Universität Stuttgart
Permanent URI for this communityhttps://elib.uni-stuttgart.de/handle/11682/1
Browse
Search Results
Item Open Access Wege zur Ermittlung von Energieeffizienzpotenzialen von Informations- und Kommunikationstechnologien(Stuttgart : Universität Stuttgart, Institut für Energiewirtschaft und Rationelle Energieanwendung, 2020) Miller, Michael; Hufendiek, Kai (Prof. Dr.-Ing.)Item Open Access Task-oriented specialization techniques for entity retrieval(2020) Glaser, Andrea; Kuhn, Jonas (Prof. Dr.)Finding information on the internet has become very important nowadays, and online encyclopedias or websites specialized in certain topics offer users a great amount of information. Search engines support users when trying to find information. However, the vast amount of information makes it difficult to separate relevant from irrelevant facts for a specific information need. In this thesis we explore two areas of natural language processing in the context of retrieving information about entities: named entity disambiguation and sentiment analysis. The goal of this thesis is to use methods from these areas to develop task-oriented specialization techniques for entity retrieval. Named entity disambiguation is concerned with linking referring expressions (e.g., proper names) in text to their corresponding real world or fictional entity. Identifying the correct entity is an important factor in finding information on the internet as many proper names are ambiguous and need to be disambiguated to find relevant information. To that end, we introduce the notion of r-context, a new type of structurally informed context. This r-context consists of sentences that are relevant to the entity only to capture all important context clues and to avoid noise. We then show the usefulness of this r-context by performing a systematic study on a pseudo-ambiguity dataset. Identifying less known named entities is a challenge in named entity disambiguation because usually there is not much data available from which a machine learning algorithm can learn. We propose an approach that uses an aggregate of textual data about other entities which share certain properties with the target entity, and learn information from it by using topic modelling, which is then used to disambiguate the less known target entity. We use a dataset that is created automatically by exploiting the link structure in Wikipedia, and show that our approach is helpful for disambiguating entities without training material and with little surrounding context. Retrieving the relevant entities and information can produce many search results. Thus, it is important to effectively present the information to a user. We regard this step beyond the entity retrieval and employ sentiment analysis, which is used to analyze opinions expressed in text, in the context of effectively displaying information about product reviews to a user. We present a system that extracts a supporting sentence, a single sentence that captures both the sentiment of the author as well as a supportingfact. This supporting sentence can be used to provide users with an easy way to assess information in order to make informed choices quickly. We evaluate our approach by using the crowdsourcing service Amazon Mechanical Turk.Item Open Access Automatische Applikation modellbasierter Diesel-Luftsystem-Funktionen in Motorsteuergeräten(2020) Xie, Yijiang; Kistner, Arnold (Prof. Dr.-Ing.)The continuous development of diesel engines for meeting the legal and functional requirements, e.g. reducing emissions and fuel consumption while taking drivability into account, has led to a significant increase in the number of sensors and actuators required for the engine. For the diesel-air system it means to introduce a turbocharger, a system for exhaust gas recirculation (EGR), an exhaust gas aftertreatment system, a variable valve control, etc. In order to control such an increasingly complex system in diesel engines, ECU-functions are developed by means of a model-based approach. The success of a model-based development methodology is based on a precise and e cient modeling of the relevant engine behavior. Because of the limited computing power of an ECU, a combination of physical models and so-called calibration parameters is usually preferred for engine modeling. The calibration parameters can be scalar or one or two-dimensional empirical models and usual ly have to be determined (calibrated) by experiments on an engine test bench. Typical examples for such calibration parameters are lookup-tables for modeling the cylinder charge (volumetric e ciency) and the e ective area of the EGR valve. In this thesis a procedure is