HFTor - HfT Open Repository (Hochschule für Technik Stuttgart)
Not a member yet
404 research outputs found
Sort by
Deep Learning Model for Wind Forecasting: Classification Analyses for Temporal Meteorological Data
No full textThis paper proposes a multiple CNN architecture with multiple input features, combined with multiple LSTM, along with densely connected convolutional layers, for temporal wind nature analyses. The designed architecture is called Multiple features, Multiple Densely Connected Convolutional Neural Network with Multiple LSTM Architecture, i.e. MCLT. A total of 58 features in the input layers of the MCLT are designed using wind speed and direction values. These empirical features are based on percentage difference, standard deviation, correlation coefficient, eigenvalues, and entropy, for efficiently describing the wind trend. Two successive LSTM layers are used after four densely connected convolutional layers of the MCLT. Moreover, LSTM has memory units that utilise learnt features from the current as well as previous outputs of the neurons, thereby enhancing the learning of patterns in the temporal wind dataset. Densely connected convolutional layer helps to learn features of other convolutional layers as well. The MCLT is used to predict dominant speed and direction classes in the future for the wind datasets of Stuttgart and Netherlands. The maximum and minimum overall accuracies for dominant speed prediction are 99.1% and 94.9%, (for Stuttgart) and 99.9% and 97.5% (for Netherlands) and for dominant direction prediction are 99.9% and 94.4% (for Stuttgart) and 99.6% and 96.4% (for Netherlands), respectively, using MCLT with 58 features. The MCLT, therefore, with multiple features at different levels, i.e. the input layers, the convolutional layers, and LSTM layers, shows promising results for the prediction of dominant speed and direction. Thus, this work is useful for proper wind utilisation and improving environmental planning. These analyses would also help in performing Computational Fluid Dynamics (CFD) simulations using wind speed and direction measured at a nearby meteorological station, for devising a new set of appropriate inflow boundary conditions.Deep-Learning-Modell für Windvorhersagen: Klassifikationsanalysen für temporäre meteorologische Daten. In diesem Beitrag wird eine Architektur von mehreren dicht verbundenen CNN in Kombination mit zwei nachgelagerten Long Short-Term Memory Verfahren zur Analyse von Windmessungen vorgeschlagen. Diese Architektur wird als “Multiple Features, Multiple Densely Connected Convolutional Neural Network with Multiple LSTM” (MCLT) bezeichnet. Insgesamt werden 58 Merkmale in den Eingabeschichten des MCLT unter Verwendung von Windgeschwindigkeits- und -richtungswerten verwendet. Diese empirischen Merkmale basieren auf der prozentualen Differenz, der Standardabweichung, dem Korrelationskoeffizienten, den Eigenwerten und der Entropie. Den CNN werden zwei aufeinanderfolgende LSTM-Schichten nachgelagert. Diese LSTM-Schichten verwenden die gelernten Merkmale aus den aktuellen und früheren Ausgaben der CNN, wodurch das Lernen von Mustern im zeitlichen Winddatensatz verbessert wird. Die MCLT-Architektur wird zur Vorhersage dominanter Geschwindigkeits- und Richtungsklassen am Beispiel von gemessenen Winddatensätze von Stuttgart und den Niederlanden verwendet und evaluiert. Die maximale und minimale Gesamtgenauigkeit für die Vorhersage der vorherrschenden Geschwindigkeit beträgt 99,1 % und 94,9 % (für Stuttgart) bzw. 99,9 % und 97,5 % (für die Niederlande) und für die Vorhersage der vorherrschenden Richtung 99,9 % und 94,4 % (für Stuttgart) bzw. 99,6 % und 96,4 % (für die Niederlande). Die vorgeschlagene MCLT-Architektur zeigt vielversprechende Ergebnisse für die Vorhersage der dominanten Windgeschwindigkeit und -richtung. Damit trägt diese Arbeit dazu bei, Wind-Analysen in Umweltplanungen besser berücksichtigen zu können. Die Analysen helfen auch bei der Durchführung von CFD-Simulationen (Computational Fluid Dynamics), bei denen Windgeschwindigkeit und -richtung als Anströmungsrandbedingungen genutzt werden, um beispielsweise das Potential von Kleinwindkraftanlagen im urbanen Raum abzuschätzen
Understanding the Impact of Measuring and Choosing RFID-Transponders for Applications in Logistics
No full textAutomatic identification (Auto-ID) is the fundament of the Internet of
Things. Besides barcodes and 2Dcodes, Radio Frequency Identification
(RFID) is used. In the different applications in logistics, the objects to be identified consist of different materials, and therefore this chapter provides basic knowledge about testing and selecting the right RFID-transponders for specific substrates
Digitalization of an Indoor-Positioning Lab Using a Mobile Robot and IIoT Integration
No full textIndustry 4.0, the Industrial Internet of Things (IIoT) as well as Smart Logistics depend on locating mobile assets. In contrast to outdoor locating, GPS is not reliable for indoor positioning. Instead, different real-time locating systems (RTLS) are used in industries for indoor locating when there is no chance of obtaining GPS-satellite signals. Students in engineering disciplines should know about the chances offered by and the limits of RTLS, for example through corresponding lab experiments. However, measuring the accuracy of RTLS is a time-consuming task. Our goal is to provide a remote RTLS-accuracy measurement experiment by digitalizing and automating the whole process.
