TY - JOUR A1 - Belaid, Mohamed Karim A1 - Rabus, Maximilian A1 - Krestel, Ralf T1 - CrashNet BT - an encoder-decoder architecture to predict crash test outcomes JF - Data mining and knowledge discovery N2 - Destructive car crash tests are an elaborate, time-consuming, and expensive necessity of the automotive development process. Today, finite element method (FEM) simulations are used to reduce costs by simulating car crashes computationally. We propose CrashNet, an encoder-decoder deep neural network architecture that reduces costs further and models specific outcomes of car crashes very accurately. We achieve this by formulating car crash events as time series prediction enriched with a set of scalar features. Traditional sequence-to-sequence models are usually composed of convolutional neural network (CNN) and CNN transpose layers. We propose to concatenate those with an MLP capable of learning how to inject the given scalars into the output time series. In addition, we replace the CNN transpose with 2D CNN transpose layers in order to force the model to process the hidden state of the set of scalars as one time series. The proposed CrashNet model can be trained efficiently and is able to process scalars and time series as input in order to infer the results of crash tests. CrashNet produces results faster and at a lower cost compared to destructive tests and FEM simulations. Moreover, it represents a novel approach in the car safety management domain. KW - Predictive models KW - Time series analysis KW - Supervised deep neural KW - networks KW - Car safety management Y1 - 2021 U6 - https://doi.org/10.1007/s10618-021-00761-9 SN - 1384-5810 SN - 1573-756X VL - 35 IS - 4 SP - 1688 EP - 1709 PB - Springer CY - Dordrecht ER - TY - JOUR A1 - Ullrich, Sophie Louise A1 - Hegnauer, Mark A1 - Nguyen, Dung Viet A1 - Merz, Bruno A1 - Kwadijk, Jaap A1 - Vorogushyn, Sergiy T1 - Comparative evaluation of two types of stochastic weather generators for synthetic precipitation in the Rhine basin JF - Journal of hydrology N2 - Stochastic modeling of precipitation for estimation of hydrological extremes is an important element of flood risk assessment and management. The spatially consistent estimation of rainfall fields and their temporal variability remains challenging and is addressed by various stochastic weather generators. In this study, two types of weather generators are evaluated against observed data and benchmarked regarding their ability to simulate spatio-temporal precipitation fields in the Rhine catchment. A multi-site station-based weather generator uses an auto-regressive model and estimates the spatial correlation structure between stations. Another weather generator is raster-based and uses the nearest-neighbor resampling technique for reshuffling daily patterns while preserving the correlation structure between the observations. Both weather generators perform well and are comparable at the point (station) scale with regards to daily mean and 99.9th percentile precipitation as well as concerning wet/dry frequencies and transition probabilities. The areal extreme precipitation at the sub-basin scale is however overestimated in the station-based weather generator due to an overestimation of the correlation structure between individual stations. The auto-regressive model tends to generate larger rainfall fields in space for extreme precipitation than observed, particularly in summer. The weather generator based on nearest-neighbor resampling reproduces the observed daily and multiday (5, 10 and 20) extreme events in a similar magnitude. Improvements in performance regarding wet frequencies and transition probabilities are recommended for both models. KW - Rainfall generation KW - Rainfall occurrence KW - Multi-site stochastic weather KW - generator KW - Resampling weather generator KW - Time series analysis Y1 - 2021 U6 - https://doi.org/10.1016/j.jhydrol.2021.126544 SN - 0022-1694 SN - 1879-2707 VL - 601 PB - Elsevier CY - Amsterdam [u.a.] ER - TY - THES A1 - Forster, Florian T1 - Continuous microgravity monitoring of the Þeistareykir geothermal field (North Iceland) N2 - In my doctoral thesis, I examine continuous gravity measurements for monitoring of the geothermal site at Þeistareykir in North Iceland. With the help of high-precision superconducting gravity meters (iGravs), I investigate underground mass changes that are caused by operation of the geothermal power plant (i.e. by extraction of hot water and reinjection of cold water). The overall goal of this research project is to make a statement about the sustainable use of the geothermal reservoir, from which also the Icelandic energy supplier and power plant operator Landsvirkjun should benefit. As a first step, for investigating the performance and measurement stability of the gravity meters, in summer 2017, I performed comparative measurements at the gravimetric observatory J9 in Strasbourg. From the three-month gravity time series, I examined calibration, noise and drift behaviour of the iGravs in comparison to stable long-term time series of the observatory superconducting gravity meters. After preparatory work in Iceland (setup of gravity stations, additional measuring equipment and infrastructure, discussions with Landsvirkjun and meetings with the Icelandic partner institute ISOR), gravity monitoring at Þeistareykir was started in December 2017. With the help of the iGrav records of the initial 18 months after start of measurements, I carried out the same investigations (on calibration, noise and drift behaviour) as in J9 to understand how the transport of the superconducting gravity meters to Iceland may influence instrumental parameters. In the further course of this work, I focus on modelling and reduction of local gravity contributions at Þeistareykir. These comprise additional mass changes due to rain, snowfall