proposed which is able to calibrate the ECU functions for stationary relationships, e.g. in the diesel-air system, automatically and with as little measurement e ort as possible in terms of the number of measurement points. The algorithm runs within the framework of sequential experimental planning, in which Gaussian models with non-stationary covariance functions are used to approximate the relations of interest. For adaptive experimental planning an active sampling strategy is developed based on the concept of mutual information and optimal system inputs (engine speed, fuel quantity, air actuators, etc.) and which determines the resulting operating points, with respect to the input space coverage, the inhomogeneous properties of the relations, the uncertainty of the estimated calibration parameters and the feasibility of the operating points. The method is able to predict the stationary engine behavior, which results from the selected system inputs, by means of the physical structure of the air system and the data-based models of the calibration parameters. On this basis the uncertainties of the application parameters are estimated using extended Kalman filters. The feasibility of the operating point is checked by comparing the predicted system behavior with the engine limits. For validation the developed algorithm was implemented on an engine test bench to calibrate the air system of a diesel engine equipped with high and low pressure EGR, a variable geometry turbocharger and variable valve timing. As a result, using the presented approach, using as little as approx. 130 measurement points is enough to obtain a comparable application quality to that achieved by conventional methods with more than 800 measurement points.Item Open Access Phase-field modeling of microstructure and fracture evolution in magneto-electro-mechanics(Stuttgart : Institute of Applied Mechanics, 2020) Sridhar, Ashish; Keip, Marc-André (Prof. Dr.-Ing.)Item Open Access Towards an underdamped thermodynamic uncertainty relation(2020) Fischer, Lukas P.; Seifert, Udo (Prof. Dr.)A recent result of stochastic thermodynamics is the so-called thermodynamic uncertainty relation (TUR). This relation, appearing in the form of an inequality, bounds the precision of fluctuating currents by the entropic costs that are required to drive the non-vanishing mean of the observable. As a consequence, the relation enables the access to parameters that are not accessible in an experimental setting via the precision of a experimentally accessible observable. For instance, it was possible to bound the efficiency of molecular machines by means of their measurable moments of motion. Albeit being generalized and modified to more general terms and dynamics, the putative generalization of the thermodynamic uncertainty relation to underdamped dynamics where the inertia is not negligible remains a puzzling problem. Although there are convincing indications for the overdamped TUR being valid for underdamped dynamics as well in some systems, a straightforward application can also lead to violations of the bound. This thesis summarizes the efforts towards an underdamped generalization of the thermodynamic uncertainty relation and shows challenges and chances that come along by generalization of the TUR. To this end, the intriguing limitations of the TUR in the underdamped domain are explored and discussed. For instance, the TUR is inherently broken for finite times where the evolution is governed by ballistic dynamics due to the inertia being present. Furthermore, it is possible to improve the precision beyond the overdamped bound in presence of velocity dependent forces such as the Lorentz force induced by a magnetic field. Beyond the limitations of the TUR in the underdamped regime, this thesis gives a thorough analysis of the proof that leads to the TUR in the overdamped regime and discusses the obstacles which have to be overcome to find the sought-after proof that is valid for underdamped dynamics. The method is illustrated by deriving thermodynamic bounds that are, however, not as transparent and often not as tight as the original TUR. Finally, a conjecture for a generalized TUR is presented which is based on the precision of free diffusion and holds for all times. The corresponding bound converges to the overdamped TUR in the appropriate limit and tightly bounds the precision, even in the ballistic regime. Being based on free diffusion this conjecture also puts the interpretation of the original TUR in a different perspective.Item Open Access Elastic parallel systems for high performance cloud computing(2020) Kehrer, Stefan; Blochinger, Wolfgang (Prof. Dr.)High Performance