das evaluierte Konzept auch in einem realen Einsatzszenario funktionieren würde.The ability to recognise different gas concentrations in spectrometer data is a common problem in technical applications. Techniques have already been developed for identifying individual or limited numbers of concentrations, which use neural networks and deep learning, for example. However, there are currently no precise methods for identifying multiple concentrations of widely varying gases in our specific use case. This work aims to develop concepts that use deep learning and neural networks to approximate up to eleven gas concentrations from spectrometer output. First, research will be conducted into different methods of detecting gas concentrations in spectrometer data using neural networks. Based on this research, concepts will be developed and trained using real and simulated data. The results will be evaluated to determine the most suitable concept for practical use. This concept will then be validated using practical spectrometer data that reflects real-world applications directly. This will confirm whether the evaluated concept would work in a real-world scenario
The barriers women face in leadership positions in Vorarlberg's manufacturing sector
Full text linkThis study explores the barriers women face in accessing leadership roles in Vorarlberg’s manufacturing sector, where traditional norms meet with industrial strength. Despite formal equality policies, informal practices, such as dialect-based exclusion, reliance on local networks, and assumptions around motherhood, limit women’s advancement. Based on eleven interviews, six recurring challenges emerged, including identity conflict, visibility gaps, and rigid work–life norms. Using Reflexive Thematic Analysis, the study highlights how informal norms still define who is seen as a “fitting” leader. Women respond with subtle but effective strategies, building visibility, staying value-driven, and reframing challenges, reflecting tempered radicalism. Findings show that change must go beyond fixing women; it requires trans forming the cultural and structural ideals of leadership toward more inclusive, flexible, and diverse models