This paper discusses adding remote experiment service to this lab, thus providing access to the lab infrastructure anytime and anywhere. A mobile robot was used to move an ultra-wideband (UWB) transponder and expose it to the RTLS measurement infrastructure. By optimizing the routing algorithm of a mobile robot, the required accuracy and appropriate safety features were justified and the accuracy of the robot reached 2 cm. It also passed all the static and dynamic obstacles with acceptable safety thanks to inbuilt sensors. The remote operation was also done in an IoT environment by implementing the MQTT data transfer protocol. For remote users to be able to operate our RTLS system via MQTT, we developed a software program. When running this program, our DigiLab4U Laboratory Management System (LabMS) is able to send commands to the RTLS system and receive positioning measurements of a mobile object (in our case RTLS tag) via MQTT messages. Thus, the real route and the measured route can be compared and the difference can be analyzed by students remotely
Fatigue behaviour of ultra high performance concrete under cyclic stress reversal loading
No full textFor the first time comprehensive experimental research was performed on the fatigue behaviour of UHPC under cyclic reversal stress loading in tension-compression. During the research project the fatigue strength and the reduction of the uniaxial stiffness due to cyclic loading were examined in 195 force controlled fatigue tests
on bone-shaped non-reinforced UHPC-specimens. Compared to similar previous examinations the fatigue tests were not restricted to 2 million load cycles. Testing an adequate number of tension-compression combinations in this first stage and an additional second stage the indication of approximately linearised limiting curves for
defined number of load cycles in terms of Goodman diagrams is the long term objective of the research project
Photorealistic versus Procedural Texturing of the 3D Buildings in Virtual Smart Cities
No full textThis paper discusses different methods of texturing techniques such as photorealistic and proceduraltexturing e.g., Dynamic Pulse Function (DPF). Based on a predefined XML schema, an unlimited number of layers can be created for windows, doors, and backgrounds. The system can use many textures for a single layer to increase the Level of Realism (LoR) in 3D virtual models. The advantage of this technique is the geospatial database of each layer in XML-schema. It can include the name of the layer, width, height, geometry, and starting point of the object in the layer with respect to the upper-left corner of the façade.
The concept behind DPF, is to use logical operations to project the texture on the background image which is dynamically proportional to real geometry. The process of projection is based on two vertical and
horizontal dynamic pulses starting from the upper-left corner of the background in the down and right directions respectively. The logical one/zero on the intersections of two vertical and horizontal dynamic
pulses projects/does not project the texture on the background image. Orthogonal and rectified perpendicular symmetric photos of the 3D objects that are proportional to the real façade geometry were
utilized for the generation of the output frame for DPF. This produces a very high quality and small data size of output image compared with the photorealistic texturing method. Complex geometries such as
coconut or palm trees can be designed as a single implicit geometry and utilized in the CityGML environment via URL to deal with rendering and lagging problems of visualization. In the current work
the texture of the façade was created based on preprocessed procedural DPF technique and the output image was utilized in 3D modelling as a texture
How Empowerment Can Help to Reduce Change-Related Uncertainty in Young Employees
No full textChange can affect employee work behavior and well-being in a variety of ways. The mediating and moderating factors that produce these effects, however, are not fully understood. This study examines the mediating role of uncertainty on affective commitment and organizational attractiveness and assesses whether psychological empowerment can mitigate the effects of change-related uncertainty. Survey data stems from 971 young German banking sector employees during a change period. Results reveal a negative relationship between change and affective commitment as well as between change and organizational attractiveness, with both relationships mediated by uncertainty. Furthermore, empowerment does not moderate this mediation for affective commitment but does for organizational attractiveness. Thus, there are fewer negative effects of uncertainty on organizational attractiveness when psychological empowerment is high. This has several implications for practice: uncertainty should be counteracted during change processes, for example, through transparent communication, and psychological empowerment should be promoted before and during change processes
iCity. Transformative Research for the Livable, Intelligent, and Sustainable City
