Hochschule Bonn-Rhein-Sieg
Publikationsserver der Hochschule Bonn-Rhein-Sieg - pub H-BRSNot a member yet
7939 research outputs found
Sort by
Current distortion study for hybrid multi-level grid inverter with active neutral-point-clamped 4-Leg topology
No full text
Generative Models for the Analysis of Dynamical Systems with Applications
No full textHigh-dimensional and multi-variate data from dynamical systems such as turbulent flows and wind turbines can be analyzed with deep learning due to its capacity to learn representations in lower-dimensional manifolds. Two challenges of interest arise from data generated from these systems, namely, how to anticipate wind turbine failures and how to better understand air flow through car ventilation systems. There are deep neural network architectures that can project data into a lower-dimensional space with the goal of identifying and understanding patterns that are not distinguishable in the original dimensional space. Learning data representations in lower dimensions via non-linear mappings allows one to perform data compression, data clustering (for anomaly detection), data reconstruction and synthetic data generation.
In this work, we explore the potential that variational autoencoders (VAE) have to learn low-dimensional data representations in order to tackle the problems posed by the two dynamical systems mentioned above. A VAE is a neural network architecture that combines the mechanisms of the standard autoencoder and variational bayes. The goal here is to train a neural network to minimize a loss function defined by a reconstruction term together with a variational term defined as a Kulback-Leibler (KL) divergence.
The report discusses the results obtained for the two different data domains: wind turbine time series and turbulence data from computational fluid dynamics (CFD) simulations.
We report on the reconstruction, clustering and unsupervised anomaly detection of wind turbine multi-variate time series data using a variant of a VAE called Variational Recurrent Autoencoder (VRAE). We trained a VRAE to cluster normal and abnormal wind turbine series (two class problem) as well as normal and multiple abnormal series (multi-class problem). We found that the model is capable of distinguishing between normal and abnormal cases by reducing the dimensionality of the input data and projecting it to two dimensions using techniques such as Principal Component Analysis (PCA) and t-distributed stochastic neighbor embedding (t-SNE). A set of anomaly scoring methods is applied on top of these latent vectors in order to compute unsupervised clustering. We have achieved an accuracy of up to 96% with the KM eans + + algorithm.
We also report the data reconstruction and generation results of two dimensional turbulence slices corresponding to CFD simulation of a HVAC air duct. For this, we have trained a Convolutional Variational Autoencoder (CVAE). We have found that the model is capable of reconstructing laminar flows up to a certain degree of resolution as well generating synthetic turbulence data from the learned latent distribution
Scholarship of Teaching and Learning - A way to identify inclusion opportunities and exclusion risks in digital learning scenarios
No full textTechnical aspects are brought into focus thinking of inclusion opportunities and exclusion risks in digital learning scenarios. However, focussing on technical limitations is not sufficient. This contribution describes another important field of inclusion, namely psychological personality traits. In a longitudinal study at the Hochschule Bonn-Rhein-Sieg (H-BRS), University of Applied Sciences, we accompanied a civil law lecture of a bachelor's degree programme, which had been digitalized because of COVID-19, with empirical Scholarship of Teaching and Learning methods for two semesters. N=55 students from the first measured semester and N=35 from the second one rated different digital teaching methods used in the developed digital learning scenario. Their personality traits according to the five-factor model were measured by using a validated psychometric short-scale (BFI-10). Moderate to large empirical effects of the students' personality traits on the assessments of different digital teaching methods, used in the digital learning scenario, could be observed. Neuroticism values influences the perceptions of the course difficulty and the preference for using an instant messenger as a central communication platform, where students can interact with fellows and lecturers in a way the students are used to in their daily life. High conscientiousness predicts a more regular execution of the weekly tasks given throughout the semester, while higher values in extraversion are associated with a preference for synchronous video conference sessions and active webcams. Higher agreeableness is associated with rating the learning atmosphere as more constructive while low values are associated with perceiving more negative consequences due to the reduced contact to fellows based on COVID-19 restrictions. Correlations between the dimension openness and any ratings of digital teaching methods could not be observed. With this insight into our students' personality traits, we were able to match the digital teaching methods used in our digital learning scenario to the psychological needs of our students, which resulted in a higher inclusion level and a reduction of exclusion risks
Plötzlich digital und noch viel besser: Innovationsbericht über einen digitalen Inverted Classroom
No full textAufgrund SARS-CoV-2 ist eine Rechtsvorlesung für Betriebswirte im Bachelorstudiengang an zwei verschiedenen Standorten der Hochschule Bonn-Rhein-Sieg mit über 300 Studierenden unter Anwendung des Inverted Classroom Ansatzes zum Sommersemester 2020 vollständig digitalisiert worden. Durch die von außen vorgegebene Lernstrategie mit wöchentlichen Arbeitspaketen und die Nutzung einer asynchronen Kommunikationsplattform auf Basis eines Instant Messengers mit adressatenadäquater Ansprache gelang es, Synchronformate auf ein notwendiges Minimum zu reduzieren. Die Ergebnisse der empirischen Begleitung zeigen, dass das neue didaktische Konzept für eine digitale Lehre die unterschiedlichsten Bedürfnisse der Studierenden befriedigte. Insbesondere konnte eine »digitale Lernatmosphäre« geschaffen werden, die von den Studierenden als sehr förderlich für ihren Lernprozess erachtet wurde. Die induzierte Lernstrategie führte zu signifikanten Leistungsverbesserungen. Es wird diskutiert, welche Maßnahmen sich auch für postpandemische Lehre empfehlen