and vertical surface displacements that superimpose onto the geothermal signal of the gravity measurements. For this purpose, I used data sets from additional monitoring sensors that are installed at each gravity station and adapted scripts for hydro-gravitational modelling. The third part of my thesis targets geothermal signals in the gravity measurements. Together with my PhD colleague Nolwenn Portier from France, I carried out additional gravity measurements with a Scintrex CG5 gravity meter at 26 measuring points within the geothermal field in the summers of 2017, 2018 and 2019. These annual time-lapse gravity measurements are intended to increase the spatial coverage of gravity data from the three continuous monitoring stations to the entire geothermal field. The combination of CG5 and iGrav observations, as well as annual reference measurements with an FG5 absolute gravity meter represent the hybrid gravimetric monitoring method for Þeistareykir. Comparison of the gravimetric data to local borehole measurements (of groundwater levels, geothermal extraction and injection rates) is used to relate the observed gravity changes to the actually extracted (and reinjected) geothermal fluids. An approach to explain the observed gravity signals by means of forward modelling of the geothermal production rate is presented at the end of the third (hybrid gravimetric) study. Further modelling with the help of the processed gravity data is planned by Landsvirkjun. In addition, the experience from time-lapse and continuous gravity monitoring will be used for future gravity measurements at the Krafla geothermal field 22 km south-east of Þeistareykir. N2 - In meiner Doktorarbeit beschäftige ich mich mit kontinuierlichen Schweremessungen zum Monitoring des geothermisch genutzten Standorts Þeistareykir in Nordisland. Unter Verwendung von hochpräzisen Supraleitgravimetern (iGravs) untersuche ich unterirdische Massenveränderungen, die durch den Betrieb des isländischen Erdwärmekraftwerks (d.h. durch die Entnahme von Heißwasser und Rückinjektion von Kaltwasser) hervorgerufen werden. Als übergeordnetes Ziel des Forschungsprojektes soll eine Aussage zur nachhaltigen Nutzung des geothermischen Reservoirs gemacht werden, von der auch der isländische Energieversorger und Kraftwerksbetreiber Landsvirkjun profitieren soll. Als ersten Schritt, zur Untersuchung der Leistungsfähigkeit und Messstabilität der Gravimeter, begleitete ich im Sommer 2017 Vergleichsmessungen in dem gravimetrischen Observatorium J9 in Straßburg. Aus den dreimonatigen Messzeitreihen untersuchte ich Kalibration, Rausch- und Driftverhalten der iGravs im Vergleich zu den betriebssicher laufenden Observatoriums-Supraleitgravimetern. Nach vorbereitender Arbeit in Island (Aufbau der Gravimeter-Stationen und zusätzlicher Messeinrichtung, Einrichtung der Infrastruktur, Gespräche mit Landsvirkjun und Treffen mit isländischen Partnerinstitut ISOR) startete ich mit meinen Kollegen im Dezember 2017 das Gravimeter-Monitoring in Þeistareykir. Anhand der iGrav-Aufzeichnungen der ersten 18 Monaten nach Messbeginn führte ich die gleichen Untersuchungen (zu Kalibration, Rausch- und Driftverhalten) wie in J9 durch, um zu verstehen inwieweit der Transport der Supraleitgravimeter nach Island die Geräteeigenschaften beeinflusst hat. Im weiteren Verlauf der vorliegenden Arbeit beschäftige ich mich verstärkt mit der Modellierung und Korrektur von oberflächennahen Schwereeffekten in Þeistareykir. Dies umfasst zusätzliche Massenbewegungen durch Regen, Schneefall oder vulkanisch-tektonische Bodenbewegungen, die das geothermische Signal in den Gravimeter-Messungen überlagern. Als Hilfsmittel verwende ich die Datensätze der zusätzlich an jeder Gravimeter-Station eingerichteten Messsensorik und von mir angepasste Modellierungsskripte meiner Gravimetrie-Kollegen. Als dritten Punkt meiner Dissertation untersuche ich die geothermischen Signale in den Gravimeter-Messungen. Gemeinsam mit meiner PhD-Kollegin Nolwenn Portier aus Frankreich führte ich in den Sommern 2017, 2018 und 2019 zusätzliche Schweremessungen mit einem Scintrex CG5 Gravimeter an 26 im Geothermie-Feld verteilten Messpunkten durch. Diese jährlich begrenzten Schweredaten dienen der Verbesserung der räumlichen Auflösung unserer kontinuierlichen iGrav-Messungen. Die kombinierten Ergebnisse beider Messmethoden (der CG5 und iGrav Gravimeter), sowie jährlich im Messgebiet durchgeführter Referenz-Messungen mit einem FG5 Absolut-Gravimeter, komplettieren das hybridgravimetrische Monitoring am Messstandort Þeistareykir. Die abschließende Gegenüberstellung der gravimetrischen Daten mit lokalen Bohrlochmessungen (von Grundwasserpegeln, geothermischen Extraktions- und Injektions-Raten) des Kraftwerksbetreibers, ermöglicht einen direkten Vergleich der beobachteten Schwereveränderungen mit den tatsächlich geförderten geothermischen Fluiden. Ein Ansatz zur Erklärung des beobachteten Schweresignals mittels Vorwärtsmodellierung der geförderten geothermischen Förderrate wird im Abschluss der dritten (hybridgravimetrischen) Studie vorgestellt. Weitere Modellierungen unter Verwendung der aufbereiteten gravimetrischen Messdaten sind durch den Kraftwerksbetreibers von Þeistareykir geplant. Außerdem sollen die gesammelten Erfahrungen des gravimetrischen Messnetzes und Monitorings in Þeistareykir zur Durchführung weiterer gravimetrischer Messungen an dem 22 km südöstlich gelegenem Geothermiefeld Krafla genutzt werden. T2 - Kontinuierliche Schweremessungen zum Monitoring des Geothermalfeldes Þeistareykir (Nordisland) KW - Superconducting gravimetry KW - Geothermal monitoring KW - Time series analysis KW - Þeistareykir Iceland KW - Geothermisches Monitoring KW - Supraleit-Gravimetrie KW - Zeitreihenanalyse KW - Þeistareykir Island Y1 - 2021 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:517-opus4-548517 ER -