No full textThe research goal of the iCity project at the Stuttgart University of Applied Sciences, and of all associated research presented in this book is to support cities in becoming more livable, intelligent, and sustainable. The research results compiled here propose tools, applications and services in the areas of mobility systems, energy, artificial intelligence, smart data, buildings, and infrastructures. The research relates to the relevant areas as well as to information and communication technologies and acceptance studies. Transdisciplinary research with mutual inspiration characterizes the research teams, which are composed of academia, industry, and municipalities. Different experiences, ways of working, and points of view lead to leaving their own single scientific research field behind and move towards life science-oriented transformative research. We believe that many of the results presented in the following chapters will also be useful in the long run when dealing with complex and multidisciplinary technology cycles in terms of sustainable use. The following contributions are organized into four parts: I Mobility, II Energy, III Simulation and Data, and IV Buildings and Infrastructure
Success Determinants of Open Innovation Partnerships between Car Manufacturers and Start-Ups in Germany
No full textIn a context of saturated markets, increasing price pressure, growing competition, changing preferences and consumption patterns, the automotive industry is evolving into a diversified mobility industry. This disruptive transformation challenges German car manufacturers to reinvent themselves and innovate in diverse directions. Since loose open-innovation (OI) partnerships with start-ups offer manufacturers not only flexibility but, more importantly, cutting-edge expertise from outside the traditional industry, they have become the partner of choice since 2010. Although incumbent-start-up partnerships are on the rise, the current literature scarcely addresses the challenges that arise from innovation cooperation with such disparate partners. Therefore, this study attempts to identify the success determinants to provide practitioners with guidance and contribute to close current gaps in literature. Based on a qualitative research design, in form of semi-structured expert interviews, key barriers and drivers concerning leadership, work methods, culture and intellectual property were identified. Considering the start-up´s, the incumbent´s and the interface´s perspectives, the paper provides a conceptual framework that illustrates the interrelations and challenges in open innovation partnerships. The start-up´s group expertise and maturity turned out to be a major driver for initiating collaboration. The key collaborative success drivers are agile work methods, the group´s internal cultural transformation, the interface´s autonomy, and a solid level of trust and openness between the involved parties. Besides practical recommendations, derived by the identified success factors, the paper constitutes a theoretical basis for further research
Digital 3D City Models Towards Urban Data Platform using OGC 3D GeoVolumes API
No full textNowadays, the digital 3D city models are the basis of the urban data platform. It plays an essential role in various fields of industry and research, from urban resource planning, environmental simulation, disaster management, and many more. Especially, there is more and more use of 3D city models in 3D visualization applications. However, one
common issue is the accessibility difficulty according to different data formats generated from different data providers. Accordingly, there is a need for an interoperable 3D geospatial data delivery method to serve data in a standardized way. In this research, we introduce the use of the 3D GeoVolumes API, the open specification from Open Geospatial Consortium (OGC) to maximize interoperability, replicability, reusability, and accessibility of 3D geospatial data.
The use cases had been implemented in the OGC Container and Tiles Pilot (2020) and OGC Interoperable Simulation and Gaming Sprint (2020 - 2021) using the 3D GeoVolumes API to deliver various 3D geospatial data formats. As a proof of concept, the 3D data from GeoVolumes API are visualized and showcased in game engines and web application clients
Comparing Service-Oriented System Management Solutions in Remote and Virtual Laboratory Environments
No full textDigitalized laboratories are gaining importance in the higher education sector. Students are being provided with remote access to physical laboratory infrastructures as well as online access to virtual labs. Due to the complexity of systems in digital laboratory environments, it is often difficult to manage the applications efficiently. Moreover, there can be multiple types of labora
tories with different system configurations. These laboratories need different management solutions based on the heterogeneity of lab systems. Therefore, different approaches are needed to create deployable software units which support multiple architectures.
We compare a microservices approach and monolithic architectures. As
regards production deployment, virtualization and containerization along with their benefits and disadvantages are considered. In our research, we compared Docker solutions as well as the main Kubernetes tools like Minikube, Kubeadm, K3S, and Microk8s. Our goal is to identify solutions that are easy to manage even in heterogeneous hardware environments. Security, high availability, and compatibility with digitalized laboratories are also considered