Computing (HPC) enables significant progress in both science and industry. Whereas traditionally parallel applications have been developed to address the grand challenges in science, as of today, they are also heavily used to speed up the time-to-result in the context of product design, production planning, financial risk management, medical diagnosis, as well as research and development efforts. However, purchasing and operating HPC clusters to run these applications requires huge capital expenditures as well as operational knowledge and thus is reserved to large organizations that benefit from economies of scale. More recently, the cloud evolved into an alternative execution environment for parallel applications, which comes with novel characteristics such as on-demand access to compute resources, pay-per-use, and elasticity. Whereas the cloud has been mainly used to operate interactive multi-tier applications, HPC users are also interested in the benefits offered. These include full control of the resource configuration based on virtualization, fast setup times by using on-demand accessible compute resources, and eliminated upfront capital expenditures due to the pay-per-use billing model. Additionally, elasticity allows compute resources to be provisioned and decommissioned at runtime, which allows fine-grained control of an application's performance in terms of its execution time and efficiency as well as the related monetary costs of the computation. Whereas HPC-optimized cloud environments have been introduced by cloud providers such as Amazon Web Services (AWS) and Microsoft Azure, existing parallel architectures are not designed to make use of elasticity. This thesis addresses several challenges in the emergent field of High Performance Cloud Computing. In particular, the presented contributions focus on the novel opportunities and challenges related to elasticity. First, the principles of elastic parallel systems as well as related design considerations are discussed in detail. On this basis, two exemplary elastic parallel system architectures are presented, each of which includes (1) an elasticity controller that controls the number of processing units based on user-defined goals, (2) a cloud-aware parallel execution model that handles coordination and synchronization requirements in an automated manner, and (3) a programming abstraction to ease the implementation of elastic parallel applications. To automate application delivery and deployment, novel approaches are presented that generate the required deployment artifacts from developer-provided source code in an automated manner while considering application-specific non-functional requirements. Throughout this thesis, a broad spectrum of design decisions related to the construction of elastic parallel system architectures is discussed, including proactive and reactive elasticity control mechanisms as well as cloud-based parallel processing with virtual machines (Infrastructure as a Service) and functions (Function as a Service). To evaluate these contributions, extensive experimental evaluations are presented.Item Open Access Die Kornstruktur und der Heißrisswiderstand von Laserstrahlschweißnähten in Aluminiumlegierungen(München : utzverlag, 2020) Hagenlocher, Christian; Graf, Thomas (Prof. Dr. phil. nat.)Die Kornstruktur einer Schweißnaht beeinflusst ihren Widerstand gegen die Bildung von Nahtmittenheißrissen. Im Rahmen dieser Arbeit wurde der übergreifende Zusammenhang zwischen Schweißparameter, Kornstruktur und Heißrisswiderstand beim Laserstrahlschweißen durch analytische Gleichungen beschrieben und das resultierende Modell experimentell validiert.Item Open Access Das Gruppendelphi : eine diskursive Methode zur Ermittlung von Expert*innenurteilen(2020) Kuhn, Rainer; Renn, Ortwin (Prof. Dr. Dr. h.c.)Das Gruppendelphi stellt eine innovative und diskursive Methode dar, die auf Grundlage des klassischen Delphiverfahrens weiterentwickelt wurde. Die Methode wird zur Erhebung von Expert*innenurteilen eingesetzt, um eine Ermittlung von Handlungsempfehlungen, eine Ein-schätzung zu (politischen) Leitbildern und Rahmenbedingungen sowie eine Bewertung von konkreten Maßnahmen zu erzielen. Diese Methode bildet den Untersuchungsgegenstand die-ser Arbeit. Da sich ein Gruppendelphi im Vergleich zum klassischen Delphiverfahren durch die Integration von diskursiven Elementen auszeichnet, wird besonders die Frage danach verfolgt, ob die Qualitätsmerkmale handlungsleitender Diskurse von der Methode des Gruppendelphis eingelöst werden können. Um diesen methodischen Ansatz zu legitimieren, sind sowohl eine theoretische Grundlage als auch ein Bemessungskriterium für die Qualität