DCS Computing (Austria)

Abstract The last decade has witnessed a rapid growth of the field of exoplanet discovery and characterization. However, several big challenges remain, many of which could be addressed using machine learning methodology. For instance, the most prolific method for detecting exoplanets and inferring several of their characteristics, transit photometry, is very sensitive to the presence of stellar spots. The current practice in the literature is identifying the effects of spots visually and correcting them manually or discarding the affected data. This paper explores a first step towards fully automating the efficient and precise derivation of transit depths from transit light curves in the presence of stellar spots. The primary focus of the paper is to present in detail a diverse arsenal of methods for doing so. The methods and results we present were obtained in the context of the 1st Machine Learning Challenge organized for the European Space Agency’s upcoming Ariel mission. We first present the problem, the simulated Ariel-like data and outline the Challenge while identifying best practices for organizing similar challenges in the future. Finally, we present the solutions obtained by the top five winning teams, provide their code, and discuss their implications. Successful solutions either construct highly non-linear (w.r.t. the raw data) models with minimal pre-processing – deep neural networks and ensemble methods – or amount to obtaining meaningful statistics from the light curves, constructing linear models on which yields comparably good predictive performance.

ABSTRACTRacetrack flumes are established experimental settings in ecohydraulics and sediment studies. Their experimental results are often coupled with numerical simulations. The two-equation turbulence closures of the Reynolds-Averaged Navier-Stokes equations are applied widely in such engineering applications. They are preferred for their ease of use and low computational costs compared with more sophisticated turbulence models involving large-scale fluvial simulations. Here, three variants of two-equation models, i.e. the standard k-ϵ model and two variants of k-ω models, were applied to tackle mean flow and turbulence in a racetrack flume and a river reach. Regarding model performances, we found an overall reasonable agreement between simulated and measured mean velocity values. Nonetheless, the simulated turbulent kinetic energy exhibited discrepancies to the measured values. The computational costs of all investigated model variants are comparable. Therefore, there is no significant preference among models. The findings confirm the application of these model approaches for mean flow-related investigations (e.g. habitat modeling) but suggest exercising caution for applications sensitive to turbulent processes.

Offshore structures are founded on submerged foundations. The excavation of submerged foundations in the sea bed is a difficult task to accomplish when it comes to the decommissioning of these offshore structures. The extraction resistance is a lot higher than the pressure acting on the structure due to hydrostatic pressure, earth pressure and its self-weight. Once the extraction begins, a negative pore water pressure is created until inflowing pore water compensates this negative pore water pressure. This depression is hindering the extraction of the submerged foundation. Additionally, the resistance is dependent on the embedment depth of the structure, the soil properties as well as the extraction velocity, which influences the dimension of the negative pore water pressure. The numerical investigation of this dynamic problem is a limitation for continuum based approaches like the Finite Element Method (FEM) due to the occurring large deformations. These results from the soil bed failing under the movement of the structure and hence starting to flow. Additionally, in order to estimate the created depression, the investigation of the water-soil-interaction is crucial, as the change of the pore water pressure plays a significant role. Therefore, it is necessary to analyze the behavior of the soil particles and the pore water pressure. In order to do this, a coupled Euler-Lagrange approach, namely the combination of Computational Fluid Dynamics (CFD) and the Discrete Element Method (DEM), is used. In these simulations on one hand, the liquid phase, e.g. the water, is considered as a continuum, while on the other hand, for the solid phase, e.g. the soil, a particle representation is chosen. Hence, it is possible to compute the particle-particle — as well as the fluid-particle-interactions. The calculations are carried out with the open source software package CFDEMcoupling®, which combines the discrete element code LIGGGHTS® with CFD solvers based on OpenFOAM®. This paper introduces the coupled CFD-DEM approach to simulate the extraction of a submerged plate in the soil bed. In this work, the soil grains are idealized by spherical particles of different diameters. In order to consider effects of dilatancy and contractancy in the soil bed, different relative densities are investigated. Additionally, a variation of the extraction velocity of the plate is carried out to examine the dependence on the creation of negative pore water pressure. For each case, the extraction resistance is calculated. The flow velocity and the pressure distribution in the vicinity of the structure are analyzed.