der Ergebnisse dieser Methode notwendig, die mittels der Theorieklasse mentaler Modelle, dem Linsenmodell aus der Social Judgment Theory sowie der Theorie des kommunikativen Handelns geschaffen werden. Das daraus resultierende Testdesign besteht aus vier Dimensionen: (1) Evidenz, (2) Verständlichkeit, (3) normativen Angemessenheit und (4) grundlegende Funktionslogik. Die-ses Testdesign dient der Überprüfung der Fragestellung anhand von fünf Fallbeispielen einer konkreten empirischen Anwendung der Methode innerhalb verschiedener Forschungsprojekte. So soll diese Arbeit nicht nur eine theoretische Fundierung des Gruppendelphis erzielen und eine Überprüfung der Funktionsweise der Methode ermöglichen, sondern eine Einordnung möglicher Einsatzfelder vornehmen und zu einer verbesserten praktischen Anwendung beitragen.Item Open Access Data-efficient and safe learning with Gaussian processes(2020) Schreiter, Jens; Toussaint, Marc (Prof. Dr. rer. nat.)Data-based modeling techniques enjoy increasing popularity in many areas of science and technology where traditional approaches are limited regarding accuracy and efficiency. When employing machine learning methods to generate models of dynamic system, it is necessary to consider two important issues. Firstly, the data-sampling process should induce an informative and representative set of points to enable high generalization accuracy of the learned models. Secondly, the algorithmic part for efficient model building is essential for applicability, usability, and the quality of the learned predictive model. This thesis deals with both of these aspects for supervised learning problems, where the interaction between them is exploited to realize an exact and powerful modeling. After introducing the non-parametric Bayesian modeling approach with Gaussian processes and basics for transient modeling tasks in the next chapter, we dedicate ourselves to extensions of this probabilistic technique to relevant practical requirements in the subsequent chapter. This chapter provides an overview on existing sparse Gaussian process approximations and propose some novel work to increase efficiency and model selection on particularly large training data sets. For example, our sparse modeling approach enables real-time capable prediction performance and efficient learning with low memory requirements. A comprehensive comparison on various real-world problems confirms the proposed contributions and shows a variety of modeling tasks, where approximate Gaussian processes can be successfully applied. Further experiments provide more insight about the whole learning process, and thus a profound understanding of the presented work. In the fourth chapter, we focus on active learning schemes for safe and information-optimal generation of meaningful data sets. In addition to the exploration behavior of the active learner, the safety issue is considered in our work, since interacting with real systems should not result in damages or even completely destroy it. Here we propose a new model-based active learning framework to solve both tasks simultaneously. As basis for the data-sampling process we employ the presented Gaussian process techniques. Furthermore, we distinguish between static and transient experimental design strategies. Both problems are separately considered in this chapter. Nevertheless, the requirements for each active learning problem are the same. This subdivision into a static and transient setting allows a more problem-specific perspective on the two cases, and thus enables the creation of specially adapted active learning algorithms. Our novel approaches are then investigated for different applications, where a favorable trade-off between safety and exploration is always realized. Theoretical results maintain these evaluations and provide respectable knowledge about the derived model-based active learning schemes. For example, an upper bound for the probability of failure of the presented active learning methods is derived under reasonable assumptions. Finally, the thesis concludes with a summary of the investigated machine learning problems and motivate some future research directions.Item Open Access Kompensationsstrategien von Rückfederungseffekten beim Umformen von hochfesten Stahlblechwerkstoffen(Stuttgart : Institut für Umformtechnik, 2020) Radonjić, Ranko; Liewald, Mathias (Univ.-Prof. Dr.-Ing. Dr. h.c. MBA)Die ständigen Bestrebungen zur Reduzierung der CO2-Emissionen führen in der Automobilindustrie zur höheren Anforderungen an den Karosserieleichtbau. Um diese Anforderungen zu erfüllen, wurde der Schwerpunkt der Entwicklungen der letzten Jahre auf die Reduzierung des Karosseriegesamtgewichts durch den Einsatz von Leichtbauwerkstoffen sowie Blechen mit geringerer Dicke gelegt. Diese Tendenzen führen zu einem verstärkten Einsatz von hoch- und höchstfesten Stahlblechwerkstoffen zur Herstellung von entsprechenden Karosseriestrukturbauteilen mit geringerem Bauteilgewicht im Vergleich zu früheren Baureihen. Solche Bauteile werden in der Regel durch das Tiefziehen oder das ziehende Biegen hergestellt. Nach der Entnahme des Bauteils aus dem Werkzeug nach der durchgeführten Umformung tritt eine Spannungsrelaxation auf, wonach sich ein neues Spannungsgleichgewicht im Bauteil einstellt. Als Ergebnis dieser Spannungsrelaxation tritt die Rückfederung bzw. die dimensionelle Abweichung zwischen dem entlasteten Bauteil und der Referenzgeometrie auf. Dies kann verschiedene Arten der Rückfederung zur Folge haben: Winkeländerung, Zargenkrümmung, Radienänderung und Torsion bzw. Verdrehung von Bauteilzonen. Bei der Herstellung von Karosseriestrukturbauteilen aus hoch- oder höchstfesten Stahlblechwerkstoffen tritt häufig eine hohe Rückfederung von bis zu einigen Zentimetern auf, welche mit derzeit existierenden Maßnahmen nur schwer oder gar nicht erfolgreich während des Umformens oder gar danach kompensiert werden kann. Im Rahmen dieser Arbeit wurden verschiedene Ansätze im Hinblick auf die Reduktion der Rückfederung am Beispiel eines zweifach gekrümmten hutförmigen Bauteils simulativ und experimentell untersucht. Um die erforderliche Genauigkeit der Simulation dabei gewährleisten zu können, wurden zunächst umfangreiche Werkstoffcharakterisierungen der untersuchten Stahlblechwerkstoffe (DP 600, DP 800 und DP 980) durchgeführt. Darüber hinaus erfolgten die Praxisversuche zur Bestimmung des Einflusses der Blechhalterkraft, Geschwindigkeit der Stößelbewegung und Werkzeugradien auf das Rückfederungsverhalten. Basierend auf den dabei erzielten Ergebnissen konnte zunächst festgestellt werden, dass mit der Erhöhung der Festigkeit des eingesetzten Stahlblechs der Erfolg der zuvor genannten Maßnahmen im Hinblick auf die Reduzierung der rückfederungsbedingten Formabweichungen sinkt, und im Fall des Blechwerkstoffs DP 980 sogar fast vernachlässigbar ist. Des Weiteren wurde die Anwendbarkeit von verschiedenen geometrisch basierten Ansätzen mittels einer nach dem Tiefziehen folgenden Nachformoperation im Hinblick auf die Reduzierung der Rückfederung simulativ untersucht. Diese Ansätze beinhalteten vor allem lokale geometrische Änderungen des Bauteils durch Kalibrierung der Radien sowie das Prägen der ebenen und leicht gekrümmten Bauteilbereiche. Dabei wurde festgestellt, dass mit Hilfe solch geometrisch basierter Ansätze der Spannungszustand im Bauteil grundsätzlich lokal beeinflusst werden kann. In diesem Zusammenhang führte der Einsatz des Kalibrierens der Bauteilradien bei gleichzeitigem Prägen der Bauteilzarge zu einer wesentlichen Reduzierung der Rückfederung. Allerdings erfordern solche geometrisch basierten Ansätze häufig signifikante Änderungen der Bauteilgeometrie. Mit dem Ziel, den Spannungszustand im größtmöglichen Bereich des trägerförmigen Bauteils ohne Änderung von dessen Geometrie entsprechend zu beeinflussen, wurde versucht, die Spannungsüberlagerungseffekte mittels gezielt gewählter Werkzeugradien durch wechselseitigen Platineneinlauf während des Tiefziehens zu bewirken. Aufgrund eines solchen wechselseitigen Platineneinlaufes werden jene Werkstückbereiche, welche während des Ziehvorganges in Kontakt mit den Stempelradien kommen, einem mehrmaligen Biegen unterzogen. Dabei werden die beim ersten Biegen über die Blechdicke verursachten Biegespannungen durch das folgende Rückbiegen mit Spannungen mit entgegengesetztem Vorzeichen überlagert, was sich positiv bezüglich der Reduktion der Rückfederung auswirkt. In diesem Zusammenhang wurde in dieser Arbeit auch dargelegt, dass eine optimale Platineneinlaufkinematik definiert werden kann, die zu einer vernachlässigbaren Rückfederung des entlasteten Bauteils führt. Eine solche optimale Platineneinlaufkinematik während des Ziehvorganges wurde am Beispiel eines zweifach gekrümmten hutförmigen Bauteils in Hinblick auf die nahezu vernachlässigbare Rückfederungsmenge erfolgreich belegt. Diesbezüglich wurde die Forschungshypothese nachgewiesen, dass die Rückfederung des Blechteils reduziert bzw. kompensiert werden kann, in dem die während des Umformens im Bauteil verursachten Spannungen mit Spannungen mit entgegengesetzten Vorzeichen überlagert werden. Die in der vorliegenden Arbeit erzielten Ergebnisse stellen einen wichtigen Beitrag für die industrielle Anwendung bzw. Methodenplanung für die Fertigung von trägerförmigen Bauteilen aus hoch- und höchstfesten Stahlblechwerkstoffen dar.